#224 – Meracious Mike Manuduction

Dave Jones: This is The Amp Hour Podcast, recorded November 12th, 2014, with special guest Mike Harrison from Mike's Electric Stuff, episode number 224. Welcome to The Amp Hour. I'm Dave Jones from the EEV blog.

Mike Harrison: And I'm Mike Harrison from Mike's Electric Stuff. Hey, Mike. We're Chrisless today. Yeah, he's at Electronica.

Dave Jones: He is. Have you ever been to Electronica? No, I haven't, but I've heard that you have.

Mike Harrison: Yeah, it was either four or six years ago, I can't remember. Right. Yeah, it's a bit of an experience. I saw, I've heard. You made some comment that it's, you know, for a lot, you wonder why companies bother actually exhibiting there, because it's so huge. Yeah. You can understand the big corporate types, but the smaller ones, you just get lost in the, you know, the bigness of it all. But, yeah, and it was, you know, it's certainly something, you know, you want to do once. Yeah, right. Okay.

Dave Jones: Right, so it wasn't worth going back again? If I had the time, I probably would.

Mike Harrison: Right. Yeah, so it's just a time thing. Yeah, that sort of thing, but, yeah, it's quite interesting. Or, you know, it's always useful to talk to, you know, talk to the people on the stands and hear what sort of bullshit they're pushing sometimes. Right. I like the ones where they've actually got real engineers on the stand and not salespeople. Right, yes, of course. Yeah, yeah. You do often actually talk to people that really know what they're doing, and as soon as they realize that you know what you're doing, then, you know, you can get some quite interesting conversations.

Dave Jones: Exactly. But, so how much time and money does it cost you to hop across the pond there over to Germany? It's not.

Mike Harrison: But there's someone I know that actually lives in Munich, so I got free accommodation. Ah, nice. Right, but, so, I mean, a lot of it depends on how early you book, but you probably get changed out 100 quid for a flight, even if you've got to be late. It's not that far. Right. For me, so it's… It's only like an hour's flight, isn't it, or something? Yeah, something like that. It's not that far. Ah, jeez. That's like going to Melbourne. The show itself was, yeah, I'd say it's not unlike most shows, just apart from scale. The only thing that you get is that these specialist halls, although normally, you know, each hall has got some sort of specialization, you quite often find just random stands that are either late bookings or whatever that look like they're in the completely wrong hall. Right. Which can be good, because you can stumble across, you know, interesting stuff you didn't know you were interested in. Right. But the one that was a real experience was a whole hall of Chinese suppliers. Oh. And that was just so annoying. Literally, you know, my normal tactic at shows is you walk down, you know, fairly briskfully, purposefully. Yes. Looking, glancing to see if it looks like it must be, you know, there's anything interesting, and then walking past. That's right. Or if you want to really, really annoy the storeholders, you glance for just long enough, you just catch in the corner of your eye that they're turning towards you, then you walk off. Yeah, yeah. But the Chinese thing, you have no chance. They're like sort of, you know, costing you in the aisles. You know, handing you business cards. Wow. You know, you walk around their stand, literally, the one stand I was walking around, they had a load of connections and stuff. And literally, there's this Chinese sort of girl, literally walking one place behind me with a notepad, just waiting for my, you know, every word. It was just so annoying. I just, you know, shit out, guys.

Dave Jones: Yeah, you don't get that from the Western suppliers, do you? Yeah, they're all sort of laid back.

Mike Harrison: I'm guessing something may have been that, you know, it's obviously cost them a lot to be there, so they put you under a lot of pressure from their bosses to get the sales and whatever. But they just don't understand that, you know, people just don't like that sort of hard sell thing.

Dave Jones: Whereas all your big companies, it's a junket, right? Woo-hoo! Woo-hoo! We're going to the trade show. I mean, it's boring as batshit to man the stand all day. Trust me. Yeah, I've done it. Not in the trade show, but computer shows, I've done it. Yep, yep. Yeah, it's not fun. But still, it's a junket, right? It's better than working, right? Especially working for the man. And, yep. No, it's... And so that's... So you don't think it's good value either for, like, a small firm to attend such a huge show?

Mike Harrison: Yeah, obviously, you know, I'm not a marketing person. I don't... If you've got some really new, amazing product that really benefits from people seeing firsthand, then maybe. Right. But I think just the problem is the sheer scale of it. Yeah. And, you know, I think most product areas, there are probably slightly more specialist-focused, you know, shows. It's just the sheer size of it, you know. You can easily miss out a complete haul if you're not particularly well organized.

Dave Jones: A complete haul.

Mike Harrison: You know? Yeah.

Dave Jones: Yeah. Whereas the show we've got here is, like, half a haul, you know? So, like, if you miss an aisle, you're pretty dumb, you know, let alone a whole haul.

Mike Harrison: Unfortunately, I'm a bit OCD about these things. If I go to something like this, I just have to have literally gone past every single... I have to have seen everything. Same here. Even if it's just walking past and glancing, yeah.

Dave Jones: Yeah, absolutely. Exactly.

Mike Harrison: Because I just thought I might have missed something. Right. So that, you know, it took me two pretty much full days. Wow. And I think I did see pretty much everything. Yeah. I think there were one or two really specialist, like, you know, there's some really specialist bits that I didn't bother going to because they didn't. I mean, some specialist stuff I like because it's quite interesting to see other fields and so on. But some things like, you know, hybrids and... No, not really. It's like in the UK, there's a big exhibition centre in Birmingham and quite often they have other shows on. So I went to an electronic show a few years ago and right next door there was a show that did all the stuff for food packaging. Oh, okay.

Dave Jones: All of the handling systems and all that sort of jazz.

Mike Harrison: Yeah, and the factory automation stuff. And I found some really interesting sensors there. So that was worth it. Nice. I'm always sort of interested in, you know, how other industries do stuff. Yep. You know, it's just seeing the other world. Also, like, you know, like reading other industries, trade magazines, you know, you get all sorts of just the level of detail that you just don't even think exists.

Dave Jones: Oh, brilliant. Yeah, I'd love to get there, but it's just, it's Germany. You know, it's probably as far as I could possibly travel on this planet.

Mike Harrison: It's almost like you drill a hole and you'll probably end up pretty close there.

Dave Jones: You drill a hole through the planet and you pop out in Germany. Not quite, but yeah, I'm sure it's not far off. Anyway, for those who don't know, who wondering who this Mike guy is, I'm sure there's very few, but yeah, we didn't really introduce you. You're Mike Harrison from Mike's Electric Stuff and you're a fellow video blogger.

Mike Harrison: Occasional video blogger. Occasional video blogger. I don't do enough stuff, but I think they haven't got it. I just do this, do it occasionally for fun when I've got something to talk about. I don't sort of feel I have to do something every day or whatever.

Dave Jones: Oh, that's yeah. I wish I could do that. I'm a bit more obligated than you to keep up the videos. Lack of pressure. Yes.

Mike Harrison: Things like real work tends to get in the way.

Dave Jones: Right. But you must work flat out on those art installation projects you get because you must get pushed for deadlines and budgets.

Mike Harrison: Oh, yeah, yeah. I mean, yes. The big stuff less so, but it's the small stuff. I mean, just a couple of weeks ago, I did a job. It was a customer. They were making a pinball machine. It was like a corporate branded thing with their own gameplay and so on. So they had problems with their solenoid control board. So actually, they needed like a new board to control the solenoid. So literally, from being asked to do it to delivering two finished boards was about 24 hours. Oh, what? No, it's a little bit more than that. It was maybe two days, including so it's PCB allowed, getting the boards done, assembling them and driving them down the road to them. So it doesn't happen often, but it does happen sometimes.

Dave Jones: Wow, you must have been, you must have not got any sleep at all. So you had to, did you have to design the circuit as well?

Mike Harrison: It was a pretty simple circuit. It was just a bunch of HC-595s and some MOSFETs. Oh, okay. Right.

Dave Jones: But still, you've got to do that. You've got to lay out the board. And in that case, would you bother doing a schematic in the CAD package? Are you kidding? No, exactly.

Mike Harrison: DRC rules? You can netlist. It gives a crap, right? No, it's fine. No, but I mean, basically, if you're, you know, if you're really cramped on a timescale, there are things that you absolutely need. You need a board layout. You need components. You don't actually need a schematic. No, that's right. You know, that can be squashed out of the critical path. So pretty much almost always. But also, if you're doing stuff, we talked about this briefly last time, but if you're doing like, say, a huge LED array with some weird ass, you know, matrixing and multiplexing, you know, there's no way you can just start with a schematic and done that as a nest list. You know, you've got to start from the layout and figure out which LEDs need to go to which drivers because that's how the PCB layout works.

Dave Jones: Yep, exactly.

Mike Harrison: And okay, you might document it later if you can be bothered. But as these are quite often one-off jobs, that does sometimes not happen at all. Right, no, it's just buggering.

Dave Jones: Did they even demand any documentation?

Mike Harrison: Oh, no, no, no. They just needed this thing to work. I mean, that's all they cared about. They needed this thing that would do the job. So that's, you know, the channel numbers are on the silk screen. What more do they want?

Dave Jones: Right. Oh, geez. You went to the effort to put the channel numbers on the silk screen. I would have just done a single-sided board permanent marker, you know.

Mike Harrison: Yeah. Oh, that's great. No, but I'd say things like, you know, just in LEDs and micro, you know, almost everything these days, I've yet to see any good solution to entering a schematic in a way that says this bunch of pins can go to this bunch of nodes, but I'm going to figure out which one's later. Right. I'd say what I tend to do is, you know, I'll throw the prints down. I'll do the rubber band. So, you know, I've got the layout with all the net list and all the net names on it in PCB form. The bill of materials, again, you know, the part, the, you know, component type and component value are in the PCB. Right. So my bill of materials is taken care of. So that when I do this, you know, the typical anatomy of a quick timescale job is, you know, I'll do the board layout. I'll then export the bomb from that board layout to produce the shopping list. And so, you know, all the parts, all the PCBs, all the stencil. And then like two days later, be assembling on my pick and place. I've also got a very streamlined process to go from my PCB to the pick and place. I've got all the common parts, all the libraries to match that. So, for example, if it's a part I've used before, I know that, for example, that dough is going to be the right way around. Because I know that my library matches the orientation in the feeders.

Dave Jones: And you've made it before, you've assembled it. Exactly. Yep. Yep.

Mike Harrison: Got it. So, yeah, that whole process could be extremely quick. And, yeah, that's one of the reasons I get my own pick and place. Because if you've got to deal with a subcontractor, you've got so much documentation to do. It's a real pain.

Dave Jones: Have you ever fed the pick and place incorrectly and goofed up and put down 10Ks instead of 1Ks or something like that?

Mike Harrison: Not to any significant quantity. Right. Yeah. Okay. Yeah. Worst case, if it's a moderately complicated thing, I'll do the first panel and test at least one board off that panel before I run all the other panels. Got it. But, yeah, most of my stuff, you know, it's unusual for the, for the, for the, like, say, five resistor values in a design. Partly because I'm really anal about trying to reduce the number of component values. Right. Because it's a pain in the bum, like, loading all the feeders. And most of my stuff is digital. So, it's, you know, it's either 100 ohms or it's 1K or 10K or 100K. Got it. Yeah. You can get those round values.

Dave Jones: Yep.

Mike Harrison: And the only time that I usually need another value is, like, for a reference resistor on a switch mode power supply. Right. In which case, like, say, if I need, like, I don't know, a 1K and a 13K, then I'll pull up the 13Ks. Yeah, yeah, exactly.

Dave Jones: And then you wouldn't bother loading the pick and place if you just need that one resistor, right? You might just hand solder that on later, perhaps.

Mike Harrison: Yeah, and if it's, like, a few pound, anything, you know, things which are literally there's one per board on a panel, then, okay, resistors I probably would load because they're a pain. But anything, certainly anything that's a weird shape that I have to spend time tweaking the vision for would almost certainly get hand placed. Got it. It's just quicker to do.

