#266 – An Interview with Ronald Sousa of Hash Define Electronics

Chris Gammell: Hey guys, Chris here. I'm going to be in Portland this coming Friday, which is September 11th, and we're going to be having a meetup for electronics people from 4 to 6 p.m. at the Lucky Lab Brew Pub on Hawthorne. So I'm really hoping people that live in Portland come out. I'm going to be there for XOXO Conference, but I also know Maker Faire Portland is that weekend, and would love to see you there, would love to meet you, would love to see what you're working on. Please bring your projects with you. So if you can make it, if you're a Portlander, or if you're in town this weekend, once again, it's Lucky Lab Brew Pub from 4 to 6 on Friday. Hope to see you there. Now, on with the show.

Ronald Sousa Of Hash Defi: This is the F-R Podcast, recorded September 8th, 2015. Episode 266. An interview with Ronald of Hash Defined Electronics.

Chris Gammell: Welcome to the Amp Hour. I'm Chris Gammell of Contextual Electronics. I'm Ronald, instructor of Contextual Electronics and Hashifunetronics. Welcome, Ron. Hello, Chris. People might be wondering, first off, Dave is on vacation. And second off, I had Ron here on as a... Yes, he is an instructor of Contextual Electronics. He's doing some cool stuff with Embedded, but actually, I wanted to talk about his background. And he's done some really cool stuff in industry. And so we're going to talk about his background in Embedded and some of the stuff he's worked on. So welcome. Why don't you tell us a little bit about yourself?

London: All right. Well, thanks for having me, by the way. I know it's kind of like a last minute

Chris Gammell: thing. Yes, my scheduling abilities are notoriously last minute, but flexibility is always rewarded.

London: I've kept track. I've kept track. Yeah, well, I live in the UK. I'm a contractor in electronics. Come from a fuel industry. That's about it, really. So I do a lot of embedded electronics. And, you know, ended up listening to The Empire. Big fan.

Chris Gammell: Oh, well, thanks, man. So the UK, like you and I were talking about before the show, we don't have too many guests in the UK. And I think I've mentioned this in the show before, but one of the reasons actually is, you know, just Europe in general, that scheduling with time zones is actually really difficult. And it's not a great excuse on our part. But, you know, with Sydney and then the US and then UK, it's all split apart by, you know,

London: weird hours. Well, considering that we set the time zone, we should have the advantage, really.

Chris Gammell: You'd think so. But then, you know, sleeping and families and everything else, that's the sticker of it all.

London: No. Yeah. So it's kind of nice to have to be invited to the show, really. It'd be nice to talk about more about the UK and engineering and this other things, really. So,

Chris Gammell: yeah. So, well, let's talk about that a little bit. I mean, so you went to school there, but we don't have many people there. And, you know, it's obvious that there is, you know, electronics industry, stuff like that. So what has been your experience in the UK, you know, throughout your career?

London: The electronics are amazing.

Chris Gammell: Okay.

London: Yeah. I mean, to be honest, though, when it comes to engineering in the UK, there's plenty of stuff you can do, a wide range of stuff. I mean, my background comes from fuel electronics. So fuel delivery trucks and all that sort of stuff. I mean, I can't really talk much about consumer electronics that not really die, they've gone into that sort of things. But it's quite an interesting place, really. You know, sort of to hear Mike doing contract work down in London. And I, oh, my gripe with London, I tell you. If you're trying to do anything, you warned me not to talk about that, did you? I'll tell you what, though. When it comes to doing contract work in the UK, if you don't live in Cambridge or in London, you're pretty much having to do a lot of commute. Yeah. And I'm so jealous of Mike for his location, really.

Chris Gammell: Right. Well, no, and that is a good point. I mean, obviously, there's a lot of overhead in, you know, the bigger city areas, but that's true anywhere, right? I mean, like, I'm in the Bay Area right now. And, you know, there's tons of overhead here as well, but people still make that stuff work. I think it's more interesting just about the, you know, just kind of the industries that are even there. So you mentioned Cambridge, that's where Arm is. And there's much other stuff.

London: It's our UK Silicon Valley, if that makes sense.

Chris Gammell: Okay. Yeah. Yeah. And then what are some of the other industries that you've seen around,

London: around even London, everywhere else? Well, we tend to be a lot of digital industries, really. So web development and stuff like that down in London, a lot of hardware, hardware-less companies. Oh, okay. So it's kind of interesting, though, because you would have thought that the UK being kind of like the hub of where software kind of became a thing. It's kind of become what London is all about, really. So marketing and chipless companies, you've got that in Cambridge. And here up in North, which is where I'm based, Leeds, we tend to be a lot more about manufacturing. So it's kind of...

Chris Gammell: Is Leeds like the Cleveland of the UK? Is that what I'm hearing? Well, well, you know, I wouldn't say... I wouldn't go as far as saying that. I don't think anyone would say that, yeah, ever.

London: Yeah, we do have cheaper houses. But... Nah, no. To be honest, though, I have to say one thing I love showing up about, and that is that I live right next to Farnell, the main hub for Farnell. So if I need anything, I can just literally just walk over and just pick it up. And to be honest, though, I wouldn't be surprised if that's the reason why a lot of the electronic companies here and what appear in North.

Chris Gammell: Ah, interesting. Right, right, right. Yeah. So any people know, I mean, Farnell is part of the... What is Farnell, Element 14, Newark, that's all kind of the same thing. And like you said, Farnell is the... That's one of the big ones. So that means there must be a bunch of stuff around there to support that. I mean, that's... I guess they kind of serve UK and then a lot of Europe as well in terms of being able to get parts, right?

London: That's right. I mean, to be honest, I've never had an issue where if I needed something, I couldn't find it on Farnell. Right. So it's... I mean, there's lots of little companies here up in Leeds, mostly in Leeds, really, where they've basically focused on very niche markets. I mean, I don't know if you've heard of Cocoon. That's one in Leeds.

Chris Gammell: No. What is that?

London: It's a home sort of security system. So like a... Oh, okay. Like an automation? No, I would say more like a smart webcam, but better than that because they actually have their own sort of microphone built-in system that are going on where they actually try and monitor sub-sounds, as they say.

Chris Gammell: Oh, interesting.

London: They were on Kickstarter not long ago, to be honest. Okay. Yeah. So right now, it seems to be kind of like a Leeds thing to talk about. Like a lot of the sort of smaller companies tend to kind of be using them as a, oh, you know, we're going to be the next Cocoon kind of thing.

Chris Gammell: Oh, yeah. Right, right, right.

London: That's in Leeds. They haven't actually gone on... You know, they have not really gone and released any products yet. They're still... And yet they've been used as a comparison. So I'm not... By the way, I'm not trying to make Leeds sound like a small sort of middle of nowhere place. It's actually quite a lively city.

Chris Gammell: Yeah. Right. Makes sense. No, that's good. That's a good point, though. And there, I mean, there's, you know, and I just get like kind of just running through it. It's like, you know, I don't really necessarily think the UK for anything specifically, but then I, you know, if I step back, I think about, you know, I guess Raspberry Pi's out there. That's kind of more on the hobby side. But, you know... Yeah, that's like Cambridge. Cambridge, right? That's another Cambridge one. And that's, you know, and there's a bunch of... There's a lot of industrial stuff. My friend told me there's actually an old chip fab in the middle of the countryside out there. I was really surprised by that.

London: I'm guessing you're saying somewhere like in the Midlands, possibly.

Chris Gammell: I don't even know. Yeah, I didn't know where it was. It was like...

London: There's a lot of PC manufacturers in the Midlands side of things. Like from Manchester downwards, it's just a silly amount. It's like every other week I'm hearing one's closed down, but another one's opened up.

Chris Gammell: Interesting. Yeah, no, I guess Manchester used to be really big for manufacturing, right?

London: Yeah.

Chris Gammell: Yeah. Okay. Yeah, because I mean, like I said, you know, geography has always been my weak point. I don't really know much about... Yeah. So...

London: I can give you a breakdown if you want to of the UK, but that's going to be a very, very

Chris Gammell: quick... Well, no, no, no. Actually, that could be interesting. Like a breakdown of like different regions even, not necessarily cities, but like, you know, so like Leeds is known for what? Like, is it manufacturing heavy, like you said?

London: Well, sort of Leeds, Bradford, sort of that northern side of things, they tend to be quite heavy in manufacturing. Okay. Yeah. Where you've got, when you've got, if you sort of follow it down towards the Midlands, that tends to be kind of like the gaming industry for some reason. Okay. They tend to kind of focus down there. And then when you go further down that, you get sort of, you kind of get a mixture just before London. Again, kind of PC manufacturers, suppliers, so to speak. And then just suddenly you hit London and then just basically everybody, it's either digital or hardware-less kind of companies. Right. And it's...