Dave Jones: And what is your go-to size package?

Mike Harrison: 0805 if there are no other constraints. Yep. Otherwise, 0603. I might pick and place it will do 0603 reasonably well. Okay. It'll do 0605, no problem till the cows come home. Yep. But, yeah, if there's no size constraints, then probably 0805.

Dave Jones: 5, yep, absolutely.

Mike Harrison: And 1206 if it needs to dissipate a little bit of power, but that's it, really.

Dave Jones: Got it. Now, for this 24-hour turn board or 48-hour turn project, did you hand assemble that? You wouldn't have bothered picking and placing it, would you?

Mike Harrison: Yeah, I hand assembled that purely because it was only two boards. I just couldn't be bothered setting up the thing. Well, I did – one thing I did get a while ago, which I only – basically, it was one of these things I bought for one job with the intention of using it for another. There was a vinyl cutter. Oh, yeah. Which I originally bought for cutting – because I had a job where I had a really weird PCB outline. I bet it was basically fitting a retrofit board into an existing thing. Yep. So, it was near Christmas, I thought. It's about time I bought myself a present. So, I bought the vinyl cutter with the long-term intent so I can cut my last stencils. But the short-term thing was cutting these weird board outlines, which worked very well. And literally, I hadn't got around to even looking at the process for cutting the outlines until this job came along. So, I thought, oh, I'll give it a try. Because it was one of those marginal things. I could have done it by just dispensing.

Dave Jones: So, you're talking about cutting an actual board? No, no, no.

Mike Harrison: Just the stencil.

Dave Jones: Oh, just the stencil.

Dave Jones: Just the silver paste, yeah. Right, okay.

Mike Harrison: And the first board of this I actually did in my etch tank. Because originally, it was just one board. Then they decided they want to cut the others for spare. So, in fact, that job, come think of it. That was actually a 24-hour job because I did the PCB here in the tank. Got it. But the – so, the stencil cut, I'll just give it a try. And actually, all I did was literally export a DXF from my PCB software, feed it into the software that came in the vinyl cutter. And it actually did a pretty good job without any tweaking, which I was quite surprised.

Dave Jones: And there's no scale in it all. It all just works. Well, no.

Mike Harrison: I'd previously sorted that out. Oh, okay. It's things like – it was cutting – there are SO16 pads with the smallest pads on there. But things like getting the right amount of overcuts so it actually cuts the piece out. And then picking all the bits out with the sticky tape to pick all the bits out. It's not the sort of thing I do routinely. But it's handy to have that good bits so that I can now, if I really need to, completely in-house do stencils and PCBs and pick and place. Nice.

Dave Jones: Yeah, you can. Awesome. So, how much is one of these vinyl cutting machines?

Mike Harrison: It was about 200 and something. It was a silhouette cameo.

Dave Jones: Silhouette cameo.

Mike Harrison: I think it's done a Python script to do stuff out legal on it. But basically, it's like an XY plotter with a blade. It's got a blade with a little – a very light – a good quality bearing on it. So, it really drags the blade across and the blade rotates as it changes direction. Oh, okay. Right. It works surprisingly well considering how simple the mechanics are. But the other sort of handy toy that I got recently – quite a long time ago, I bought a laser cutter one. The really cheap Chinese ones. It was really –

Dave Jones: And they're garbage, aren't they?

Mike Harrison: Yeah. Let's buy the cheapest one I can find just purely to see how crappy it is. Sure, yeah. I mean, the laser side, it was fine. The problem was that they were using this really crappy motor controller, PC parallel port running on two AML 8051. It was just horrific. It would just – you know, you'd get halfway through the job and then just stop and set fire to the piece with the laser on. So, and also, it took a lot of space up. So, I eventually sort of fogged that. But I got a little CNC. It's like a CNC engraver type thing there. There's two really common ones on eBay. There's a 3020 and a 3040, which are – the 3040, it's about an A3 size. It's bad. It's designed really as an engraver, so the vertical travels only about 15 millimeters. But for things like cutting PCB outlines, it is brilliant. It works really well. I've also been using it for cutting out little acrylic cases. I don't know if you saw my smallest thermal imager video. Oh, yes. I did that case just cut out of acrylic. So, that's actually been one of those sort of toys that – and just simple things like making up – I had to do a test jig where all I wanted was like 20 holes at an exact pitch. Yep. Which, you know, doing that manually is a real – it's almost impossible to do that accurately. But this literally, you know, a few minutes and done. And obviously, I can use it for things like cutting front panels and so on. So, that was definitely worthwhile, certainly much more than laser cutting. Of course, the other difference between that laser cutter is you've got a much wider choice of materials. Laser cutting is only really a small handful of materials you can cut without gassing yourself, if nothing else.

Dave Jones: Well, I was going to say, you have to have like a ventilation thing for these things, don't you? Yeah.

Mike Harrison: Yeah. Right.

Dave Jones: So, you can't put them in like an office environment like I'm in here. Not usually. Right. Yeah. Yep.

Mike Harrison: I had like a bit of dry adduct that I hung out the window, but it's still – Right. You know, cutting acrylic, it does get a bit – you start getting headaches.

Dave Jones: Yeah, you would sort of turn on and leave, you know.

Mike Harrison: But, you know, for me, it's much more useful to be able to mill out a PCB outline than laser cut in one dimension, you know, in 2D on acrylic. So, that was definitely a recommended thing.

Dave Jones: Yeah. Any other – so, do you have a 3D printer? That would almost complete the set.

Mike Harrison: No. I mean, the thing about 3D printers is 3D prints always look crap. I've never – yeah, apart from the expensive industrial stereolithography ones. I mean, it really annoys me. You see these projects on Hackaday where they've just made what is basically a project box.

Dave Jones: Yeah.

Mike Harrison: 3D printed. It looks terrible. Where, you know, for three quid from Hammond, you could get one that looks much nicer.

Speaker ?: Exactly.

Mike Harrison: You know, so the thing about the CNC is that you'll cut – you know, you've got the nice factory finish of a sheet of acrylic, for example. And you can cut, you know, PCB – you know, in particular PCB, both PCB outlines and things like test jigs and so on. To be able to just drill that is really – I did try drilling, doing automatic drilling on those – that controller board that I did. Yeah. And the problem is that because I printed the artwork on the laser printer, the accuracy isn't quite as good as the CNC accuracy. So I started drilling holes not quite in the right place. I abandoned that and hand-drilled it.

Speaker ?: Got it.

Mike Harrison: So – It would be nice to get that happening.

Dave Jones: What is your method for making your own boards? Do you use the overhead transparency on the laser printer, positive photo resist?

Mike Harrison: I use tracing paper and a laser printer that works really well because you get really good – the problem with transparency film is you don't tend to get really good solid black because it's too shiny.

Dave Jones: No, no, that's right. Yep.

Mike Harrison: But the tracing paper – I've used drafting film in the past, but that tends to crinkle a bit too much. But the optimum thing I found – this is years and years ago, I've spent ages trying all sorts of materials. And thick tracing paper, 90 GSM or above, because the thinner stuff crinkles too much through the fuser. It's dirt cheap, it's got good stability, and it just works. It's brilliant. And then just stapled together into an envelope to do double-sided. And I don't do nearly as many homebrew PCBs as I used to, maybe because most of my jobs I can afford to do a fast turnaround from my local friend and PCB guy.

Dave Jones: But for example, that 48-hour one, you had to. There was no choice.

Mike Harrison: Yeah, yeah, yeah. And so my local place, literally, if it's really desperate, I can get them artwork before about 11 in the morning and they'll ship out the same day. Oh, nice. And then close enough, I can drive up there in an hour and pick them up if I need to.

Dave Jones: So how much does that service cost?

Mike Harrison: Not as much as you'd. They're actually surprisingly cheap and their tooling costs are quite low. So to get, let's say, one Eurocard done in 24 hours would be, it wouldn't be more than 150 quid all in. Right. For double-sided. Well, that's quid.

Dave Jones: So, you know, but hey, when you're charging a client, that's fine. You wouldn't do it for your own hobby project, of course.

Mike Harrison: I mean, basically, this place, it's got to the point where I hardly ever even bother asking for a quote. I just phone it to them saying, I need this many in this days, and I know it's going to come out reasonable. And also, I've passed them on to a few of my clients. So quite often, what I'll say is, okay, here's the artwork. Send it to me, invoice it to whoever, you know, the customer is. Just saves me dealing with the paperwork.

Dave Jones: Yep.

Mike Harrison: Which makes my life a bit easier.

Dave Jones: I hate paperwork. So you're pretty much fully set up then for doing, I mean, but this is your business, right? Yeah. Is being able to turn projects, like these sort of, you know, lowish volume custom projects.

Mike Harrison: Yes. And also, yeah, doing the hardware and the software and just, you know, they just need a solution. So, you know, it's in a market that I understand and I'm enthusiastic about and everyone knows everyone else. So I've got a reputation as being the go-to guy.

Dave Jones: Got it. Yep.

Mike Harrison: Doing funky.

Dave Jones: Do you end up with more work than you can handle or?

Mike Harrison: It generally tends to work out. I mean, what tends to happen is if I've not got much on, I tend to just slow down a bit and don't get as much done as I would otherwise. Yeah, drag it out. Yeah. But no, it seems to, by accident rather than planning, really, seems to. Got it. Yeah, quite often, I'll just bat something away and say, just now I can't, you know, fingers in the ears, la, la, la, I'm not listening sort of thing. Yeah. Do you have it? Sorry. Okay. The, you know, one issue is the projects that do have a sensible time scale, you know, by the time you've got all the other stuff out of the way and they've made out their mind, they then become urgent as well. Right. By the time you have to get into them.

Dave Jones: Do you have any clients that you're going, oh, that was such a pain in the ass, I'm never working with them again?

Mike Harrison: Not really. I tend to filter those out. I can now generally filter those out before they become clients. Got it. You know, yeah, they give you what is clearly, you know, they haven't thought about it. Yeah. I mean, so I just sort of. They haven't thought it through.

Dave Jones: They've got no clue.

Mike Harrison: I usually get involved at an early enough stage to point them in the right direction saying, you know, no, you know, you're not going to, you know, plaster that power supply into a wall so it can never be accessed again. Right. All this sort of stuff.

Dave Jones: Yeah. It's not, once you've got that engineering intuition, it's not hard to know when a project is going to be viable or not.

Mike Harrison: Yeah. Yeah.

Dave Jones: Which is a perfect segue into, after two years, the saga has finally finished and, well, we think so anyway, what you and I both called the Mu Thermal Imager. They finally ran out of money, apparently. Well, duh.

Mike Harrison: Yeah, I mean, that was just, I mean, that was weird. I mean, I've met, even from the video, I've met people like that and they just, I don't understand them. They seem to think that they can just, you know, get the right people in and make something happen despite the fact they don't have a clue what they're actually doing. Yeah. And, you know, I just recognized almost immediately when I saw that first video of the guy saying, you know, just the way he was talking about it, it just, you know, bullshit. Yeah. This is never, ever going to happen.

Dave Jones: The vibe just comes through. Same here. I got the same vibe and it's just went, nah, it's not going to happen. It's not going to happen. They're not going to be able to pull this off.

Mike Harrison: I'm just amazed that they strung it along for that long, to be honest. I really am.

Dave Jones: I know. It's incredible. You know, I thought, oh, this will last six or nine months and they'll run out of money and it'll all go tits up and they'll vanish. But it dragged out for two years, apparently. They haven't admitted it yet, by the way, for those who've been following this.

Mike Harrison: Well, I mean, Seek Thermal has to be the last nailing that that cough in.

Dave Jones: Oh, totally. Totally. They just, yeah. But even if Seek didn't come along, they still had no chance, I think.

Mike Harrison: It just went so long for not showing anything. Exactly. That was the thing.

Dave Jones: Yeah. It's, nah. You just can't. You can't do that. You can't run a project like that, you know.

Mike Harrison: No, no. It would be interesting to actually hear some inside story from that. It would be. I'm sure a few things might leak out at some point.

Speaker ?: Right.

Dave Jones: And show their real prototypes of when they actually started on it, you know. Because you and I both knew that they didn't have Jack, right, at the start.

Mike Harrison: Yeah. I have a suspicion that maybe they were talking to a sensor manufacturer who maybe believed some of the bullshit. Yeah, right. Of, you know. Of course. Believing they're going to sell, ship however many hundred thousand or whatever.

Dave Jones: Oh, because that's possible. Because you can just show them your Kickstarter campaign and, hey, look at all the hype surrounding this, you know. Yeah. And, yeah, I can picture a sensor manufacturer actually talking to them and being rather excited. But, yeah, they're, you know. But often they're not going to have the engineering and systems intuition to know that they don't have the chops to pull it off.

Mike Harrison: But if you're making thermal imaging sensors, you ought to know whether a potential client knows what they're talking about. I mean, it's ridiculous that they work.

Dave Jones: Oh, yeah, I guess. Yeah.

Mike Harrison: But. Yeah, especially if you're not one of the big boys, you know, if you're one of the small, you know, I'm sure. Oh, right. Yes. Yeah. If you discount FLIR and all the companies that FLIR have bought, there can't really be that many left. No.

Dave Jones: Well, whose sensor do you think is in there?

Mike Harrison: I have no clue. Do you have any clue? I'm not sure there is. I really don't know. I mean, I don't know enough about that market because it's not, you know, it's not the sort of thing you see in DigiKey. Right. So I really don't know. You know, pass. Right.

Dave Jones: Okay. If any.

Mike Harrison: You know, the fact they said it was a readily available sensor and then they changed to another one. Yeah, I know.

Dave Jones: And that was, there were so many, how many red flag, I lost count. I lost the number of fingers on my hand at the red flags in that project. I mean, that just, you know, things that just didn't make sense. You know, it was one of the most bizarre projects I've ever followed, really. Yeah. Anyway.

Mike Harrison: Well, I think it's one of those ones that, you know, a lot of the potential customers wanted it to be true. It's a bit like, you know, sort of psychic, you know, people want it so much they're blinded to all the bullshit. Yeah.

Dave Jones: That's a similar thing. It's not that this project's bullshit. This project's awesome. It's the $300 pick and place machine on the Hackaday prize, right? Yeah. I think the problem is that. It got the most votes. Yeah. Because everyone wanted a $300 pick and place machine. Yeah.

Mike Harrison: And most of them are wrong.

Dave Jones: Yeah.

Mike Harrison: Exactly. Exactly. Tell us why they're wrong. Well, you know, until you've used a pick and place machine, you don't, yeah, there are so many literally just things like, you know, how long it takes to load the feeders, how long it takes to set up. Yeah. People think, okay, you know, I'm really bored with hand assembling these two boards. I want a machine to do it. But they don't think through all the implications of what it takes to do. You know, even if you have the best design, most perfect, amazing pick and place in the universe, you know, you've still got to load parts into feeders. You've still got to export the data from your PCB software. Yep. You know, you've still got to set stuff up.

Dave Jones: And you've got to stand there while it's doing it because, you know, if there's a mistake, you can't just come back five hours later and go, oh, it's stuffed up 10 minutes into it and stopped.

Mike Harrison: But again, even details like, just for example, you know, what my machine will do is if it gets like three missed picks in a row that say fail vision, you'll say, okay, well then just carry on with the next part and just do all the stuff it can. And then at the end say, hey, you know, go and sort this out, which is, you know, it's an obvious thing once you realise it. So I think, you know, apparently from what I gather, the guys behind this have actually used some real pick and place machines.

Dave Jones: Oh, no. Yeah, yeah. No, totally. No, it's quite a professional thing they're doing. It's just, you know, the public, I don't think, understands how viable a $300 pick and place machine is.

Mike Harrison: Yeah. But also, looking at their mechanics, I don't think their mechanics are up to the job. You're using that. Yeah, I know. Using that Delta thing.

Dave Jones: That Delta arrangement.

Mike Harrison: I just can't see how you can get the precision. Because one of my sort of thoughts that if you could figure out a way of measuring the position of a head in three dimensions really, really accurately, you could use cheap mechanics and do a closed loop. But the problem is with picking places, you can't vision the board because it's covered in solder paste. So unless you can come up with some really neat, you know, some sort of laser interferometry sensor or something, you know, basically some way of moving all the difficulty into software instead of expensive mechanics. Now, if you can do that, then that's, yeah, I don't know if it's possible. Right. So maybe it is, maybe it isn't. It probably needs a lot of weight. It probably needs somebody who doesn't know that it's impossible to try and actually do it. You know what I mean? Yeah.

Dave Jones: Something that springs to mind is Jerry's cast AR thing. She's claiming like 0.1 millimeters resolution with the IR tracking and all that sort of jazz, you know? Yeah. Like something novel out of the box like that. Yeah. You know? But still, it's, yeah. But this is going to have the linear, is it going to have the linear accuracy over, you know, a 300 millimeter build area? You know?

Mike Harrison: Well, I mean, it doesn't matter if it's non-linear as long as it's predictable. And you can compensate it. Yeah, exactly. Because you can compensate in software. That's right. In some respects, it's easier than a CNC because it's not carrying any load. True. So, you know, you've got a really hard pickup. All it has to do is pick the component, rotate it, and drop it down. But it has to do it really accurately. But also, it has to be reasonably quick. I mean, a pick and place that does 100 components an hour is a waste of time. It's completely useless. You know, it has to be at least as quick as doing it manually. Otherwise, it's just pointless. Yeah, there's no point. If nothing else, your shoulder place is going to dry out. Yeah, yeah. That's right. But, I mean, the other thing is that, you know, I've yet to see a good solution. It is not a pick and place until you've got tape feeders. You have to have tape feeders. Yes, totally. And, you know, I don't think that's a fundamentally difficult problem. The problem is that a lot of people haven't realized that that's what the problem is.

Dave Jones: Yes.

Mike Harrison: You know, moving an XY head, you know, even, for example, that CNC machine I've got, you could probably almost modify that to do it. Okay, it'd be a bit slow. But without feeders, it's a toy. Yeah, totally. You know, you have to be able to stick a rig. It's pointless. You know, you don't. Yep. You can't be having to stick a, you know, strip of, okay, for, you know, you don't need necessarily everything to be in feeders, but for all the jelly bean stuff, you need to be able to just have it spit components out virtually forever rather than reload the things every, you know, sort of hundred components.

Dave Jones: And people don't realize that you become a slave to that machine. Like, you know, like, you know, like, even at the professional pick and place houses, right, the professional assemblers, they stand there just, man, all they're doing is constantly massaging these machines, making sure that, you know, everything's working. Well, they shouldn't be.

Mike Harrison: If they're doing it right, they shouldn't be. Once it gets, you know, they should literally be able to press the button and let it go and it will flash a light, you know, if it has a real problem.

Dave Jones: But, you know, if you do it in a reasonable run of boards, you're always changing parts, you know, especially like if you're using tubes or something like that, you know, you can only fit, you know. Yeah.

Mike Harrison: That's where it probably becomes difficult actually sizing, you know, if, you know, if it's really well set up, one guy can probably run three or four machines. Oh, yeah, yeah, of course. If the process has been really super set up. But, you know, it's a case of having the, you probably need to be running fairly high volumes to do that because otherwise you've got so much set up time to deal with. Either running large volumes or running, you know, say a job that say, you know, every Monday you run a hundred of one particular job or, you know, some regular schedule. So it's all predictable. You can plan it all. And so like one job, like one customer, you've got his own bank of feeders that, you know, his bank of feeders goes onto the machine, they press the button and that job comes out the other end. Got it. But if you're doing small runs of stuff, there's so much set up involved. But the thing about $300 pick in place, you know, okay, as long as I can spend another $2,000 to add feeders and add all the stuff to make it a real machine, fine. Exactly. If $300 is your entry cost for a machine that can be made usable, great. But, you know, if it's $300 for a toy, you know.

Dave Jones: A useless toy that can't be expanded.

Mike Harrison: Yeah. Yeah. It's a very difficult market section because, you know, below a certain amount, it's a toy. Above a certain amount, it's, you know, your market has really disappeared. But also your actual target market for your own pick and place machine, you know, because, you know, why not ship it out to a subcontractor? Yeah. Yeah. Your market is, you know, that band between stuff that isn't worth or is too much hassle or is too urgent to subcontract. Yep. And is big enough to not just hand assemble it. So you're starting off with a fairly thin band in the market.

Dave Jones: And that's never going to change, is it? I can't see that ever changing. That sort of band.

Mike Harrison: It might move a bit, but... It will change if you can move the entry cost to a usable machine. I think 300 is too cheap. So $1,000, I think, is maybe a bit more of a...

Dave Jones: Totally. If there was a $1,000 one, I would buy one. Right?

Mike Harrison: Just for kicks. Right? Because I think, at that price, I think you can have to make something which is useful. And it's the sort of price level where, you know, a couple of, you know, a few set-up charges to a subcontractor will probably cover it. Yeah. And I think that Chinese one, that's about $6,000 or $3,000.

Dave Jones: What's your opinion on that?

Mike Harrison: My main issue with it is it doesn't have vision. You know, I hear surprisingly good things about it, considering how crude it is. If they get vision on... Because the thing is, vision these days costs peanuts. Yeah, I know. And in terms of the complexity of the problem, as vision problems go, you've got, you know, a fixed position, a very finite number of different things you need to look at, complete control over the lighting and background. You know, all you need to know is rotation and dimensions of, in basically 2D. So, in terms of vision, it is not... You know, I'm sure you could do it with something like OpenCV, you know, without any problem at all. Totally.

Dave Jones: Is it... I think there's probably a niche market there for somebody to develop a vision add-on for that machine, because the base machine's there, and apparently it is, by most reports, it is quite, you know... Yeah. Actually, you can...

Mike Harrison: I've heard that they are talking about adding vision too. I think at the moment it's got some sort of Windows CE or something platform in there. But again, the problem is, you know, because it's a limited market, you know, it's a big investment in time and effort to do that. Of course. And how many are you going to sell? Yeah. You know, okay, you know...

Speaker ?: Yeah, exactly.

Dave Jones: You might sell a thousand if you're lucky. Yeah.

Mike Harrison: And if it's open source, well, who's going to make money on it? And would you trust, you know, reels of expensive components to a software that's been coupled together? Exactly. Open source doesn't exactly have a glowing reputation for reliable, you know, dependable, usable software. Well, exactly.

Dave Jones: Just my microcurrent alone uses a reel that costs $6,000, right? A reel of resistors, a reel of precision resistors. Am I going to trust that to anyone but a professional assembler? Yeah. No. Not really, right?

Mike Harrison: Yeah. Well, again, some of my lead jobs, you know, you're putting down... Well, I've been doing a few little demonstration units for a lead manufacturer I'll talk about in a minute. And one of those, it's a 32 by 32 array, so that's 1,024 leads on each board. So if you then... I was thinking of panelizing that maybe four-up panel. That's still, you know, 4,000 leads. Yeah. Even if they're only 3p each, you know, if you drop that or stick your finger in the solder paste or something, it's a lot... It's a big, big risk. Exactly.

Dave Jones: Yeah. No, you're just doing it in-house. Yeah. I think we're both in agreement there. They're very limited scope for an in-house pick-and-place. You're probably the ideal case where an in-house pick-and-place is actually viable. Yeah. With your standalone contract design business, you know.

Mike Harrison: Yeah. I suppose, yeah, for some prototypes, certain types of prototyping on moderately complex things, maybe, or if you do lots of variants and so on. Actually, one of the reasons I got that is one of the projects I was involved in a while ago was looking at a design where it had something like 100 boards, which were all going to be populated differently to basically produce, say, like graphic outlines on a PC, which we spent a lot of time figuring out exactly how we could actually do it. Right. But unfortunately, the project never actually happened, which is a shame. But it was for that sort of thing.

Dave Jones: Did you get paid for that, or is it one of those...