Chris Gammell: Well, we've talked about that a little bit before too. Like why do hardware engineering companies always seem to live out in the suburbs, right? And it's because, you know, if you need a big lab space or something like that, if you need a manufacturing floor, you're probably not going to want to pay London-type, you know, real estate prices. Same thing out here, Bay Area. You don't want to... You're not going to put a huge manufacturing facility in downtown San Francisco, but, you know, it's just... That's just the reality of the situation. So you're going to find more hardware-type companies kind of out on the outskirts, that kind of thing. Maybe a little bit less now because you can get smaller shops that can be, you know, in an office or something like that. But still, the manufacturing, it feels like that stays away from the big cities.

London: Yeah. I mean, so with projects that I've worked in the past though with regards to embedded stuff, I always find myself that whenever I want to get something manufactured or if I want to sort of deal with a customer with regards to some sort of truck and that, you always end up having to kind of drive further down south, find yourself like a sort of like a sort of the niche, smaller fuel delivery companies. And they're the ones who are willing to try out the prototypes. And but then when you sort of move further up north, they all seem to be a bit wiser. They seem to be kind of, well, if you're going to try it out on our trucks, well, you're going to have to do a lot more than just sort of borrow a truck for a day kind of thing. It's like, whoa, okay.

Chris Gammell: Okay. Okay. So, so how did you, how did you end up in Leeds in the first place? I mean, like what, you went to school there or what?

London: Yeah, I came to Leeds University where that's where I studied and got my master's in. To be honest, it sort of when I, because I actually used to be based in, I used to live in Essex with my family and that. And I kind of thought, well, I want to kind of venture out in the world and sort of do my own kind of thing. How far can I sort of move away before, you know, where I can start making my own decisions? And I figured, well, let's, out of all places, miles, we'll just go to Leeds. How far is that even? I don't even know. It's about 200 miles. I didn't really go far, to be honest. When you compare the distances in there to America, sort of city to city, it's a stone throw away, really. Gotcha. Gotcha. But no, to be honest, I came to Leeds University because when I went to college down south, there was a teacher there, Bruce Kelly, amazing guy, amazing guy. He used to always tell the story about how he helped design the automated machines to, well, you know the Bank of England? Yep. The point where they were facing out hard labor to kind of check, you know, each individual notes to see if they'd been printed correctly. They were sort of going through a phase. He was the guy who designed, him and a few others, designed the machine that automate the whole process. So it's been, so it'd be done by machines rather than by people.

Chris Gammell: Oh, wow. So like machine vision type stuff or what?

London: Exactly. So that and sort of also the automation side of things, so the stuff that kind of throw away the notes when they're corrupted, so when they're not being printed correctly or it's like stuff like that. He used to always kind of boast about that and he kind of turned around and said, you know, so I kind of asked him like, what's your advice? It's, you know, if I want to carry on my path and kind of do the sort of stuff that you're doing, what can I do? And he says, well, first of all, find a good university. Then once you've done that, stick to it and work hard and say, okay, that makes sense. And I went ahead and asked him what university and said, well, you could go to Cambridge, but if you really want to learn what engineers really like, you need to go to Leeds University. And I kind of went, all right, I'm not going to question that. I'm just going to do it.

Chris Gammell: Okay. Just look at a brochure even.

London: Well, a brochure just kind of showed up one day and went, teach me, please.

Chris Gammell: All right. All right. That's good. That's good. And so, oh, so you went, I didn't realize you went to undergrad at a different school. I thought you went there at Essex as well, or sorry, at Leeds as well.

London: No, no. I, well, obviously I was born in Venezuela and then, so when we moved over to the UK, pretty much all my family ended up living in Essex. And so, yeah, there was quite a bit of, you know, life in Essex kind of thing before I came to Leeds.

Chris Gammell: Right, right, right. Yeah. Cool.

London: But my real engineering sort of background came, started when I, when I, as soon as I moved to the, to Leeds really. So. Cool. Projects and stuff. Like, to be honest, I tell you what, I've, I've mentioned this a few times to some friend of mine. But I tell you, if it wasn't for Farnell, if it wasn't because Farnell was based in Leeds, I would say that a lot of the people that I, that I went to unit with would have possibly not carried on with engineering. And I mean, sort of to, the reason why I'm saying that is because back when I first started, back when I first started university, Farnell used to kind of, what's the best way to kind of say, throw away electronics. But they were, they were kind of throw it away at the university. So they basically kind of get, get all their old new stuff. Yeah. Like their dump site.

Chris Gammell: Right.

London: But yeah, pretty much. But they would kind of, they would say kind of, as a charity sort of thing, kind of give it to the least university. And I, I remember that, the day that kind of stumbled across that massive box, I would say you can fit quite a few people in there. A lot. I remember kind of, I'm, I'm not drinking a full of junk. I'm talking about like some really good stuff. I mean.

Chris Gammell: Like just for like scrap parts, stuff like that. Like you could just.

London: Exactly. Like people will be buying parts and then what they would do is they just return it. And what the people, Farnell would do is well, either they'll try and get it sent to the manufacturer for them to check it. But if it's something quite low cost, they just wouldn't bother. They just say, here's a new one and we'll just keep this to one side. And I think it was a, it wasn't until a few years later that I realized that they had a trade counter at, well, in Leeds. And what they would do is all the return stuff would go there. The, the people in the trade counter would kind of go, well, we'll leave it here for a few weeks. If it doesn't sell, we'll just end up putting it into a box. And then after that box build up, they'll just ended up kind of giving it away to Leeds University. Nice. I mean.

Chris Gammell: So like what kind of stuff though? Like you just find real, you get like reels or these would be, these would be reels of components or what?

London: It was, or it was just a huge mix of stuff. I tell you. So you, you would, in some places you kind of find like a massive reel of 10k resistors, you know, 0.1% that somebody's returned and they, I don't know what, why that got returned. It was just too good to be kind of thrown out. So I take in there. Um, one of my favorite things, it was this massive, um, siren. Uh, it's one, it's one of those embedded ones, you know, the way you just put power and then you just toggle the dip switches and you get various different sounds and stuff. Yeah. Right, right, right. And the number of pranks that me and my other two friends used to do with that.

Chris Gammell: Yeah. I can imagine.

London: It became a mission for us, for example, to try and embed that on every project at university. And so I remember one time, I mean, I'm not kidding. Like on the first year we had a, a, um, a project, uh, which is basically it was a burglar alarm. You know, most people who kind of go into university either one way or another end up doing something very simple or very similar to that. And, uh, you know, at that point we were already experimenting with the Chinese. We knew what we kind of were doing. We transitioned. We understood what was going on and we thought, well, we could blink an led or, or, yeah, we could scare the shit out of the professor. This, this thing was, uh, was like, um, uh, I would say about three lessons worth of, uh, or three different lessons at any one time would send somebody out to try and figure out where that sound was coming from. And it was just, it just, it, cause the thing is, I should probably point out. So the burglar alarm was using ultrasounds to trigger. Um, so if anybody kind of went by after a certain distance or a certain speed, I think it was using Doppler shift. So it would have been a certain speed. Um, it would trigger the alarm. And so what we did is we, well, actually what my colleague Alma did, he kind of put it together and he's just going to put it to one side and we just, we'll be testing it and testing it. And, and there'll be other lessons going on at the same time and they'll be complaining about, Oh, you know, the sound and everything else. And I just, I remember one time, sort of me turning around and I just finished putting together a, um, a little board. It's just a light detector, you know, something very simple and just kind of driving, driving this siren against that. And I just remember putting it in one side and forgetting all about it. And then somebody came along and opened a box that I just kind of left it there with a battery. I mean, I, you know, just kind of straight forward and someone just opened it up and just ended up scaring this shit. Sorry, I shouldn't really swear. Yeah. That's okay. Yeah. It ended up scaring the hell out of, um, uh, what eventually became one of my friends. Um, not because of that reason, by the way, I doubt if she realized that was because of, you know, she was frightened by that, I doubt she would have been my friend, but it was, it was brilliant. But going back to the final box, I, I remember getting power supplies. I remember getting, um, I, I, I, I think it was a broken down multimeter, but I think it was just a battery that was dodgy or a connector. I can't remember what I did with it, but it was all working.

Chris Gammell: Yeah.

London: Uh, eventually. And, um,

Chris Gammell: That's amazing though. Like just like how that, how that kind of stuff, uh, can, you know, it's almost like when, you know, there's a secondary industry behind, uh, you know, like a recycling type industry, you know, that's, it's like, it's like your, your university experience was the, the result of that kind of thing. So that's great. That's really great.