Mike Harrison: I got paid for doing some demo boards. We did some handmade prototypes and so on.

Mike Harrison: Okay. Yeah, I got paid enough to do those, but I'm the sort of person that I just have to think in background. It's a difficult problem. Sure. I just have to solve that problem if I'm in my own head for my satisfaction that I've figured out how to do it. And sometimes it happens, sometimes it doesn't, but that's just me, you know.

Dave Jones: So just as an aside there, when you're doing jobs like that, do you give a new customer a few hours freebie design evaluation?

Mike Harrison: Yeah, it varies a little bit. Yeah, what generally happens is that, you know, we agree on a rough spec and I'd figure out a price to get to a certain point on that spec. Right. I almost certainly don't charge people.

Dave Jones: So you would spend like half a day on it for free or something like that?

Mike Harrison: Yeah, it depends a lot on all of them. Yeah, if it's an existing customer, I've usually got a pretty good idea of something happening. And in some cases, you know, we'd actually develop something that's a bit speculative. And if it doesn't get used on this job, it'll get used on a different one at some point. And that's happened a few times as well. You know, we've got this hardware pretty much developed. So let's pitch it to another client. Yeah, exactly. It's one of those markets that's not really spec driven. It's, oh, that's a cool thing. What can we, you know, what art can we put on it? Yeah, yeah, that's right.

Dave Jones: So how much of your business is that art installation stuff these days? Most of it?

Mike Harrison: Pretty much all of it. All of it. Okay. But that's partly my choice purely because, you know, since I've been doing that sort of stuff, I find the industrial type stuff just so boring. Right, yeah, totally. It tends to work on a much like anything. I'm pretty much nowadays turned down anything which is for production. Because in terms of just timing, you know, you do the process, then you don't hear anything back for three months, then you've got to do all the production stuff. And then, you know, in which time I've done like 10 other jobs in the meantime and I've forgotten about all the other stuff. I know.

Dave Jones: Yes, I've been a production, designing production test systems for like a decade or more. Yeah, it's not fun, people. Just take your word for it. Don't go there. It's not a glamorous industry. It's not fun.

Mike Harrison: Most of the jobs I do now, it's literally, you know, design it, maybe prototype, build it, and that's it, you know. And that's it. That's it delivered. Watch hands. But, you know, there's other aspects. It quite often means that, you know, you have to push the button on the production run of a few hundred boards. Right. And there's no chance to, you know, to get that wrong. Okay. Of course. Yeah. I will generally insist on doing a prototype, even if it's something I'm quite comfortable with. Yeah. Because there's always things, you know, you stick like a thousand heads on a board and it's pulling like 10 amps. There's always risks of things that you didn't quite expect happening. Exactly. Exactly. Okay. Well, exactly how many capacitors do we need to put on this board to stop it crapping the data out? That sort of thing.

Dave Jones: And while you can like modify and hack the board, you know, if you goof up the PCB, you don't want to do it on a hundred boards, right? Absolutely not. Like, you know, if it's 10, okay, it's not too bad, but a hundred?

Mike Harrison: Yeah. I actually had a customer who was wanting to do, I think it was about three or four thousand. It was a fairly simple thing. It was basically a...

Dave Jones: But still, that's big volume.

Mike Harrison: Yeah. It was something like a badge that would light up when it got a radio signal. Right. And they wanted to turn this around in like two weeks flat. And I just said, look, you know, if you're building 2,000 or anything, you cannot do it without a prototype. Because, you know, if you're building like 20 or 30 or 100 boards, they did a few tweaks. Well, then, yeah, someone's got to work. Yeah, it's okay. You can sit down for a day or two and... If you've got 3,000, you are screwed. There is no way out of that. Yeah. Yep.

Dave Jones: Yeah. You would have to hire a subcontractor to do the mods for you. It wouldn't be doable.

Mike Harrison: You know, you'd be talking about probably delivering like two or three days before the thing had to be shipped out to the end customer. So, you know, it wouldn't be hard. It would just not be possible to do at all. Yep. And so I just said, look, no, just don't forget it. You know?

Dave Jones: So tell us about this new lead project with a couple of thousand leads per board.

Mike Harrison: Basically, this came about in a slightly odd way. I was looking through Mouser trying to find a few things to make up the free order, you know, the free shipping amount, as you do occasionally. And I found some 0402 white leads, which were actually quite cheap. Hang on. Do you get free shipping on Mouser? If you order more than 50 pounds, yes. And DigiKey.

Dave Jones: Oh, okay. I don't think I get – no, mine's always 30 bucks shipping, always.

Mike Harrison: I think basically DigiKey styles, and I think Mouser have followed. Oh, okay. It's really odd because things like power supplies, I can actually buy a power supply from Mouser slightly cheaper than I can buy it from the UK rep from the same company and get free shipping. Oh, interesting. So somebody must be losing something along the line there. I don't quite know.

Dave Jones: So what percentage of your stuff do you buy from DigiKey and Mouser as opposed to, you know, RS in the UK and Funnels?

Mike Harrison: Well, my first stop is for odds and ends rather than production, my order is usually RS – not RS, sorry, Funnel, maybe RS if I can find anything on their damn website. Right. Then Mouser, then DigiKey because Mouser are quite cheaper. And for – basically for production stuff, anything that's more than like pennies, I'll look at all four to see who's the cheapest. Right, yeah. Yeah, for stuff where it matters. If I'm – I tend to – if I have to go to DigiKey anyway, I'll tend to do all the order with DigiKey because, again, it's only out of principle because I'm so annoyed that Farnell – and there are so many little details on their website that they could so easily improve and they clearly – I've told them so many times. Yep. I know. They're not interested in it. So I'm almost on principle by elsewhere just because I don't want to give them the business for prototyping. I mean, I couldn't live without Farnell for prototyping.

Dave Jones: Of course. Of course.

Mike Harrison: Yeah, because literally, if I'm doing a stupidly urgent order, if I place an order today, I know Farnell will come tomorrow, DigiKey will come the day after tomorrow.

Dave Jones: They used to have – well, they still do. They have a same-day service here, Farnells. And I think RS might do too. Like if you order before 10 a.m., you can pay a bit extra and have it delivered same day.

Mike Harrison: I don't think they've got that as a standard service. I'm sure if you ring them up and they can – they're quite a long – they're like halfway down along the country. But I'm sure if it was super urgent, yeah, they would hire a local bike courier and it would probably cost you a couple hundred quid or something to do. Oh, okay. Because they're quite a long way. Oh, okay.

Dave Jones: They're a long way.

Mike Harrison: It would be like a good like two and a half hour drive at least from where I am to Farnells.

Dave Jones: So you don't have the luxury of having anyone that you can just go pick stuff up from a tray counter?

Mike Harrison: No. I mean, RS have a tray counter, but they don't really do any electronic bits. You know, if you want a load of spade connectors or a roller cable, then fine. Right. Yeah, it's fine. But if you want a resistor, forget it. Yeah.

Dave Jones: Yeah, exactly.

Mike Harrison: And the bastards have even stopped opening on Saturdays because what you can do is you can place an order and pick it up the next day from the tray counter. So when the tray counter was opening on Saturdays, that was sometimes an option, but they don't even open on Saturday anymore, which is really annoying because they're about 15 minutes away from me. But yeah, that would be nice.

Dave Jones: Yeah, well, here they're about 20 minutes drive or something. And you can just turn up and go, I would like this part, please. You don't have to pre-order. You can just sit there and wait.

Mike Harrison: Do they carry a decent amount of stock locally?

Dave Jones: It's getting worse and worse every year. But, you know, for your basic stuff, yeah, sure.

Mike Harrison: You know, it's, yep. The basic stuff I tend to have in stock anyway. For the sort of quantities I need for prototyping, you know, I've either got or I can improvise or I can pull on off of a board. But, you know, I've got some of my component stocks been going for probably 30 plus years. So, because it's only me, you know, pretty much anything I'll know if I'm likely to have it or not without even looking. Right. But also I'll know exactly where to look if I, you know, if it's something I've used in a project. Generally, I'm quite good at filtering the leftovers out of the project box into general stock as soon as the project's finished. Got it.

Dave Jones: And we were talking before the show about how you got a huge score of components. Can you tell us about that?

Mike Harrison: The basic, this was a, it was a fairly small subcontract assembler up in Cambridge. They, you know, they went out of business and they did a, it wasn't an auction, it was a sale by tender because they didn't have enough space to get everyone in for an auction.

Dave Jones: And how did you find out about them?

Mike Harrison: I can't, I think it might have been advertised in some of the electronics press and I'm on a few main lists for a few. I used to go to loads of auctions years and years ago, pretty much before eBay. I used to go to quite a lot of factory closure auctions. Sometimes I just wanted to have a look around the factory, even though I wasn't going to buy anything. It was quite interesting to look around. Yeah, yeah. Yeah, totally. To these factories. But also quite often you find that a lot of the bargains are like in the weeds, like this box of assorted stuff that they haven't bothered lotting, or in some cases in the dumpsters outside. You know, they've gone out, you know, the auctioneers, they go in, they clear all the crap off and, you know, shovel stuff they don't think is worth anything into a big pile. And you quite often, you know, I picked that tender sale on the viewing day. They had a big skip out. In the UK we call them skips, not dumpsters, by the way, for you non-UK people. But they had a big skip outside and I went on the viewing day and had a look around and made notes and stuff. And there was, yeah, I picked up about two or three dozen, you know, plastic bin bins and loads of bits and pieces and components. And I got, yeah, I asked them, I said, yeah, is it okay if I have a look around the bit of skip? They said, yeah, no problem, you know, I'm saving them money, I'm creating space in their skip, you know. Yeah, yeah, of course. Yeah. And so this was a tender sale where you just put your written bids in and then they come back later and tell you whether you've won or not. And a lot of the lots were like, all their stores were like big rows of rack, you know, steel racking. So, you know, one lot would be like this bay of like five shelves, everything on that bay of five shelves. Oh, right, yes. Or everything in this box of assorted stuff and so on. So a lot of those components, there was quite a lot of tools and like crimp tools. I got a load of crimp tools and stuff. And, yeah, a lot of the big stuff goes for serious money. Yeah, some of the stuff goes for ridiculous money. Those auctions I've never quite figured out why. Yeah, I know. Some of them, yeah. The bargains are often in the weeds. But this one worked out really well because there were quite a few like bays of components I bid almost like, you know, 20 quid on the basis. Someone might not even bid anything. So it ends up I got the call from them. So I won about, I think it was about 3,000 quids worth of stuff.

Dave Jones: Oh, okay.

Mike Harrison: And some it was like relatively expensive single boxes and others literally bays of components and stuff.

Dave Jones: But you knew that you could flip them for a profit?

Mike Harrison: Yeah, I bid based on what I mean, you know, I'd see like, let's say there's a rack of five shelves and there's one bin that had, say, a load of, I don't know if you've seen those old unit pack of alphanumeric LED displays. And this bin was full of those things. And I knew those were, you know, I could probably get a tenner a piece of those on eBay. So, you know, as long as there's one thing in that rack or either something that was useful or like a cell, it was worth it. At least putting a nominal bid in if nothing else. Of course. Yeah, yeah. So, yeah, I got all this stuff. So I went up there, I borrowed a transit van. This was like a long wheelbase, high top, high roof transit. You know, the sort that you can actually stand up in the back of. And sort of went up there, got it up there sort of as early as I could because I knew there'd be lots of other sort of people running around clearing their stuff and stuff quite often goes missing in those scenarios. And I got there and one of the things I'd bid on, it was a laptop was part, it was something like a device program with a laptop. And by the time I got there, the laptop had disappeared. So I had to work with the auctioneer and he was like really embarrassed. Oh, sorry about this. And I said, well, there's a pile over there that has, that's not been lotted. You know, do you want to swap it? Yeah, yeah. Take that instead. And so there's me and a few other. Unfortunately, I didn't have anyone to help me out. And they just, I was on me and I'd previously bought a big box full of those fold together cardboard boxes. Yeah. Flat cardboard boxes. So I had about, maybe about a hundred of those. And literally I was going, literally, you know, unfold these boxes, unload these shelves into the boxes, close the box into the van, next box. So literally it was a whole day of literally loading stuff into boxes. And this van, yeah, this van, high top long wheelbase transit was full. You could hardly get another resistor into it. I did leave, there was some, you know, lots were like the shelving and the stuff, weren't it? And I just had to leave some of the shelving. I just didn't have to. And at the end of the day, I was actually, the auctioneer wanted to go home. So basically I ended up outside unbolting some of these shelving, the bits of shelving I could get into the van outside.