London: I, I tell you what, I, I wish there was more companies who would do that rather than just throwing away the electronics. They just put it in a box and send it out to some, you know, soon to be engineers for them to play with because I, the amount of fun that I had with the stuff, I mean, even till today, I'm still going through some of the stuff that I collected that I, I, to be honest, I just, I just kept it. It's, it's a bad habit.

Chris Gammell: Right. Well, yeah, I think a lot of people can probably, uh, identify with the, uh, having the boxes of junk at home, you know, I'll use, I'll use it someday. Right. I mean,

London: like, yeah, I'm going to use this, uh, nine volt battery that's already leaked. Uh, but it has a funny connector, so I have to keep it. I can't just throw that away. Right. That would be a, a travesty. Exactly. Exactly. Until, uh, eventually my fiance kind of turns around and says, uh, well, you either throw that away or you're sleeping on the sofa, in which case you, uh, you end up throwing it away. Yeah. Right. Right. That's when you, that's when you finally learn. Yeah. That's hilarious. I, I, I can't remember how many different projects, uh, span from those, um, from those components. I mean, uh, I think, um, um, one of my favorite ones. So going back to the, um, uh, the light detector one, for the light detector one, um, yeah, that was using various parts from that box. I mean, there was a, uh, pick 12 that came from that box, uh, brand new. I mean, it, it, it, yeah, the pins were a little bit bent. I mean, that's, wasn't hard to kind of do any, anything with it. So just, yeah, it's just, there was stuff like that, LEDs, bi-direction LEDs, and pretty much were able to, to pretty much, like, even, even the strip board, even the, um, Vero board came from that box. And I, I, I, I have to say till, to, to this day, I probably wouldn't have been as, I probably wouldn't have had as much fun during university if it wasn't for the,

Chris Gammell: for the stuff that came from that box. Yeah. So how, uh, so how, what, was it the pick 12 or like what, what actually got you into the embedded side? And, and what, what were, were we, uh, what were your experiences getting into the embedded

London: in this side of the industry? Oh, right. Um, sorry. I should have probably kind of started with that earlier and just kind of going off a box, a final box related stuff. Um, to be honest though, when I was back in college, uh, I wanted to get into robotics. Um, I, it was just something I always wanted to do. And, and once you kind of get into it, you, you, you end up kind of making, you, you end up kind of having to do a trade-off between, um, either going full on digital or, or full on analog. Um, and, and to be honest though, I mean, I have met, um, I, well, I do have a few colleagues who've gone and done PhDs into, in robotics and they have actually kind of kept both sides. But, um, the reason why I kind of ended up being into embedded was really, was because I wanted to do robotics. And two, as soon as I came out of university, uh, the first job I had was just purely, uh, embedded electronics, um, which is a, a local company, uh, nearby, uh, in Leeds, uh, which basically they just design all the, uh, embedded electronics for a few delivery trucks. So it's not just electronics. They also did mechanical stuff, but the stuff that I dealt with was, uh, pure digital stuff. So, um, screen driving, sensor reading, um, relaying that back into a server, what, you know, that various sort of stuff. Um, but yeah, it was, it was basically because of the first job that kind of made me get into the whole embedded side of things, what really focused my skills down to embedded skills. Um, so it, it, yeah, it's really all came down to that really.

Chris Gammell: Okay. Yeah. So, so let's, let's talk a little bit about it. I mean, so, um, so you mentioned what, so fuel delivery, what, I didn't even know this was an industry before, you know, you and I started talking. What's massive. It's probably one of the oldest, to be honest. Right, right. So what, what is the industry and what are, what are some of the

London: challenges that you ran into with that? Okay. Um, so in the industry I worked on and still kind of on and off depending on contract, uh, basically it's all about, um, the electronics that you may need, uh, for various different types of fuel delivery. So for trucks, which is what I tend to, uh, focus more on automotive. Um, so, uh, if you imagine that you're going to transport fuel from one point to another, you're going to need some sort of electronics to, well, traditionally you're going to need some sort of electronics to actually meter that, that, um, that fuel to kind of see how much you dispense. Right. Well, actually, what I, what I would say is I'd usually think,

Chris Gammell: uh, fuel does not want electronics because electronics make sparks and sparks make fuel go boom.

London: Exactly. That's, that's, that's, that's, that's an ATIC side of things, which, uh, I'll, I'm happy to mention that in a minute, uh, which has been much of my headache in this industry really. Yeah. Um, well, you'd be, you'd be surprised. It's not actually the designs of the, of, to be honest, it's not usually the designs of electronics that causes an engineer a headache. It's usually the certifications. Right. That goes with that. That's, that's the headaches. And I remember when I was working on, um, when I was still doing full-time work for that company, um, we were just getting into ATEX. Sorry. I hope you don't let me kind of, uh, I'll come back with a whole more about the fuel. Okay. Actually, no, tell you what, let me actually do that first. So it's one of those things I've, I guess. So in the fuel industry, so yes, you need to, um, deliver the fuel, but quite often you want to deliver to smaller houses. So here in the UK, uh, you tend to have, um, sort of places like farms, uh, or small, uh, uh, states where they may not have access to gas.

Chris Gammell: So this is, uh, including like heating, this is heating oil as well as other types of, uh,

London: That's right. That's right. So gas oil, kerosene, diesel, gas oil and diesel, technically the same thing, but priced and taxed differently and all that. Yeah. Um, so for heating, but

Chris Gammell: also for, for, you know, farm equipment, stuff like that. Like you're just saying the, it just in general, getting the go-go juice to the customers who need it.

London: Yeah. The go-go juice. Right. Um, what is it? The call it? The, uh, it's, it's, it's, it's liquid gold. I think that they call it. So Texas tea, black gold. Exactly. You need to get that expensive gold to the customers in a way that guarantees that people aren't stealing. And that's basically the industry that I, that I was kind of going into. What can we design so that we know that the drivers aren't going to steal, which is, uh,

Chris Gammell: Oh, interesting. Okay. So that's, that's just a problem as well. So it's about, it's about certifying that. Uh, so one way would be like, like weights. Is that right? Or is that one way to, to handle how, you know, how you, you know, that a truck leaves, it's like one of those, uh, school math problems. A truck leaves station A with, you know, however

London: many kilograms and then, and then suddenly the driver is a, it's, it's a gram or two heavier than, uh, then, uh, what's the question at the end of that? Yeah. It's one of those problems. Yeah. Right. Um, to be honest though, when it comes to metering, uh, there's various different ways to do it. Uh, but surprisingly the one that's traditionally done so now is literally just, um, uh, slip gears or, um, it's essentially it's just, you've got some gears where the fluid is trying to kind of push through and rotate those gears as, uh, basically essentially costing a metering sort of, uh, sort of costing a, uh, an encoder to count how many cycles, uh, those, the, well, essentially those gears are doing.

Chris Gammell: So is that like, you mean like on the outflow? So like as fluid ticks past a certain point, it's, it's, it's saying, oh, well these gears have ticked five times and that equals, you know,

London: four X amount. Yeah. Yeah. Okay. And compensate for X temperature and, you know, uh, whatever flow rate you're on and so on and so on. Oh yeah. Cause I guess, yeah. If you like any kind of liquid

Chris Gammell: has different expansion characteristics, stuff like that, right? Yeah. Yeah. So temperature

London: compensation became quite a big thing about five years ago. I think, I don't know about the, about, about the USA, but here in the UK, you're, you're, you're pretty much anywhere you go, you're going to see temperature compensated. Uh, actually even, even on the fuel courts, you end up finding that the, when you kind of next thing, next time you come to put fuel in your car, have a look to see if you can find that 15 degree sticker that tells you that fuel has been compensated to 15 degrees. Um, so yeah, it's like, it's kind of became a big thing back in,

Chris Gammell: well, I'd say about five or so years ago. So what it's standardized on 15 degrees as the, as like the standard, like coldest kind of, well, that's not true. Cause it wouldn't be coldest.

London: It's just a standard. Well, what it'd be, it's like, they're trying to figure out how much energy they've, they've, they've, they've, they've basically, uh, charging you for. And so the way to do this, regardless of what temperature the fuel is, you bring it back to a known temperature or, or you bring it back to a kind of standard temperature. Um, and so it kind of like, if you imagine about, I mean, you'd be surprised how many fleet managers are a bit, uh, you know, quick for the trigger kind of thing. Um, you imagine if, if, if you're kind of delivering, uh, fuel when it's cold, by the time it gets to the customer, um, it may not have the same energy. It's going to be more dense, isn't it? So you're, you're actually delivering more fuel when it's cold, but when it starts warming up, you're actually, the fuel's going to expand and therefore you're going to be delivering for the same area. You're going to be delivering less energy.

Chris Gammell: I see. I see. So, okay. So if, if you put a, if you had a very tiny truck and you put a gallon into it, but the tank could hold two gallons, you're saying, and then it heated up that when you meter it back out, when you're pouring it back out into your, wherever it's being delivered, if you measure out and you say, Oh, well I'm going to give you a gallon, there could still

London: be some remaining in the tank. That's the idea. Exactly. And then you, you ended up finding that, uh, quite a few sort of smaller, uh, fleet companies or smaller fuel delivery companies would take advantage of that. They will basically try and see if they can, um, because when they introduced the whole temperature compensated thing, um, you were actually able to, well, the drivers were actually able to enable and disable that feature. Um, so it kind of became a bit of an annoyance really, because, uh, some less honest people would go and enable it during the winter time and disable it, um, during, um, the warmer times. So that, so that, that way they always gain, but because the, the, the theory was always that if, if everybody wasn't temperature compensated, then it should, in theory, average out across

Chris Gammell: the year. Right. Oh, I see. Right, right, right. Yeah. And if you're always for the same delivery

London: people, that kind of thing, that's interesting. Exactly. But then as they slowly started introducing the whole temperature compensated side of things, you ended up finding that some people are trying to, you know, you, you get customers kind of saying, is there any way we can disable that at will? It's like, yes, but that's not morally right. Uh, right. Yeah. But that said, there's not become a thing now. So national wet measures, they, which is how, I don't know what it's like in the USA, but in the UK, national wet measure basically turn around and said, it has to, if you're going to have it on, it stays on. If it's going to have, if you're

Chris Gammell: going to have it off, you have to have it off. Yeah. Right. That makes sense. So a little bit more about the, the actual logistics here though. So like when you say, I, when I think about tanker trucks and stuff like that, I think about the ones that I see around

London: the U S at least, like mostly rigid ones, aren't you? Yeah. The ones that are going to

Chris Gammell: gas stations that are like basically going to fill the tanks there. You're saying that these are, these are more home delivery, like smaller trucks that go. Yeah. Smaller trucks.

London: I mean, I, I do, I do also deal with the larger ones as well, but, uh, for the actual electronics, uh, most of the trunks that I designed for actually went for the smaller ones. Uh, but there was a reason for that actually, because, um, for the larger ones, like the ones you're used to, usually you don't tend to have power on the back other than the tail lights and indicator lights kind of thing. Um, and so usually if you try to put any electronics over there or in the back of those larger vehicles, you have to have some sort of battery method, um, to try and kind of get rid of it or trying to power your electronics. And so for a while, when we started this releasing products or when I still worked at the place, when we were releasing products, we kind of stuck them. We kind of ended up focusing more on the rigid or smaller ones, um, just to make sure that, uh, we got all the electronics working and it's kind of proven technology. And then eventually we will move on and kind of introduce battery power one. And so, so like the truck that you're kind of used to seeing those tend to kind of fill up completely and discharge completely. So they don't tend to, um, do smaller drops if that makes sense.

Chris Gammell: Yeah. I guess you wouldn't need metering so much as like just put everything, everything where you're going.

London: Yeah. I know this is the case in the USA because, um, so I, cause I, I do have some, um, dealings with America. So what they do over there is that actually the fuel is actually owned by the customer. So you're, the customer is not, so what, so let me kind of take it back a bit. So ordinary here in the UK, what tends to happen is that you pay somebody to deliver this much fuel for you. So you're actually buying the fuel of somebody and they will take care of the logistic to get that to you and dispenser and so on. Uh, in America, the way it is, is actually they already bought the fuel. The customers already own, owns that fuel and they're only paying you to go and pick it up and drop it off. And so what ends up happening is you end up getting like, um, kind of like either one guy who owns, happens to own a truck, he'll get, he'll get paid and do the fuel delivery kind of thing. Or you will end up paying kind of like a, kind of, kind of like a smaller or I'm not really sure how distributed it is in the, in America, but you end up paying like a fleet, a driver, uh, sort of like a company who, who has a fleet of people who basically own the trucks and they're ready kind of to do the delivery for you.

Chris Gammell: And then it's like a charge on top of the actual fuel cost, that kind of thing.

London: Yeah. And so the only, the issue they have, which is kind of the reason why we started going into that market is that they, what happens if they've gone and bought a full truck full of fuel and the driver's gone over there and went to deliver it, but it's a frustrated delivery. So, you know, there's like a thousand gallons, whatever it is that they weren't able to deliver. What's the driver did with a fuel now? Because in his point of view, he has to go back and pick up other fuel for other customers. So what is he going to do with that? And so we started kind of going for is, um, uh, we're basically taking the product we already, well, they already had here in the UK and kind of give him a way to kind of measure that and send that information back to that cup, to their original customer. So they know this is how much fuel you got left. And so this is how much you could potentially get off this fleet driver from, you know, next month or next week when you actually do have the space for it kind of thing.

Chris Gammell: That's interesting. It's almost like a, it's almost like, like liquid accounting, you know, like there's like entries and there's, you know, withdrawals and entries and that kind of stuff. But yeah. And I was just thinking about it too. I, you know, this, this does exist in the States. I mean, it's, it's, you know, my, my, my, uh, my wife's family is from the country and they do, I think it's propane heating, but sorry, I have to say propane.

London: What do you say? I'm trying to do the Kendall Hill thing. Oh, I don't, I don't watch that show. I'm sorry. That was bad. It was supposed to be Bobby. All I'm going to say is I'm sorry guys. That was really bad.

Chris Gammell: It's fine. Uh, I, I don't know what the delivery or what kind of actual gas it was, but yeah, I mean, I, this is, this is, I'm sure this is existing a lot of places where there isn't, uh, infrastructure for piping, for piping, you know, natural gas or other type of fuels around. So it's very interesting.

London: I mean, my, my sister lives in, uh, I think he's in somewhere in Kansas. Uh, there, yeah, they have a massive, uh, tank of their own, which once a week or once a month, however long, uh, they get somebody just to kind of top them, top them up. And yeah, that's yeah. That plenty of people live for that. And it doesn't seem to be an issue. It's, it's, I mean, it's nice to actually have, um, gas coming straight to your house because you're not going to be worried about running out of fuel during winter. Right. But, uh, to be honest, it's actually going to be cheaper for you to actually buy the fuel yourself and store it.

Chris Gammell: Right. Cause it kind of stuff like that. Yeah. Yeah. Yeah. That's interesting. Hmm. So, uh, let's talk a little bit more about the instrumentation itself. I mean, so, so you mentioned, so is it mostly flow metering or is it weight or how, how, I mean, you mentioned those gears as well. So what are, what are some of the methods for, you know, just, just even measuring liquid in general, right? Because that's kind of a broader term that I think a lot of people would be interested in for

London: kegerators, stuff like that. Um, well, there's quite a few different ones. I mean, I can tell you the ones that I had the experience with. So the, the standard one, which since it's kind of seems to be the sort of the norm here in the UK is, uh, the flow meter. And that's the one you seen, um, uh, I forget the name that I put it up. I'll send you a link to that later. Okay. Um, so essentially, essentially there's two sets of gears that kind of, um, are acting against the flow. So, uh, you know, the, the flow itself is kind of rotating

Chris Gammell: those gears and you've got that one, like a, like a, like a windmill or something like, or like a turbine almost where you're kind of spinning the turbine. That's, that's the other

London: method. So that's another method. So using a turbine. Okay. Yeah. There's a, there's a difference. So this one you literally have like, um, uh, it's, you know what, it's going to come back to me after the show. I'm pretty certain this is really bad. Um, so with it, where this one is basically, it's kind of like a flatter, um, set of gears, um, usually two and essentially you put, you're basically trying to put pressure against these gears to rotate. And, and basically, um, assuming that there's no slip, uh, you can measure pretty, pretty accurately how much fuel you pump through the turbine. It's more of a, so the, the, the

Chris Gammell: flow meter is more, I'm sorry, there's a lot of noise here. Um, the, the flow meter is more about kind of measuring or having a known, as you apply pressure to these gears, it creates motion, uh, like a certified motion type of thing almost. Yeah. Yeah. I mean, uh, where,