Dave Jones: And I can picture you driving home with a huge smile on your face with a van full of moot, right?

Mike Harrison: Oh yeah, a van that's, you know, always scraping along the ground. And then, you know, I sort of assembled a lot of the shelving in the garage. And it took about, it must have taken at least six months where I had all this shelving in the garage. I had to sort it all out, figure out, you know, and then eBay out. And the amount of stuff that was going through eBay was ridiculous. It probably must have been about 2,000 lots altogether over maybe a year. Wow. And I ended up with lifetime supply of like pin headers. I've got this enormous like filing box crammed full of every pin header you can imagine. And the other thing about eBay, it's really surprising the things that really sold well. It's really bizarre. I mean, things like little, you know, those little test points you can get on PCBs, like a little plastic ring and a loop of wire. And a hook on, yeah, like a ring on top. I was setting bags of those for going for like 10 quid a piece for a bag of 100. Wow, okay. What on earth? And that one that went really well. Heat sink pads, you know, sheets of like A4 size sheets of heat sink pads. Wow. And they were just flying out the door. It was really, really weird. But no, most of it did get flogged. I'd, you know, I'd probably, I know, made about, I don't know, five or six grand out of the lot and ended up with lifetime supply of various bits. But that was a quite fun experience. But with it, yeah, almost all of those sort of sales tend to be on eBay these days or at least online. Yeah, totally. Very much old school, you know, write it down on a form and stick it in the post tender form type thing. No internet at all. eBay was still going strong, but a lot of the, you know, the traditional business auctioneers hadn't really caught on to it. Right. So, yeah, that was fun. I did another one actually quite a few years before. It was quite a big distributor. And it was, they, there was, I got loads and loads of leads. This must have been at least 25, 20, 25 years ago. And again, that was literally, I came back with a car absolutely full of leads. And a lot of them were just bags of like sort of, you know, just like red leather. Oh, yeah, yeah. This was before the days of blue and white. You know, these were, but all sorts of weird shapes and sizes. But the real score on that, there were a load of electroluminescence matrix displays in there. Oh, nice. Yeah.

Dave Jones: They would get a premium.

Mike Harrison: Well, the sort of thing that was so specialist, bear in mind this, this was well before eBay. This was like before, before the, almost before the internet. Right. So I had loads of these displays. They were quite, you know, quite, I had all the data sheets. They were quite an interesting yellow color. I actually put one in the top of a BBC Micro to make a portable BBC Micro. But that was purely by luck. I just phoned up another distributor for the same manufacturer and said, look, I've got all this stuff. I'll just send you a stock list. And it just so happened that they had a customer that was buying this particular display. So I shifted all of those for like twice what I paid for the entire lot of everything else. Oh, brilliant. And what I did with the LEDs, it was, I took, again, this was really very much before eBay. And we, me and a friend took a stand in a radio rally or Hamfest. And what we did is literally, we filled the car for all these LEDs, laid them all out in boxes. And we just gave people a cup and say, okay, for a pound, you can fill this cup with LEDs.

Dave Jones: Yeah, with LEDs.

Mike Harrison: Excellent. And I've still got a few of them. Most of them, I think, have gone by now. But I've still got a few bags of like weird triangular LEDs and stuff.

Dave Jones: Yeah, yeah. But it's fine.

Mike Harrison: And again, that was another way where I pulled a load of stuff out of their skip, you know, for free. Because they just want to get shot of it. So I haven't really seen many, many sales like that for a while. I'm still on a field mailing list auctioneers. But I think so much stuff goes on eBay these days.

Dave Jones: Yeah, exactly. No, it's hard. Anyway, we got sidetracked there. Grossly.

Mike Harrison: You were talking about the LED project with a thousand LEDs. This one came about in a really unusual way, actually. As I said, I was sitting on Mauser, noticed these 0402 LEDs. Oh, that's fun. Be fun to make a nice little dense array of those. Let's see. I don't know. Maybe 32 by 30. So I bought, you know, 1,500 of these things. They weren't that expensive. This is from a new company to the LED business. It's a name that some of our older listeners may remember. Plessy. Oh, yes. Who used to be very into industrial military stuff. And basically, they are now...

Dave Jones: Technically, I used to work for GEC Plessy. So there you go.

Mike Harrison: I actually worked for a part of Plessy when I was at school as a holiday thing. They had a big sort of factory up the road where they made, like, military radios. Basically, there's one of these companies that's changed hands a gazillion times and split and merged. Yeah, oh, yes. I know.

Dave Jones: It's split up dozens and dozens of ways.

Mike Harrison: Their main business was a semiconductor foundry and custom chips, which is why no one had really heard much about them. But they developed this process for doing gallium nitride on silicon, which basically means they can make blue LEDs cheaper than everyone else. Yeah, they've really reduced the cost of... Right, okay. So they've got a few products out now, including these 0402s, and they're targeting them, trying to target them at, like, the wearables market. So what happened was I bought, you know, a bunch of these LEDs. I was going to make up a little... I made up a little matrix board, which was 1.27 millimeter pitch. So it's a little thing about 40 mil square, 32 by 32. And obviously, I've got... You know, I did all my control magic to do nice graphics and stuff on it. So my intention was I was going to build this, probably show it to a few of my clients, some of whom would actually wet themselves when they see it with excitement.

Dave Jones: Okay, so you're proactively developing projects you think that these art people might like. Absolutely.

Mike Harrison: Loads of projects have come out. Like, you know, here's a cool thing. Let's make something. Show them. They pitch it to the client. But what actually happened, so the intention was to do that and also just to bung an email. Yeah, I'll probably stick it to a YouTube video and just bung an email to the UK office of the company, the lead manufacturer. But actually what happened is it actually sort of preempted that about a week after I ordered them. So I ordered them, they were sat on the shelf and I didn't have time to do anything with them or I was waiting for the boards to come back or something. And I got an email from them and said, we noticed you ordered these leads from Mouser. What are you doing with them? Which I actually was a little bit surprised that Mouser...

Dave Jones: How did they know that?

Mike Harrison: It must have come from Mouser, I'm sure. Yeah, it must have done, which...

Dave Jones: But Mouser are giving your details to the manufacturer?

Mike Harrison: Yes.

Dave Jones: Of who bought what?

Mike Harrison: That's a bit dodgy. Yeah, I think yes and no. I'm in two minds about it. It's their product that, you know...

Dave Jones: Yeah, but that's got to be a breach of confidentiality.

Mike Harrison: Yeah, I think it's a tricky one because it's their product you're buying. So, you know, it's not like they don't know that their product's been bought. So, yeah, put that aside. Yeah, maybe, maybe not. But it was a new product. So, obviously, I doubt that would happen on a routine product. But this is a product that had just gone into the catalogue. So, they were probably quite interested to find it. And also, they're new into the market. So, market research, whatever. So, whatever. But, yeah. So, what happened is they contacted me saying, you know, what are you doing with these leads? We're really interested. I think they may well have seen my website as well before making the contact.

Dave Jones: I was going to say, they knew who you were perhaps.

Mike Harrison: I don't recall the sequence of events. So, I said, yeah, I'm going to do this thing, this lead board just for the hell of it. But, yeah. The original idea was to contact them after I'd done it and say, here, you know, do you want me to do some of these as samplers or demos for you sort of thing? But it happened sort of the other way around. And I said, well, look, I'm making this thing. I'll do it. I'll send you one to have a look at and then decide. So, I ended up, I did actually make a few of these little units they're using as demonstrators, which are. So, they're little about 50 mil square by like about 12 mil thick little, you know, acrylic case. Again, that was where the CNC machine came out really nicely. 32 by 32 white display. Playing video off an SD card. Lithium battery in there. They're quite nice cute little units. I did a few for them.

Dave Jones: And so, these are full-color displays?

Mike Harrison: No, no, no. Just white.

Dave Jones: Just white. Okay. Right.

Mike Harrison: Because the nice thing about white is it's so much more efficient in terms of power drawing. The amount of light you get for the amount of power you put in for a white mono is like a factor of umpteen compared to RGB. But also, if you actually want white, RGB is rubbish anyway because the color mixing is always bad. Oh, yeah. No, it's awesome. But no, a lot of the stuff I've done recently has been white. So, I've done a few of these demo units. They've wanted a few more for their distributors. Yeah. And they're actually exhibiting. There's a UK show called Lux Live next week in London. Oh, okay. Which is a lighting exhibition. But there's also quite a lot of lighting components in there that they'll be using those, probably showing them there. But there's also, there are a few in Electronica as well. So, if you're wandering around Electronica and you see this nice little square, it might be on Mouser. I'm not sure which other distributors are out there, but maybe on Mouser you might actually see some of these things. So, that worked out quite well. I turned them around really quickly. So, I charged them. Obviously, as a company, they're not used to stuff being turned around really quickly because they're like a big industry company. So, they were quite impressed with how quickly I turned these things around. And I thought I had to get the PCBs from EuroCircuits because they're quite fine. They're like a 0.125mm track of space and a 0.15mm drill. Right, yeah.

Speaker ?: That's something you can make in the house.

Mike Harrison: Just like it. So, yeah. It's beyond the cable. It's in my go up the road as well.

Dave Jones: Oh, okay. Right. Jeez.

Mike Harrison: I think it was a 0.15mm drill. So, they were quite small and quite dense.

Dave Jones: That is tiny. Yeah. That is. Yep. Yeah. Usually, like 0.3mm, you know. Like a small house. Yeah.

Mike Harrison: 0.3 is about the normal limit. Not quite laser drill. And again, there's a trade-off between the VIA ring size and the drill size. And the EuroCircuits are quite good, but they're really fussy. They've got this super automated process. But if one track is like 0.001 of a millimetre too wide, they'll bounce it back at you and you've got to redo it. Oh, really? Yeah. I hate it.

Dave Jones: Okay. So, you've got to be very strict with your DRC.

Mike Harrison: Yeah. On those. Okay. This was literally pushing the limits of…

Dave Jones: The boundary of their capability. Yeah.

Mike Harrison: Of basically their finest design rules. And I was pushing it to the… One of the annoying things is they give you a spec for the annular ring, but they add a 0.1mm tolerance on the hole size as well. So, if you're down on small hole sizes…

Mike Harrison: I mean, if you look at the artwork, these look like ridiculously big VIAs, but it's purely because they're 0.1mm hole size tolerance on a… I think I started off with a 0.25mm hole. Right. But I say, okay, fine. Just shrink the hole size down and you get your annular ring. So, yeah. But also, I'm now doing another job for them, which is basically they wanted a bigger demo unit for their larger LEDs. And what they originally wanted was a big rectangular display. I said, don't be stupid. You know, if it's a rectangular display, it might as well be a monitor. Let's do a round one. And especially as their logo is round as well. Oh, okay. Yeah, yeah. So, they could actually do their logo quite nice. So, I've done this. It's a 500mm diameter, done in like four quadrants with white. It's basically a white and… If you imagine, like, it's a 32x32 grid of white LEDs, you know, clipped into a circle, but with then another array of blue offset by half a pixel. So, it's, you know, it actually looks really nice. But sort of being round, it looks quite unusual. There's a really thin board around the edge. And it was done as four quadrants to keep the board size central. It was like 250mm square board so I could run it on my pick and place.

Dave Jones: Oh, I was going to say, how did you fit a 500mm board in the pick and place? Yeah, in four pieces.

Mike Harrison: But also, obviously, being in four quadrants, if it was a simple matrix, I could probably have done a single board layer and then rotated it four times.