London: with the, with the turbine, uh, as most people are going to tend to picture with a turbine, you literally just put fuel through, uh, this turbine and starts rotating and the number of revs you get equates X amount of fuel. Right. And so with that one, there's a couple of issues you tend to find, uh, first is you tend to get slippage. So you get, sorry, you get, um, product kind of going around the blades themselves. So you need to kind of have like a very tight gap between the blade and the outer walls, um, to make sure that none of the fuel kind of escapes through that. Um, so because ideally, because the problem is if you've got fuel kind of going through the blades, or sorry, going through the, um, the turbine, if you've got quite a few of them kind of, quite a bit of liquid kind of going around blades or kind of escaping between the blade and the outer walls, you're not going to be able to measure that. And so that could potentially add up to X amount of error, you know, many liters worth that you're not really counting for accounting for. Um, so that one, that one tends to be a bit difficult to kind of, to deal with. And usually, uh, when you end up buying those sort of turbine meters, they tend to be quite expensive because there's a lot of good engineering that's gone into that to kind of guarantee that the workings of it and the life, the lifetime of it and you know, all that sort of stuff. Um, but once you get it right with those ones, uh, you tend to have, uh, very little errors later on. Um, so like with the first one I was mentioned, uh, mentioning to you, um, that one needs to be at least calibrated once a year. I mean, to be honest, all, if you're delivering any fuels here in the UK, all the meters have to be calibrated once a year anyway. So it's a requirement, but as far as

Chris Gammell: actually, that's like government, government mandated type thing. Yeah. That's national wet

London: measures, um, requirements that you have to calibrate it once a year and you have to have a certification for that. Uh, you have to have a, you have to have a seal to kind of prove that nobody can tamper it and so on and so on. Right. Um, but I've seen those,

Chris Gammell: I've seen those at a U S gas stations as well. Usually there's a local commission that does that same kind of thing where it's just like, it's certified. I always wondered, I always wondered if it was more of a, you know, uh, that's how they make part of their money type of thing as well, but it's also probably protecting consumers, that kind of thing. Right.

London: Yeah. I'm sure it's a bit of both. Um, to be honest though, with most certifications, people are going to want to make money. Um, I'm pretty certain that, for example, like with, um, if I, if I tell you that, like the third one, for example, of a way of metering, uh, using ultrasounds to actually, uh, measure the flow. Uh, in fact, Mike, uh, he did a video, a teardown video on a gas meter that uses ultrasounds. Um, uh, he put it up on YouTube that, uh, yeah, that sort of method is far more accurate as long as you can assume that you can remove all air bubbles in the liquid. Um, and when you think about it, that, that shouldn't really need any, that shouldn't have any wear and tear. So yeah, yeah, yeah.

Chris Gammell: And yeah, exactly. The versus the mechanical type of thing. Could you explain a little bit more though? So is it like you, well, I don't understand the actual measurement.

London: Oh, okay. Oh yeah. So, so with the ultrasound, so what you, you have, so if you imagine you've got yourself a pipe, um, and you've got, and you need two ultrasound sensors for this to kind of accurately measure the flow rate. And so what you do is you've got sensors lined up so that when one, uh, pings, it pings with the product in the direction of the products flowing, bounces off the walls and hits the other sensor. So the basically you're, when one sensor is pinging, the other one is reading, uh, the pulse that is sent. Okay. And so they're both, they're facing

Chris Gammell: one another. Is that the idea? It's what's like kind of like, like a crossing laser line type of

London: thing. No, they're, they're more, usually what tends to happen is that you tend to kind of have them mounted on one side of the pipes and the same side of the pipe, uh, bouncing off the wall from the other side of the pipe and kind of following in the pipe along. So usually you tend to have like, like, uh, kind of like reflectors to kind of have like a, like a two point contact on the actual, uh, pipes kind of directed to the other sensor. So you wouldn't normally kind of face them together. You tend to kind of try and the idea behind this is so that the actual, um, the actual ultrasound ping is following the path of the liquid at least for X amount of distance. Okay. So are you getting that? I think so. Like it's, it would be like a, like an,

Chris Gammell: almost like an angle, like you'd set up like, uh, so it's in the direction of the pipe, but it might be angled at like not directly along the pipe, uh, like at zero degrees, it might be like 20 degrees off and then you put a reflector and then you put another one 20 degrees. Exactly. That kind of thing.

London: The key thing with these sensors, Oh, with, with this, with the ultrasound sensor is that you, you actually want at least, uh, X distance where the ultrasound is traveling with the liquid. If that makes sense. Yeah, I think so. I think so. So the idea is that, um, if the product is flowing in one direction, you ping from one sensor and then you ping the other sensor going against the liquid and the difference between the two will give you the flow rate. Oh, okay. Okay. So it's like a, like a time of flight type of measurement or what? Exactly. So like with lasers, you do the same thing with this. So you measure the time, sorry, I should have said you, you measure the time, um, that the, the, the ping takes to get to the other sensor and, and by pinging both sensors, you know, with and against the flow, you can figure out the flow rate. Um, so with Mike's video, he did a pretty good job kind of tearing it apart and kind of showing you how, how that works for a gas meter, but that same principle works for fuel. So assuming there is no air bubbles, um, in the actual liquid, assuming you can pretty much get rid of that or at least minimize it to a point where it's not too much of a problem. You can very accurately measure, uh, the, the, the flow rate of the fuel. And now if you know the flow rate and you know what the density it is or what the product is, you can work out how much that's flowing or how much you've actually, uh, metered through. Yeah. Right. So let's talk a

Chris Gammell: little bit about the, uh, so it sounds like, first off, it sounds like these sensors. So like you buy like kind of like certified completed sensors, you, you buy kind of more raw sensors than you do the

London: calibration. All right. For the, is this for the meters in general? Yeah. Just in general. I mean,

Chris Gammell: like, is it cause I've, I've had situations before where like I've been using sensors where it's like you have to buy, you buy a fancy sensor with a, with a certified output or you buy a raw sensor and

London: then you make the certified output. Ah, right. I see where you're getting at. Yeah. So, um, like for this, for example, for the stuff that we did, so with the first one that I mentioned, that's the one with the gear stuff. Um, yeah, but you know, the company I work for, they just buy that, um, that, uh, basically all that, what is just an encoder, uh, a quadrature encoder. Uh, you just buy that to you and what you're paying for is the quality of the meters of quality of the gears and stuff. So the housing and stuff. Yeah. And so you know that that's going to be, that's going to have X amount of errors and so on. Right. And then you design your own, uh, electronics to basically read from that and do your calculations and you basically display that. Right. Um, but for the company I worked for, uh, they were buying that already made that, that side of things. So the, the official national wet mesh of certified of all product, they'll buy that in. And then what I was doing to sign in the extra electronics to basically anything that he might need. So, um, like for example, I was mentioning, uh, the air bubbles and the liquid. Yeah. So I would have designed the electronics to guarantee that we know when there is air bubble. And if there is then open X valve that would allow to, you know, for the antrine air to be removed and so on. Um, but that was just one aspect of it. Um, so yeah, you could actually buy those, uh, already, you know, those sensors, uh, sorry, the meter and the driving electronics and you got yourself pretty much a certifiable product, but then there is the second part that you need to deal with. And that is actually

Chris Gammell: getting it certified because, uh, right. Yep. Yep. Because you can't just go like, uh, you can't

London: just go there and just put it together and say, yeah, this is going to be fine. Uh, we, you know, we, we're happy that's going to work. You actually have to go back and first get your product approved by national aware measure and obviously meet their requirements. Right. And then once you've got that approved, then you, then you need to also either have approvals for you to certify your, your, your equipment or you pay somebody to calibrate and certify your equipment. And it's like, Ooh, okay. It's that there are just so many points of, uh, certification kind of thing.

Chris Gammell: Yeah. Yeah. I guess, I guess the thing I think about it, so like I, I used to do industrial and it would be like, you could either have someone who wanted to use, uh, you know, they could use. So, so in industrial, there's like four to 20 million type communication stuff, but then some people, yeah, some people would also do, they would actually want to use like plus minus 10 volt input and then have like a raw sensor kind of thing. And the four to 20 is almost always, you know, like a, like a hockey puck module thing that has them processing on it externally. And that's just sending back that current signal over the line. So that's kind of my experience with that kind of thing, where it's a lot of times these days, it felt like more people were just buying the hockey pucks. They were already certified to a certain, a certain spec. Oh yeah. And then, cause to be

London: honest though, it takes a lot to get stuff certified, to be honest. I mean, that, that same company, I mean, there's two products that I can mention that the, the same company did a separate product of together and it wasn't sold as a metering system. It was sold as a gauge system. And so there's one, what's the difference? So the difference is that one of them you can sell buy and the other one you use it for like stock, uh, for stock checking or, or, or to verify how

Chris Gammell: much you've got in a truck. Oh, okay. So meter is actually certified. Yeah. So the meter will be

London: the one that you use to sell to the customer. That's the one that you say, well, we've given you 10 liters. So therefore pay for us, pay us that kind of thing. Um, I should say a hundred liters cause that's the minimum. Most of these trucks tend to do, uh, or they're allowed to, uh, the other one, um, which is a tank, a tank gauge system. That was actually a sensor that sat inside the actual truck itself. And that one measured the product directly and told you how much you actually had inside the tank. Right. And so that one uses, uh, pressure sensors. Uh, so there's various ways you can do this. You can actually measure the actual static, uh, height of the, of the liquid. But the way, the way this one worked is that you had two pressure sensors, one, which is pretty much sat right at the bottom. And the other one was, which was precisely 200 millimeters apart from the other one, precisely, you know, with intolerance, obviously because you calibrate it out. Yep. And so what you, the point you made there about, um, you know, buying already made sensors or already made calibration sensors, it's a big difference because like for the stuff we were doing, we were trying to push the technology we were buying in. So we were having to actually recalibrate the sensors to try and get that optimal, um, uh, you know, reading because the stuff we were buying, basically the, the issue they were having at the time was that they wanted to have a low cost, low cost alternative to what, uh, the other industry, other people in the industry to do in like there's, um, companies in Canada, for example, which they're using, um, uh, well, ultrasounds to measure the product, uh, the height of the liquid and their system costs many, many thousands of pounds for, I think like 10, 20 grand. Um, so the objective of the company and work was, they put basically when it has, they wanted to have a system that would cost no more than a few hundred quid or a few hundred pounds or so. Yeah. And so having to calibrate your own sensor actually ended up being a need because we were buying in what seemed to be relatively cheap sensors, pressure sensors, uh, which were technically classes calibrated. Um, and so we'd taken that and we'll go ahead and, you know, I built a whole, uh, test rig and a whole, uh, pressure chamber and everything else and temperature compensation and all, and all that stuff. And, you know, it's a pain trying to kind of like calibrate every single point and trying to compensate for various different, uh, various different problems from the sensors. So things like, for example, you wouldn't, you wouldn't, you wouldn't really think about it, but a resistor, it will change its value, uh, based on temperature.

Chris Gammell: I used to think about these things all the time, man. Oh, right. Sorry. Yeah.

London: So yeah, little things like that. It's like, it, it made a sense. But the thing is that when, when you buy calibrated sensors, um, you really, really have to pay attention to where the manufacturers tell you what you, what you're going to get from it. Yeah. Because they'll tell you, like you can talk to a sales guy, tell you now, and this has bitten me a few times, you can talk to a sales guy and you, you will ask him directly and say, is this good for XXX? They'll go, yeah, yeah, you can do it. It's fine. It's fine. It's fine. It's fine. It's fine. It's fine.

Chris Gammell: It's fine. What's the big deal? Right. Exactly. Right.

London: You come back a week later and you go, well, hang on, I just looked at your data sheet and just checked again, but that doesn't make sense to what you're saying. Oh yeah. Well, we calibrated to X amount and to guarantee that all that we can sell enough sensors, you know, we always calibrated better than what they really are. So you can actually do what you're saying. It's like, no, no, you can't. Kind of, it, it, it, it, if you're having to sell the sensors and you haven't to kind of, you're going to, you're going to calibrate them at order magnitude better, aren't you? But you, you can't expect that to be a realistic value. Right. Right. Anyway, so yeah, so I can appreciate having to actually buy your ready color rate sensors. Right. And I tell you, if you can do it. Yeah. Right, right,

Chris Gammell: right. So what about on the, so, I mean, so on the output of these things, so you, you were obviously building embedded systems around these. I mean, in terms of, you know, outputs and, and handling these sensors, what, what were the, what did the systems look like?

London: All right. Um, well, the industry standard, um, well, the industry standard for automotive is CAN. Um, and so basically pretty much anything that we design would one way or another end

Chris Gammell: up outputting data by CAN. Right. And we were talking about that last week with Mike Osman. Oh yes, yes, yes, yes, you were. You were. Hacking in the, getting on the CAN bus, that

London: kind of thing. So. Yeah. Although interesting enough though, um, even though you could potentially, like, even though we worked in the automotive industry, sorry, even though I worked in that industry, you don't often tend to connect your CAN system onto, uh, the truck CAN system, for example, you would, you, you, you only do that if you had to, if you, if for some reason, some customer said, well, we already have a system, you know, stiffened data from now, so please connect it to that. So you're saying it's more about a, it's more about having a standard

Chris Gammell: than actually being hooked into the, the larger bus. Exactly. Because the thing is that the,

London: the, the, one of the nice things about CAN is, is how easy it is to create nodes. How easy it is to kind of split, uh, processes. So like for the smaller rigid trucks, you, you may have like five or six different compartments. Uh, and so you, you may have like one sensor per compartment and each one may spit X data. And, and so having a, a modularized, uh, having a, a CAN system that supports multiple nodes, um, it's, it's a must really, because you can do all sorts. I mean, you can have a sensor reading from the truck, then you can have a node, uh, relaying that information back to the driver and you can have another node, send that data back

Chris Gammell: into the, into the headquarters. And does that make it more extensible than two, because you might have one truck that has five compartments and one that has 10. Is that kind of the idea?

London: Exactly. Exactly. So that was actually one of the major driving system for the, for us to go into CAN because, um, because one of the alternative we had, uh, with one of our competitors, he was literally trying to shoehorn, uh, RS-232 to actually do that. And I know you can put transceivers and make the sound like it is. Yeah. That's a good standard. I like it. Yeah. So you can do, you can try and do that, but that's not what he was doing. He was, he was literally having to kind of have microcontrollers per, um, what's it called? Per hub. So you had X sensor sending RS-232 to a node and that node was reading from various sensors because he had like five different serial ports

Chris Gammell: and. Oh, yikes. So it was kind of like a hub and spoke versus having like a, like just a string.

London: They're all kind of on a string. Yeah. I mean, there were some advantages to that because, um, like with the ATEX side of things I was mentioning, uh, I think I mentioned this to you before. Um, so in, with, with, if you're going to work in this industry, you have to be aware that you can't just stick your electronics into these trucks and hope that everything's going to be fine because if there's a spark, someone's going to die. Right. You, you, you, you literally, the drivers are literally driving a, a time bomb. You know what I mean? Yeah. Right. Exactly. Yep. And like what we most standards, the ATEX standards, what you want to try and do is make sure that you've done everything right. So you're not the guy that's been blamed for what something's gone wrong. Right. And so, and basically what ATEX standards is doing is, it's kind of defines things like, um, what can you, what can you have? What can you, what sort of materials can you have in exposed in this environment? What sort of electronics can you put in there? Um, if you are going to put electronics in there, what sort of, uh, what sort of protections are you going to put in there? Is it going to be, are you just going to pot it and hope that that's going to be enough? Or are you going to have some sort of flame proof closure? And so basically the standard kind of controls exactly what you can do or what you can and cannot do with the electronics. It actually, to be honest, it goes as far as actually even the clothing the drivers can wear.

Chris Gammell: Really? Wow.

London: So it controls, it controls everything. It's not just the electronics. It controls, um, you know, what the, uh, employer gives to their employees, uh, what sort of, um, notice, what sort of billboards and labels and, uh, you know what I mean? You can have in the office to make sure that people are aware that that's going to be explosive. And I mean, that's, that's really,

Chris Gammell: I mean, yeah. And it's for a reason, right? I mean, that's just for safety and stuff. So,

London: well, that's the thing though. Enough people have to die before stuff like this happens. So it's, yeah, it's sad. Yeah. Yeah. Yeah. It shouldn't happen, but eventually we got a standard that actually helps prevent issues. Yeah. But the problem I was having, right, when I first started out with ATEX, uh, the company I worked with, they, they didn't have direct this before. They basically, they were saying, we're going to go into this. I mean, I should probably point out that if you're delivering a kerosene or gas oil, you know, something like that, because of the temperature, uh, temperature range that we tend to have in the UK, it technically doesn't actually fall within, uh, the ATEX, uh, round of things. It's more of a, if the customer wants it, they can request it and, and, you know, enforce it, but it wasn't a requirement. Uh, at least not until recently anyway. And so at the time we were able to design electronics. And as long as we can prove that the electronics were going to be powered off during an hazardous, uh, situation, then, you know, you can stick whatever you want to there. And so very quickly, I mean, it didn't last long, but the company went straight away to, okay, we've got to go to ATEX and just guarantee that nothing's going to happen and so on in case for some reason, the driver decided to turn the system on, you know, too many, do you want to remove those variables kind of thing? Right. Um, and so one of the interesting thing though, is that when you, when the company wanted to go into ATEX, they didn't just go, they didn't just went ahead and say, okay, let's just bring somebody in who's done it before. Tell you what, let's get one of our engineers and just firm in the deep end. Let's just give them the standard and hope for the best. And that's what happened to me. They just kind of, yeah, yeah. They just kind of like, it was kind of like just one morning say, okay, it's official. We're going to do this. We're going to force it. And from then on, I just got given a copy of a standard because, uh, my boss went off and, um, this is before I joined, they had a meeting with some company who was trying to kind of sell them their, their ATEX services. And they bought the ATEX standard at the time, I guess, like it was an EN number. Yeah. And then when we finally... I'm sure a lot of, a lot

Chris Gammell: of people could, could, uh, identify with this, you know, like having to do other types of ISO type standards, stuff like that, you know, just being handed a standard, being like, now you have to parse all of this and figure out what we need to do, that kind of thing.