Dave Jones: I was going to say, were you able to use the same one and then just physically rotate?

Mike Harrison: The LEDs are slightly asymmetrical, but because each pixel is basically, you've got like a white and a blue next to each other. I had to do two versions. Oh, okay, okay.

Dave Jones: So, you're doing white and blue combo.

Mike Harrison: Yeah. Right. The top right and the bottom left are the same PCB, but they're populated with the blue and the white LEDs the opposite way around. So, that matches up. And the top left and bottom right are a mirror. Now, I thought, well, obviously, all the LED layout, I could just flip the board out and get in the mirror, but obviously, all the drivers are on the back. So, I thought, well, I really can't be bothered to redo the layout. I'll just mount all the drivers upside down. Because I was only making four of these boards. So, I thought, you know, it's way, way easier than relaying out. So, each of these quadrants had 1,500 LEDs on it and 13 drivers and some row drivers and a few other bits and pieces. So, about the only things I actually had to rotate and redo the layout for were some MOSFETs and the connector. All the other SLICs were just mounted upside down.

Dave Jones: Nice. Which just saved a lot of time on board layouts. Yep, just flip them. Yep, flip the chips. Physically flip them.

Mike Harrison: Plus, I'd spent a lot of time. Because I'd already done one iteration of that board and decided halfway through this was the wrong approach and I had to rip it up and start again. And these were like four-layer boards, like three-single layers on the ground plane. But that eventually worked out quite nicely. Then the fun was all the software mapping because obviously with a circle, okay, you start off with basically a rectangular grid and like a block of drivers. So, each driver was running eight by 16 LEDs on the other side of the board. But obviously, once you start cutting the corners off, you know, if you've got a small corner cutoff, then fine. But if you've got a big corner cutoff, you can't get a driver there. So, you have to start stealing the cutoff drivers from the other bits of the curve. Oh, nice. So, it's sort of muddled up. Yeah, it's muddled up at the corner. A lot of the blocks are the same, but everything on the curve, there are some LEDs that come from different rows or different drivers and so on. Oh, nasty. So, it took about a day to sort out the software mapping of that thing. Oh, yeah, yeah. But the nice thing was, once I'd done it for one quadrant, the mapping for the other quadrants was just a mirror and a flip. So, that didn't take it as long. Exactly, easy. But I was literally, I'd had a bit of tracing paper over the display, literally lighting one LED at a time and making it where it was. And then making a mark where it was. Yes, okay. Very nice. So, again, if you're going to Luxe, the other thing they're doing on that, I've got a camera. I'm probably in my next video is connecting a little cheap camera module up to a PIC32 using some sneaky tricks with DMA. So, this has actually got a camera, a phone camera module. And there's actually just enough space to put it in the middle. Where the four boards meet, there's enough of a gap. Oh, okay. Yeah, yeah, you can put a hole in the middle. Yeah, yeah. So, it's like a circular LED mirror, which is quite a nice effect. So, if you're going to Luxe Live, I'm sure you'll see that. I'm just about to ship all that to them tomorrow. So, hopefully. Very nice.

Dave Jones: And so, but it's only a white and blue display.

Mike Harrison: Yes, because they only make white and blue LEDs.

Dave Jones: Right. Okay.

Mike Harrison: Yeah.

Dave Jones: So, what does like an image look like in white, in a combination of white and blue?

Mike Harrison: Well, I don't think there's much intention of mixing it. I think it's one, for example, on the camera mode, having the blue is just like a soft background, which actually looks quite nice. I mean, combining white and blue in the image doesn't make a great deal of sense unless it's like a graphic logo. So, I think the idea is that it'll crossfade. You do an image in white and then you crossfade that to the same image in blue. Fade, right. And for the camera image, obviously, the white's a lot brighter than the blue anyway. Got it. Yep. And so, the camera image, it doesn't make sense to do the blue. So, I'm just giving them a sooty they can load a background image in so they can put some check. I can maybe have their logo as a blue watermark and then have the white camera image displaying. But no, it should look a bit unusual, certainly. Hopefully, it's going to grab some attention.

Dave Jones: So, each LED is individually PWM controlled as well?

Mike Harrison: Yeah. Basically, it's a 16-channel PWM driver and then an 8-to-1 multiplex. So, you've got 12-bit control. So, you can get… One issue with monochrome LEDs is that you need to get a nice gray scale. You basically need to apply gamma correction. So, you need a minimum of 12-bit control. But what you can do is you can start from 8-bit data and then feed that through a correction table. So, you can have… Your 8-bit linear scale looks like a nice white, you know, gray scale. Got it. But in fact, it's pre-gamma corrected. So, you… I know squaring is a good first approximation. But sometimes, you can use a lookup cable. But particularly, if you want a nice… Particularly now, if you do it linearly, you find like the first 32 codes sort of go from white… From like… Off to almost white. And then, the top 75% range, you just don't see any difference. Because it's… Got it. The end is another issue in that if you're looking at LEDs rather than LEDs through a diffuser, because you're looking at small points, the eye tends to saturate. So, the perception of gray scale is actually quite distorted. Even if you take a display and put a diffuser on it, things like color temperature become a lot more notable. So, you can look at a bunch of LEDs and they all look… You know, they're white dots. But you put a diffuser over it, you then start seeing all the differences in color temperature because the eye is no longer saturating on those little points. Got it. So, it's one of the slightly non-intuitive things that you don't find until you actually… The other issue is going back to what I was saying about doing prototypes and so on. Is that, you know, if you've got a board with, you know, this thing got 6,000 LEDs in it, you really don't know how much current you're going to need to put in that to get the right brightness because you've got so many LEDs. You know, if you're… Okay, if you run it at full blast, it looks like the Death Star, which is no bad thing. Yeah, right. Of course. So, you know, things like if you say you've got a LED board that's got a DC-DC converter on it, you need to know how, you know, okay, if in theory these LEDs run at full blast could pull 20 amps, chances are you're not going to need that. So, it probably isn't worth spending the time or money to put a 20-amp DC-DC converter on there.

Dave Jones: Exactly.

Mike Harrison: So, what power supply did you end up using? Again, I don't really know how… Yeah, this is going to be run in an exhibition. So, I tend to design most of my stuff so it can run at full brightness if necessary, but it almost never does. And so, I mean, if you turn this over there on this board, it pulls about 40 amps at 5 volts. Oh, okay. Jeez. And it doesn't actually get particularly warm because it's spread out of, yeah, it's a 500-millimeter diameter.

Speaker ?: Yeah, yeah, right.

Mike Harrison: So, it's spread out of quite a large area. But I find with most sort of graphic content, it's, you know, you're lucky if you get to 25%, but also because of the gamma correction in that, you know, a quarter of the power looks half as bright. So, unless you're doing really solid graphic content, anything that's vaguely photographic or smooth or textured, the power constantly just shoots way down. So, what I quite often do is, if there is, let's say, for example, there is a DC-DC limit, I'll put some code in the driver that sums up the total power. And if that's going to be too much, it scales the whole thing back automatically. So, you know, you can't physically overload it and it will scale back gracefully to the point that almost, you know, no one is ever going to even notice it. But it means you're not going to get things tripping out or doing anything crazy.

Dave Jones: Terrific. And the reason you're able to do all this stuff so quickly is because you've done it countless times before.

Mike Harrison: Yeah, I know what all the problems are. I know which drivers are. You know, I've got one of my sort of go-to drivers is made by a Taiwanese company called Macroblock, which are, they're similar to some of the TI-16-bit drivers, but they're a lot cheaper and there's a few very subtle ways in which they're better. So, you know, I buy those by the real of 2,000 at a time. Right. So, again, you can't get those in less than three weeks lead time. But, you know, I know I've almost always got like 2,000 on the shelf.

Dave Jones: So, got it. Yeah, yeah. So, you can turn any project. Yeah. Yeah.

Mike Harrison: Easily. And obviously, I've already got all the code to drive. Yeah, they're a bit fiddly to drive. So, I've got code on the PIC and the AVR and whatever to drive them. So, a lot of the time it's, but also, as well as, let's say, doing a display board, it's the other stuff. It's like the test software to send the data to it. So, you can, for example, work out all your pixel mapping and so on. So, again, I've used, you know, standard pro. I tend to use like RS485 or TTL. Right. Serial, like, you know, up to 6 megaboard to drive these things. But it means I've already got this stuff that I've hacked up in Visual Basic years ago that can take an AVI file and spit it out to this matrix, having done all the address mapping and everything. So, in terms of demonstrating, getting stuff up and running. And, you know, I've now got stuff for doing all the SD card playback and, you know, loading a mapping file. Now, what I can now do is I can actually give someone a player that will play off an SD card and they can generate their own mapping file. So, it reads the mapping file off the SD card that figures out. It takes your rectangular AVI file and you say, okay, this pixel is displayed at this node on this bus on the actual whatever the output format is. And I'll use that in countless installations.

Dave Jones: And this is why you're probably very smart to stick with a niche field like this and only accept jobs in that field because you can then be more impressive that you can turn, yeah, you've done most of it and you can impress clients, turn it over.

Mike Harrison: And it's low risk as well. Exactly. It's something I know I can do. I know, you know, I know what the issues are. And if the client says, you know, I can give the client a reality check in terms of things like, you know, power and brightness and whatever. And, you know, there's some subtle issues that I'm actually surprised that some of my clients are supposedly qualified lighting designers. But the best one was, this was actually a slightly client that was maybe not quite as familiar with it. But they did this, basically it was a meeting table made out of Corian. Do you know what Corian is? No, what's that? It's basically, it's a blend of acrylic and silica. It's used for things like kitchen worktops. It's a really nice, very, it's a very nice high quality. It's all, it's like a bit like a marble sort of feel. So it's quite expensive, but it's, if you, if you machine it down, it's quite translucent. So if you put like lighting behind it, you get some really interesting effects. So basically they made this big corporate meeting table out of Corian and it must have cost them well into four figures just for the piece of material.

Dave Jones: Wow.

Mike Harrison: Yeah. And what they did is they machined the back out of it to make like a bar graph out of LED strip. The idea was, this thing was called a timetable. So the idea was to keep meetings trudging along. Oh no. At the start of the meeting, you press a button and this bar graph appears on the table and counts down over like 30 minutes or whatever and then sets an alarm off when the meeting's finished. Which, you know, it's a quite nice concept. They had like some quite cool speech effects and there's a special button you can press. They had the speech so it would like, it would say like five minutes, 10 minutes of speech. Got it. And there was this secret button you could press which put it in a disco mode where you played Saturday Night Fever and all the lights started flashing. Yeah, yeah, right. But basically they, you know, they were machining this out in the factory where they did all the Corian machining and they showed the client and they had some made strips. They said, yeah, that looks fine. In this factory environment, they said, yeah, it looks brilliant, great. Come to installing it, it goes in a corner office with windows on two sides. And guess what? It isn't bright enough. Yep, yep. Because the various people didn't realise that the actual difference in light level between indoor and at-natural lighting. Oh, it's huge. It's massive. Because the eye is so good at compensating. I know. If you didn't know and didn't use a light meter, you wouldn't know that it's like an order of magnitude, you know, or several orders or maybe two.

Dave Jones: This is why people don't realise when you're shooting video, there's a vast difference between shooting video indoors and outdoors. Vast difference in the performance of the camera in terms of noise level. Vast difference.

Mike Harrison: So at the last minute, we doubled up on the lead strips and they sort of grudgingly accepted it.

Dave Jones: But some people ought to have known better about things like that.

Mike Harrison: Yep.

Dave Jones: Oh, traps. Yep. Oh, goodness. Shall we talk about briefly something that's on the list here?

Mike Harrison: Yeah, what's that?

Dave Jones: Which I find personally interesting. Apparently, there's some capacitor price fixing going on. And I'll link this in. But apparently, all the major manufacturers, I'm talking Panasonic, Sanyo, Teo Uden, NEC, Kemet, Nippon Chemicon, right? AVX, Rubicon, Vichay, right? Everyone, basically. Everyone has apparently colluded. I don't know how long this has been going on. But they've colluded to price fix the market on capacitors.

Mike Harrison: I can only imagine that applies to the more experience. Because things like your box-in electrolytics are so cheap. I just wonder. Yeah, yeah. But when you start looking at stuff like the aluminium polymer and tantalums. Oh, yeah.

Dave Jones: Those really low ESR, top quality, top brand ones.

Mike Harrison: I'm just saying, in the performance of those, they've got like millions of ESR. They're just incredible capacitance. But they're like, you know, one pound 50 each. But they can cost $5, $10 each. Yeah. That's a cheap one. On this big round lead board, my default, I tend to use ceramics. But the problem is, if you've got PWM ceramics, the things buzz. You actually, you've got these LEDs PWM yelling.

Speaker ?: Oh, yeah.

Mike Harrison: They are seeing. Yes. And I was actually getting quite that thing. So I yanked the ceramics off and just stuck some 100-off polymer caps on there. And they just killed it stone dead. It just worked beautifully. But, yeah, in terms of pricing, I can't imagine it's worth price fixing a 5p electric. Maybe. No, I don't know.

Dave Jones: No, no, it's not. It's got to be the high-end stuff, sure.

Mike Harrison: I'd imagine it's the more expensive. I can't understand why polymer caps are that much more expensive. But, yeah, I'm not heard about that.

Dave Jones: Yeah, no, I hadn't heard either. But so thank you for whoever put this on Reddit. And basically, a company called Chip Tech Limited is the complainant. And a class of similarly, oh, right, no, they sort of started it. But then it's a class lawsuit or whatever, however, you know. So they've got lots of all these, you know, purchasing who are purchasing these components. They all started this lawsuit against all these companies for price fixing. And apparently, one of them has admitted it, rumored to be Panasonic, according to this law firm article from this law firm. Yeah, Panasonic have come out and admitted that they've done this, but they want amnesty from prosecution from the U.S. and Chinese authorities. Yeah, we'll dub the other ones in. Thank you very much.

Mike Harrison: Well, it's got to be at least two of them doing it if they're price fixing, isn't there?

Dave Jones: Yeah, yeah, exactly. And they're going to, apparently, they're going to dub them in and admit to it. And it's probably that open and shut case. And, yeah, so that's really fascinating. There you go.

Mike Harrison: It must be a bit hard to figure out what the actual damage is. Yeah, the amounts involved are, though.

Dave Jones: Oh, yeah, I have no idea. But that's illegal in, like, that's under federal law, like actual price fixing. So, yeah. So you can get a federal car.

Mike Harrison: Yeah, because some of the flash manufacturers or memory manufacturers get done for that not too long ago. Yes. I don't remember.

Dave Jones: Yeah. Yeah, there's, I'm sure. But in capacitors, who would have thought? Right? Like, you know, it's such an innocent sort of industry.

Mike Harrison: That's what they think. Hmm. Yeah, yeah. No one notices capacitors. Let's, uh. Yeah, yeah.

Dave Jones: But these are buyers, all buyers. They're, they're really sticky, you know, every last cent, you know, and they're going, geez, these prices, you know. And, and they would see patterns, right? All of these buyers would see patterns. Huh? I'm getting them from five different manufacturers and they're suspiciously all just gone up in price recently, you know. And, like, yeah, they probably put two or two together and, and realize that this is going on. But it takes balls for, you know, somebody to actually start a lawsuit against that. So, but apparently.

Mike Harrison: Yeah, and I have no idea all the sort of mechanics. Yeah. I'm sure the lawyers are the ones that will make the most money out of it.

Dave Jones: Oh, I know. Of course. They always do. But anyway, that, that's a fascinating tidbit from the industry. So, there you go. I like that.

Speaker ?: Yeah.

Mike Harrison: I'm surprised I don't see anything in the, I still get some of the, you know, the paper press. So, I'm surprised I haven't noticed anything about that. But, uh, I'm sure it will probably all start hitting the fan at some point. I do sometimes wonder on, yes, things like the lower end passes, how people can make money making resistors and stuff. You know.

Dave Jones: Oh, they, yeah. They, they can't. It's a, it's a, it's a, it's a, it's got to be a lost lead in business.

Mike Harrison: Can it be, you know, some companies that do little else. So, they must be making money out of it. I know. I know. But, you know, sometimes you go around a trade show and you see this poor little guy on this resistor stand. You feel sorry for him. How can you get excited about selling and making resistors? I know. About resistors, yeah.

Dave Jones: Like, a company like, oh, no, Rome's a bad example because they make chips as well. Yeah, they, they do also. Well, who's like an only, who only manufactures?

Mike Harrison: I think most of them do. I think they do other like ceramic-y stuff, you know, stuff that uses similar tech. But even so. Right, yeah. But it's one of those old things that.

Dave Jones: Well, they probably make their money on like the precision, you know, the really high-end precision ones that cost 10, 20 bucks a pop.

Mike Harrison: Yeah, but it is one of those things that everybody needs, you know, we couldn't live without them. But they're so cheap, there must be just so much competition that, you know, you can buy, you know, it's literally when I'm doing a production order. If I need more than, I think, it's usually about around the 700 or 800 mark, it's cheaper to buy a reel. Absolutely.

Dave Jones: Yeah. I know. It's, well, what is a reel of resistors called? You can get it for like five bucks, can't you?

Mike Harrison: Yeah, it's about that, yeah, yeah.

Dave Jones: Like for a reel of like 5,000 resistors.

Mike Harrison: Yeah, and that's 1% as well. Yeah, yeah, yeah, exactly. I know. I don't know why they even bother making like 5% resistors.

Dave Jones: Yeah, I know. It's just, I know, it's a mad, the market's just, yeah, crazy. I don't know how they can do it. And, and.

Mike Harrison: I'd love to see the factory. And capacitors as well. I'd love that much. Oh, yeah. I'd love to see the factory and see the machines. I'd love to do a tour as well. Pitting out machine gunning, you know, capacitors. It must be millions a day, the productions.

Dave Jones: Oh, it's got to be. It's got to be. Oh, I know. It'd be fantastic. But, yeah. Anyway, if we ever get a chance, yes, we must shoot some video at a resistor or capacitor manufacturer. Absolutely. Boy, all right. Well, we're, our hours, our amp hours well and truly up. That's great. Mike, we'd love to.

Mike Harrison: Where did that go? Maybe.

Dave Jones: Well, it could always come back. Yeah. Yeah. We can always come back. You're always, you're welcome back any time either of us aren't here. But it's a time zone difference because it's what, 11 p.m. over there or something?

Mike Harrison: Yeah, but that's fairly normal working hours for me.

Dave Jones: Oh, okay. Yeah.

Speaker ?: Right.

Dave Jones: You're just, just getting started for the day, right?

Mike Harrison: Yeah.

Dave Jones: How many hours a day's sleep would you get?

Mike Harrison: Oh, I don't know, about seven-ish, seven or eight, something like that. Oh, that's, that's not bad. Yeah, because of course I don't have to get up early for anything, so I'm quite often just get up about, you know, nine, half, nine or whatever, having sort of, you know, sort of work through.

Dave Jones: But you've got the advantage of living and working in your lab, right? That's, yeah, exactly.

Mike Harrison: Yeah, 30 seconds walk down the garden.

Dave Jones: Well, yeah, down the garden, exactly. So, yep.

Mike Harrison: Yes. Got a few very good. The luxury. We've got time for a few viewer questions. I put a thing on YouTube saying. Oh, yeah, no, sorry. Yes, yes. Viewer questions. Yeah, please. About half of them were actually already answered in my previous appearance, so go and listen to that. Oh, okay.

Dave Jones: Well, let's do the ones you haven't answered. No, just a quick thing.

Speaker ?: Go for it.

Mike Harrison: When are you building your big boy power supply? We want flames. I think Photonic Induction's got that covered, so I've got no immediate plans on that one. Right. Just make those bastards talk about electronics for a change. You've got this. We've done a bit of that.

Dave Jones: I think we do that. Yeah, mostly 90% today.

Mike Harrison: Yeah, what else is we going? Weirdest and most interesting. It's hard to say that. I mean, again, I discussed a few of those last time.

Dave Jones: Yeah, we talked about oddball projects, didn't we?

Mike Harrison: Yeah. Yeah, yeah. Oh, here's a good one. Have FLIR stroke Sik offered you a job yet? Ah, right, yes. The answer isn't actually completely no. I actually had quite an interesting conversation with a guy at FLIR after the FLIR one tear down, and everyone at FLIR really enjoyed it, and they sort of said, oh, yeah, he's got that guy. Maybe we should hire this guy, sort of thing. So it wasn't quite a job offer, but… Right. It was rumblings of it. Right.

Dave Jones: But you wouldn't take it anyway, would you? Well, I don't need a job, yeah.

Mike Harrison: No, exactly. I think it might have been semi-serious if I'd not said, well, I don't need a job, sort of thing. Interestingly, their response to the E4 business was congratulations, which I said, I'm not really expecting that, but I'll take it.

Speaker ?: Right, yeah.

Dave Jones: Well, yeah, I'm a bit surprised by that, because they are very paranoid about their security on that sort of stuff.

Mike Harrison: Well, yeah, not enough to have fixed it really quickly, though. No. It took them a long time to fix it.

Dave Jones: True, but they have fixed it. What's the latest update on that FLIR E4? Is it still hacking?

Mike Harrison: The current status is the latest fixed. The resolution has been hacked again, but not all the extra functionality is the current status. Right, right. In post number, it must be up to about 6,000 posts or whatever. Oh, that's crazy. That's really nice. I've had almost no involvement since the original thing. I mean, people much cleverer than me in terms of looking at encryption have done all the hard work on that. You know, I just lit the blue touch paper and everything else. Yeah, that's it. A bit like when I tweeted the thread on the FTDI thing. Right, yeah. I think everything has pretty much been said on that that's going to be said. Yeah, that's all the FTDI thing.

Dave Jones: Yeah, that was, yeah, because it was your tweet that started the whole Twitter thing.

Mike Harrison: I wouldn't change what started it. I'm sure it would have happened anyway, but I think I sort of had some small part in the plus the backwards and forwards tweets between me and FTDI that they then deleted, I think. Yes, they then deleted.

Dave Jones: That's just ridiculous. That's a joke. Why would you?

Mike Harrison: I don't know if they've actually done anything yet. They said they were going to do stuff. I don't know if they've actually changed the driver yet. I have to investigate that.

Dave Jones: Well, they announced they would. I don't know if it's actually happened.

Mike Harrison: They pulled it from the Microsoft update, but whether they've changed the… Yes, that's right. I haven't even.

Dave Jones: Yeah, no, I think it's still a work in progress.

Mike Harrison: I think that one, on Hacker Day, I think it was either 500 or 600 comments, which I don't think I've ever seen happen before on any subject. Yeah, exactly. What's your favourite microcontroller FPGA at the moment and why? I've actually got a couple of quite good answers on that. FPGA Lattice Mac XO because it's got on-chip flash, on-chip regulator, and on-chip oscillator, reasonable range of sizes. They do a 32-pin QFN version, which unfortunately doesn't have any memory blocks or PLLs, and you have to go to 100-pin. Oh, right. But they do these really nice little development boards, like 17 quid for a board that's basically got just what you need. It's got the FPGA, voltage regs, JTAG programmer, and all the pins broken out onto headers. It's cheap, which is really nice.

Dave Jones: And how is this software?

Mike Harrison: Pretty good. I mean, I've used Xilinx in the past, and the software was a 3-gig download, most of which you can't actually use because it's for part libraries and parts that it doesn't even cover. Yeah. But the Lattice Diamond seems to be pretty reasonable. It will do compiling a minimal design in 10 seconds, which is quite reasonable. Every time I've looked at Altera, they haven't really had anything at the really cheap low end, but the new Max 10s look quite interesting as and when they appear. Because, again, they're a bit of a copy. I think they're trying to take some of the market from the Lattice XOs because they've got the onboard config memory and the onboard voltage regulators. Yeah, that's right. That's a huge advantage. Parts in QFP as well. But, again, no small pin count QFPs. No. No, they're hopeless. It's really annoying.