London: Yeah. And, and to be honest, that, that really does tell you a lot about the person though. Can they handle it? Are they going to break, you know, are they going to make a break? Right. Exactly. And I, I'm not kidding. I was, I was working with this colleague of mine. Uh, he's probably gonna be listening to this. I don't care. Um, and I, that day, right. So he finally joined and everything else. And I said to him, look, I'm going to do the same thing that was done to me. Um, but I wasn't going to leave him in deep pain. I was just kind of trying to see what's going to happen. And I just gave him the stunner said, look, read it, memorize it, do your thing and do your electronics. Um, you know, if don't worry about it when we got to, when we get, we'll go through the certification process, we'll all be around, but this is down to you to get it right. Yeah. And he didn't take a while. Uh, in fact, there was, uh, let's just say that, uh, it, it, it could have, his career could have ended at that point, uh, at least with that company anyway. But, uh, but that said though, cause I've been friends with him for years and he knows what I'm like anyway. So he kind of eventually realized what I was trying to do there. It was more, look, you, you just joined, I mean, I should probably try out by fire basically like just trying to, yeah, I mean, I should point out that I managed to get him the job, uh, at the place that I was working on. I was working because we needed somebody quickly. And then, you know, he just finished off his PhD or he was just in the process of kind of doing something else altogether. And I kind of went, well, I know you don't have any skills in electronics. Don't worry about it. I'll help you out. You know, I'll, I'll mentor you. I'll teach you everything you need to know. And you know, you learn by experience. And then like the very first week I'd say here, he's just to get, you learn, right? You start, you're, you start learning through this. You memorize like, like if it was, you know, like if this was your only way to live kind of thing. And it was, you know what, I, I, I don't, I can't say that we should do this to people, but it's, it's amazing to see the reactions. But please, if you're going to do that, just bear in mind that you have to have some compassion and tell them eventually that actually you're just testing them. You don't expect them to actually do it, but you eventually handled it quite well anyway, but that's good. Well, I was just, uh, some more about the,

Chris Gammell: the interfacing with the systems. I was wondering more about, uh, so you mentioned Canbus, but then actual interface, like off of the sensors, was that mostly like parallel serial interface type stuff? Was it like I squared C or what, what, what did that do?

London: So, so the sensors that we were buying in, uh, were detailed. So they were SQS C. So you design your board to connect directly to it as well, as close as you can to the thing. Um, and you take the sensor because they were, they were being sort of calibrated. Right. And so you can take that information, you, you get your, your mind controller, which at the time I was using an ARM processor, um, the LPC, uh, 11 C14, I think it was, um, which was quite convenient because it did everything you needed. Like you can read from SQS C and you had a, uh, a canned peripheral in there. So you can send the data straight away kind of thing. And so basically the objective was at that point is to read your sensors, your two sensors, because I think I mentioned that there was two, uh, pressure sensors in there. You take your readings, uh, you do your next, um, calibration process, whatever it is that you've, you, you had to do to calibrate whether that was temperature conversation, which was at the time. Um, take your reading, you convert that to a known value. So, uh, in this, at the time I was doing like kilopascal kind of thing. Uh, and then I was basically taking that and then working out what the height of the product is, because the thing is that because you've got two pressure sensors in there, you know, the distance between the two, so you can take the pressure difference between the two of them and work out density. Right. And so you're saying

Chris Gammell: though that, that you want to send back that calibrated value to some other node then?

London: Is that the idea? Exactly. So once you got that information, you can put it through the can and some other node could be a display screen, for example, which, uh, another thing that I had to develop. Uh, and that would display the information. It'll tell the driver things like, oh, uh, you're carrying kerosene because we had density. You're carrying kerosene. Um, you've got X amount of product because we've measured the height and he's, and we've got density. So this is how much it got left. Okay. Um, but the thing I should point out with trucks like this though, you can't just go ahead and take height and take density and here you go, you've got yourself a volume because it's the inside of those trucks on, uh, aren't perfectly round. If that makes sense. They're, they tend to be kind of like made out of baffles, like curved walls. Right, right. So you can't, you can't assume some uniform shape basically. Yeah. It's not going to be linear or it's not going to be predictable enough anyway. So you have to have a second set of calibration that takes your pressure and convert that to volume. And so that's, so it's quite, it's, it's, it's quite an involved system, but the point is that because you're using set pressure sensors and you got yourself a really readily cheap interface can, you can expand it to do whatever you want. So you, you don't have to just put pressure sensors in there. Like, uh, the most recent product that, that was released was, uh, had color sensors. So that's where I basically took a color sample of the, of the, of the product. And with that, cause I mean, I should point out that there's multiple products that have the same density, but it's just their color. They're going to be different. And so it's hard to tell. Yeah. It's like, it's hard to tell if you're carrying diesel or gas or if they're quite close to each other, if their densities are really close, if not the same, I mean, they are technically the same product. Um, so like one in the UK is more like, it's a bit more yellow. The other one is a bit more brown and, and so on. So you can take that color sample and, and so what did you need to,

Chris Gammell: so you, yeah, that's actually really, that's a good idea. The, the color sensor idea. Uh, but what did you use that output? So you, now, you know, it's diesel versus being kerosene. What is, is that just for an information type thing? Like you're carrying just as a verification

London: you're carrying diesel. Well, the, well, the thing is though, is like, um, like going back to what I mentioned to you about the, uh, sorry, the calibration output. So the volume output, um, it, you actually be able to use that to tell the driver one, what you actually carry in because, uh, one of the things that they have to do is because one of the things that you don't want to do, because unfortunately the way the trucks have been designed is you could actually accidentally, uh, load on top of another product that probably isn't the same product you're trying to load on. Oh yeah. That'd be bad. Right. Cause that would

Chris Gammell: basically make it a big tank of uselessness, right? If it's mixed together. Yeah. And

London: believe it or not, there's a huge industry, um, around that, around what happens when you do actually mix two products and, uh, an industry, an industry where they actually go ahead and try and recover as much of the product as possible. Interesting. Um, by splitting chemical wise. I mean, I'm not really too familiar with it, but no, I mean, it's amazing that if

Chris Gammell: there's an industry for it, that means it probably happens a lot, right? You'd be surprised.

London: And it actually, it might not necessarily happen a lot. Um, one of the things that actually can happen is like in the actual depot themselves, the places where you're doing your fuel delivery, you tend to have like small, small amount of spillage anyway. So when you're connecting and disconnecting connectors or, or you have a little bit of waste here and there, what they tend to do is they tend to funnel all that to one storage tank. So, cause I mean, they have to, they can't just throw away the waste into some sort of landfill and hope for the best. Right. They tend to just collect it. So the industry not only does it deal with, um, when people had accidents, but they also deal with actually trying to recover that waste, um, trying to split that up, split that up and try and get as much of it as possible. Yeah. And so, yeah, it's quite interesting though, cause, um, I've seen the electronics that got in there. So it's like you, you actually have various different sensors running on that, trying to get that information. I said, what is in there? Is this purely, is there water in here, for example? Can we remove that straight away kind of thing? But, uh, so yeah, it's, it's not just purely for accidents, but, uh, although that's it, I do remember a colleague of mine, um, he, cause when you want to test these sensors, you know, these pressure sensors, uh, you've got like a hose, um, that you connect from the side of the truck and you can plug that back into the truck and into the compartment. And what he was doing, he was, uh, taking out from one tank in the same truck and moving into a different tank. But what he failed to realize is that one of the tanks, I think he did realize, but he just dosed off kind of thing. He failed to realize there was already X amount of products in the tank, you know, and that, you know, what he was feeling into. And so he thought, he figured, Oh, I got five minutes. I'll just bugger off, quickly do something, come back. He left the truck on and turn it. And when he came back, there was just fuel coming, being spilled from the top of the truck. And it, it, it, it, it, he wasn't, he, yeah, we didn't like, we didn't let, like, let go of that for a long time. Nice. Yeah. It wasn't too bad, by the way. It was kind of like a savior that they did manage to recover all that fuel anyway, but it was like, no, no, no, that's a good

Chris Gammell: point. There's, I mean, there's just a lot of money in that too. So that I would imagine that that helps, you know, also sell, sell these types of products because you can quickly say, Oh, well this will save you money because you're not going to, you know, obviously you're going to measure stuff, but also you're going to know what's in there. You're not going to ruin stuff, that kind of thing. Anytime you have a product that's going to, save people money, they're, they're usually pretty ready and willing to buy that kind

London: of thing. So. Yeah. Yeah. I mean, it's surprising how many little, little industries kind of spill off from that sort of stuff. So it's, yeah, it's quite, it's, it's good for people. It's good for people. I mean, to be honest though, I would hate to kind of see that sort of waste product that's been kind of being put to one side. It's bad enough that we throw away electronics all the time. It's just to sort of see that sort of crude oil that's kind of been mixed with all sorts and crap and don't let it go to waste. Right. Right. So any other, any

Chris Gammell: other big challenge? I mean, so obviously there's the sensor interfacing, there's the, there's the connecting nodes together, that kind of thing. Did you have like a, a master node for that kind of thing for like on a canvas? Is there a master node or how does that work?