Dave Jones: But as long as they're in a QFP package, that's…

Mike Harrison: My primary issue is two-layer boards. That's why I want those packages. Yes. Yeah, totally. Because I want to do stuff on a two-layer board because it's a 400mm long string of LCDs or something bizarre that I really can't go to four layers on.

Dave Jones: And something that you could possibly hand solder if you need to. So you can just do a drag solder, you know?

Mike Harrison: Yeah. And, you know, my PCB guy up the road only does two layers. So that's another reason. Right. Yep. Exactly. Like a controller, I can actually give a very specific answer to this. Pick 32MX170F256B. Also, why? A number of reasons. People keep arguing about architecture. The architecture just doesn't matter. It's almost all about the other stuff. What I like about it, it comes in DIP, SSOP and QFN and they're all available easily. The latest one in particular, the most recent version of that family was like a 130, 150, etc. That's been around for quite a long time. But recently they brought out the 170, which gives you 256K of flash and 64K of RAM, which is very unusual on a chip that small. And because I'm doing stuff with video, RAM is almost always the first thing I run out of.

Dave Jones: RAM is the bottleneck, right?

Mike Harrison: Yep. Yep. But also I can buy them pre-programmed from microchip for like another 30p. So when we did those, that project with the nano LCDs.

Dave Jones: So that's actually easy to do, just order and pre-program.

Mike Harrison: Yeah, literally, you upload the file. And who do you?

Dave Jones: Actually, Mouser and DigiKey offer that service, don't they?

Mike Harrison: Yeah, I suspect it's an awful lot more expensive, though. Oh, okay. I mean, the microchip, the small parts, like the little SO23s, it's about 2p to get that programmed and marked.

Dave Jones: Oh, that's crazy. Why wouldn't you? Exactly. Is there a minimum volume with that? 150.

Mike Harrison: Oh, is that all? Okay, right. And it adds about two days to the order time, and it's about two days setup time. So basically, the way it normally works is you send them a file, you then order a sample, they'll send you five devices, but you can actually bypass that if you're in your own house. Don't bother with the approval order, I trust you. And they'll put a paint dot or a marker, like a label on it. But for things like the iPod Nano thing we did with like 2,800 QFN devices, even programming those on board would have been a pain in the arse. So to be able to get them, even if all that's in there is a bootloader, that's all you need, because you can then bulk program them in production. But say other reasons I like PIX, dev tools are all the same from literally from the 8-bit up to the 32-bit. I'm using the same environment. I'm using the same programmers for production. If I'm doing, say, a production job where the subcontractor is programming them, I can give them a 30-quid pick-it-3 with the code preloaded. So literally, they connect it, press a button, and get a green light, and it's in. And it's got, yeah, it's quite a nice, I've used it on quite a few jobs. But I'm sure, yeah, there probably are, you know, are cortexes that do a similar thing. But, you know, familiarity, I think the answer is, what's your favorite processor? It's the one you know.

Dave Jones: Exactly, exactly.

Mike Harrison: It's got some nice stuff. It's got DMA. It's got some reasonable peripherals.

Dave Jones: Yeah.

Mike Harrison: The UART, a nice feature about the UART is that it will do board rates up to a quarter of the clock rate. So you can do, like, 12-megaboard UART on it. Yeah, yeah. They slightly screwed up the FIFO implementation, but it's workaround, which is not a huge deal. Right. Basically, you've got this nice, you know, FIFO UART, which you need if you're running it 12-megaboard. Yeah. So you get an interrupt, like, and it'll, you know, you pull eight bytes out of the UART. The problem is, what it hasn't got is a timeout interrupt. So that if you only get three bytes come in and then nothing else, you never get that interrupt until some more bytes come in. Which is, yeah, firstly, it's, you know, for things like sport, that's less of a problem. What is usually the problem is getting the last few bytes out of a big transfer. So you just have to set up a timer interrupt. So you just, like, set up a timer interrupt a few microseconds after the last bytes. Yeah. If there's anything still in the FIFO, just call it. So it's get-around-able. It's just irritating. But other than that, it's, you know, there's no particularly nasty errata on it. And it's, you know, I've used it in quite a lot of chip. But also, as well as being the thing you know, it's the thing that you've written all the bootloaders for. And it's the stuff that's a real fiddly thing to do. And half the battle on that is not so much figuring out how to actually do the code updating. It's how to fight the linker to get it to put your bits of code in the plate as you want it to get on. And that's...

Dave Jones: So that's not necessarily a recommendation from you. Use this chip. It's just the one that you know. It's the one I know.

Mike Harrison: You've done all the work in. Yeah. There are a lot of things that suit me really well. And, you know, there's no... Yeah. Okay. There's other chips that do different... Yeah. It's horses and courses. But that's been... Yeah. It now tends to be my default choice. Unless there's a reason to use anything else. Yeah. I've done a few odd things.

Dave Jones: Does it have the... Does it have the peripheral pin swapping?

Mike Harrison: Sort of. It's not as good. And the Pic24s were brilliant in almost any peripheral. Yeah. It was great. Here it's a subset. You've got... Each peripheral has a subset of pins. So... Ah. You actually have to be... Right. You sit down really carefully and work it all out. Yeah. Most of it you can make it work, but it's not as flexible. You know, it's got so many... Yeah. It's almost impossible to not have that on a micro these days, especially in 28P. You've got so many peripherals. You just don't have enough pins for all of them. Yeah. No, no, no.

Dave Jones: That's right.

Mike Harrison: I've yet to fail to map something to all the pins. But in a couple of cases, it's actually been quite close and been quite... Right. Especially as the way I like to work in, you know, throw all the footprints and what... Yeah. Generally, what happens is I end up with a picture... The, you know, the pin-out pin of the data sheet, handwritten, all the... All the... Yeah. I know, say, I need two EARTS on an SPI and whatever. So I just write on each pin which pins I can use for those functions. Yes. And then cross out the ones that clack and then work out with the few permutations that I've got left. So, yeah. Yeah, that's right. It's not as good as the PIC24, but it's a lot better than some others I've used. So it's a neutral on that. Yep.

Dave Jones: For those who don't know what we're talking about, the PIC24 F-Series, they have this feature. It's almost like an FPGA pin mapping feature where you can just choose your UART, like you have your UART block, and then you can route it to almost any pin on the chip you like. It's not a fixed pin. So there are some limits to that, but basically, yeah, that's what you can do.

Mike Harrison: Yeah, but on the PIC24, actually very few, almost any pin will go to almost any peripheral. And it's quite handy, like, you know, for example, one of the ways I like to drive some of these LED drivers is that I have a PW and I'm actually clocking from the SPI port. So you can do stuff like share a pin and do weird stuff. But also you can change it dynamically as well, which is... Yes. I have an example where I did that on the Heathrow Airport job. I basically needed four SPI ports, but I didn't necessarily need them all at the same time. So I just did some pin juggling to sort of spit the SPI out of two pins to feed one half of the ARC and then swap the SPI, remap the SPI to the other pins, and then spit the packet to drive the other LED out of the other one. So, yeah, there is some scope for some further juggling and playing around with it. So, yeah, it's quite a nice chip.

Dave Jones: Have you had any problems with the microchip tools? A lot of people complain about the new versions of the tools.

Mike Harrison: Yeah, I've not even looked at MPLAB X. I've tweaked the linker in MPLAB config files to deal with the extra RAM. But I've not even started... Yeah, MPLAB 8.whatever, it works. Ain't broke, don't fix it. Okay, so you haven't changed. You haven't moved on. No, no.

Dave Jones: Right, you haven't moved on from that. Because that's like, what, four years old or something? Yeah, probably.

Mike Harrison: But it does the job.

Dave Jones: Yeah, right, so you're still using that, yeah. And what C compiler?

Mike Harrison: The microchip one.

Dave Jones: Oh, right, okay.

Mike Harrison: I mean, it's the... I've actually bought the upgrade for the paid-for one. Because I'm just... Got it. On the later ones, though, the early ones, it's actually... It's so trivial to hack that it's obvious they didn't really care about it. Yeah, that's right. And the later ones are slightly better protected. But because I'm a microchip registered consultant, I get ridiculous discounts on dev tools. I can't remember what it is. But it's, you know, a very significant discount on the dev tool. Got it. Nice. Any more viewer questions, quickly? Has Dave ever admitted to being wrong?

Dave Jones: I always... I'm sure. I always admit to be wrong. Yeah. Absolutely. I'm one of the few people who do it. I mean, a lot of other people out there will not admit that they're wrong. I do it all the time. Jeez, I've done it maybe hundreds of times since I started the video blog.

Mike Harrison: It doesn't sound like it, but... If you had six months of completely free time and £50,000 of available, what would you do or make? I have no idea.

Dave Jones: Me? Me either? Jeez, that's a tough question, yeah.

Mike Harrison: I can't find it. Someone said there's some issue about, you know, what are the limitations. But almost all my primary limitations in terms of gear and so on is space. Right.

Dave Jones: Have you thought about moving to a bigger space?

Mike Harrison: Not really. It's just... It would be such a traumatic thing. It's, you know... Yeah, yeah. The place I've got is adequate. You know, anything isn't used a lot either gets moved out. Some, you know, for example, when I got the pick in place, the PCB tanks got moved out into the conservatory because I don't use it very much. But I still use it occasionally. So, you know, generally... It's like in the last few days is a really major job. The place is just bedlam and stuff everywhere. I'm quite good at clearing, you know, fairly ruthlessly clearing out and retrieving space. So, yeah.

Dave Jones: Have you thought about renting or buying? Like a big warehouse space nearby? No, not really.

Mike Harrison: So, again, it's, you know, quite a lot of cost, business rates, all that sort of stuff. And, you know, this place is big enough and, you know, it's 30 seconds walk down the road and, you know...

Dave Jones: Got it.

Mike Harrison: Yeah, storage. Yeah, storage. I've got most of the house for storage. My whole back room is full of shelves and that's just offline storage. So, here's a good one. How big is the pick-and-place machine that did the 1.2-metre PCBs? Well, a bit bigger than 1.2 metres. Yeah, exactly. This company up the road, they do a lot of lead stuff and they can actually do boards up to 1.6 metres.

Dave Jones: Well, that's in length, right? Yeah. It's not like you can put a 1.6-metre wide board in.

Mike Harrison: Oh, no, no, no. That would be good. No, right. People have to understand that. No, it's only that far. Basically, the way it's designed, it's like, if you think of a standard pick-and-place, make it wider. But it's got two heads in it so that they can run in almost two independent machines for small boards. And then they can put one big, long board and then both the heads are placing onto this single board. So, yeah, that works quite nice. They're quite expensive for small-sized boards, so I tend to only use them for the really big ones. But it's quite handy to have that capability. Got it. Again, we use that on the Ether Airport job. I think they were one-metre boards, actually. Yep. But we stuck two of them together, so the whole light fixture was, like, two metres long. Right. Any more questions? I think that's about it for the sensible questions. So, the ones that haven't already been answered all, I'll ignore the one about UFOs.

Dave Jones: Right. Okay. Fair enough. Let's not go there. Jeez. All right. Well, thank you very much, Mike. Thank you. Thank you. Thank you for coming on and filling in for Chris, who's in Electronica at the moment. Yeah. Excellent. Well, we'll get... You're certainly invited back anytime. You might be able to go to, Guy.

Mike Harrison: I'm sure I've got at least another show's worth of stories from my past. Right. Okay.

Dave Jones: Excellent. Well, we'll have to get you back on.

Mike Harrison: Yeah.

Dave Jones: I'm not sure if Chris is back next week. I'm not sure what the... I don't know. I think he may not be. Anyway, I'll figure it out. Yeah. All right. Thanks, Mike. It's been awesome. All right. Cheers. Thanks, mate. See you. Cheers.