London: No, no. So you can, if you want to, I mean, it depends how you design the code, but sort of like it's real, sort of like a, as this basics kind of view, as basic setup, you basically every sensor talks, you know, every sensor will have his own unique ID and basically the one with the lowest bit value for the ID wins. So if you've got multiple nodes set up, you've got, you know, sensor A, which has ID one and you've got sensor B, which has ID two, then the one with the lowest number, uh, well the, the one with the lowest recessive, I think it's recessive. It's been a, it's been a few months now, um, ends up winning the communication. So if you have two nodes trying to talk at the same time, then one ends up talking and then...

Chris Gammell: You know that we did something similar with RS-45 and there's a reason for that, isn't there? Where it's like, you're like XORing the, the... Exactly, exactly. The ID numbers or something?

London: That's, that's right. So what happens is that all the CAN devices, all the CAN nodes are in sync together. So they're all start talking, they're all basically, whenever they're going to talk, they wait for that, uh, start sync and then basically they start pumping out the bits together and the one with the, basically the lowest zeros or the most zeros. Basically when they, by the time they finish, uh, pumping out the, the ID, they end up reading it back and seeing, okay, is that my ID? It's not. Okay. Let, let whoever, whoever else want, uh, do their talking and then I'll try again in a minute or however long you, you will leave it to. Yeah. Yeah. But you'd be surprised though, um, if you don't design it correctly, how easy it is to one, um, overcrowd the, the, the bandwidth. So for example, if you've got one sensor who's sending data every millisecond or however quick you've got the CAN set up, if you don't give enough time for other, uh, nodes to talk, then you can quite easily, uh, stop or you can quite easily kind of, uh, essentially span the network enough so that you can't read from other sensors.

Chris Gammell: And it's like the ones, cause the lowest, the lowest sensor, the lowest node number is basically flooding the system.

London: Yeah. And it's always coming back straight away, especially when you've got, uh, especially when you've got similar architectures, uh, all around the truck using what, like if you imagine using the same processor and they're running the same code and they're virtually in time, together kind of thing, you're going to, when they're all kind of trying to talk and then they'll go away and then they'll try and talk again at the same time, again, that one sensor keeps coming back. So, um, one of the things that I did for the first, uh, product, sorry, for the first, uh, node system that I had kind of set up, I had a master and slave system. So you, it's completely down to software at this point. So what I had is like, they'll all be listening to one particular ID, like a global ID. Um, and when that gets sent out with whatever data I've sent out, they all kind of straight away look at that and send their information.

Chris Gammell: So it's like a broadcast packet almost.

London: Yeah, precisely. Like a broadcast. Yeah, it's precisely that. And so the way I had it set up at the time, like the first time I designed the hardware for this, um, I would have it like every 500 milliseconds, I would sample the network. And so any data that I want to read from all of them, they should be able to come back and send their information within 500 milliseconds before I do the broadcast again. Um, but the reason why I'm saying that you have to be careful is because when I came to the second, uh, version of the system, I wanted to go for more of just send your data once a second, we'll see how it goes. And so what would happen is that the way the code was set up is that it would try to send the data. If it can't send the data, it would just pull back and just leave it and we'll try again next time around. And so what would happen is if I'm, if they're all being set up to send data once a second, because I, I stupidly, uh, wrote the code so that it will, it will, it will be basically every second, precisely every second, it will be, there'll be an interrupt asking for the data to be sent. You basically had like one sense of talking all the time and the rest kind of going, I can't talk right now. We'll, we'll just leave it. Right. Exactly. Yeah. And the next time it was, uh, I mean, you can set it up so that the can would all, the hardware proof itself would automatically reset the message. But at the time I didn't build up with that, I just kind of wanted to kind of control every aspect of it and let me decide if I want to retry and let me see what the error is. And yeah, that very quickly.

Chris Gammell: Yeah. Does that, so does that, so two questions. One is, uh, first off, how do you actually set the node numbers on those kinds of things? Is it like a jumper or how do you do that?

London: You know what? That's actually one of the hardest thing to do when it comes to deciding when you have to assign an ID to something, it's, it's, it's quite a hard thing. So what I went with at the time was that, uh, basically you plug in one sensor at a time and you send it a master ID, which is the same for all of them. Um, like this, like for example, the same broadcast ID that I would have asked, I would have sent out to tell all of them to sample. They would have been, I would have used the same ID to tell it, here's your new ID. And so I know it kind of sounds a bit cumbersome, but basically you had a master broadcast ID that could override any of the settings from the other ones with whatever ID that you wanted to have. And so the idea is that when you come, when you go into production, you plug in one sensor, you power up, you set the ID, you disconnect that, you plug the next one and you, you know, you just do that. Oh, okay. So you just have to put them on

Chris Gammell: the network, talk to them and tell them what they are. That's the idea. Exactly. Now we can get away

London: with that by the way, because, um, usually what tends to happen in this industry is that you, you design, you, you, you, you put together your kit that you're going to send out to a truck to be put to get, to be, to be mounted into a truck, sorry, to be fitted to a truck, I mean. And so what you would do at this point is that, okay, let's leave them all set to the same ID because the customer, which is the people who put together the truck, they will have to put one by one anyway. And they'll just set the ID because they're having to connect the physical sensor rate. And so what I ended up doing, it's just a sort of like a second thing to us. It is that once you've been, once you've got an ID set now, from now on, you, you have to stop talking and you're, you have to enter your power out after you, you have to enter the power cycle to get back into the network. And so, right. Yep. So basically all I did at the time is save the end user or whoever it is that's setting it up, an extra step of disconnecting the sensor and plugging it back in. So basically the, the filters will go in there, put a probe in, connect into the network, save the ID, and then the probe will automatically disconnect, uh, but just basically went into power cycle anyway. And then you plug in the next one and just kind of follow the flow.

Chris Gammell: Right. Right. Right. Yeah. Cause you don't want to, you want to make sure that you also don't assign the same number to two, uh, two, uh, two devices in the network, that kind of thing.

London: Cause then be surprised how often that happened. Yeah. Right. Right. Cause the, cause you'd be surprised cause you have one, one guy coming along, plugging one sensor and you're disconnected and go, Oh, but did I set that correctly? And then I'll plug it back into the network going, no, wait, no, I think you leave it on. Oh, Oh no, don't, don't. Right. Right. So that's a tricky problem for sure.

Chris Gammell: For sure. And, uh, I know I've, I've heard some stuff about like wireless networks. They have an even harder time because with wireless, you don't have the opportunity to power stuff down and have like physical connections and say, okay, well this one's plugged in the network right now. I know I'm talking to this one when it's wireless. It's like, uh, well, if it wakes up and it's on the, and it's trying to communicate, then you got to kind of broker who's going to be who. And then you have to do some kind of weird, weird naming, you know, like just kind of picking a random, random ID or, or using a unique ID chip on board, that kind of thing.

London: That's right. I mean, I've seen some people use, uh, one wire devices to basically, because they got unique Mac IDs and they'll just use that. Um, incidentally the, the device that I mentioned, the LPC 11 C 14 or virtually all NXP, um, quarters and zeros, as far as I'm aware, anyway, I think the same for the M3 versions, the NXP have, um, dedicated Mac IDs in them. So there will be unique. So I, I, yeah, but that's, that's tricky too, because that's another thing

Chris Gammell: I've looked at in the past. And then how do you keep track of it? Because then your manufacturing process has to be so tight, tightly coupled with your, your code, right? Because now you know, oh, well I have this tray of chips and it's IDs, you know, six, seven, four, two, zero through six, seven, four, two, nine. And you know, now I have these chips and I have to assign them somehow. And it's like, how do you actually utilize that unique number and keep track of it? It's almost like a serial number system you have to do. And it's, cause you have to guarantee that they're unique or else, you know, if two, if two of them have the same code or something, then you're screwed.

London: With the ATEX, you have to track the product. You have to have serial numbers, unique serial numbers. So, you know, if something's gone wrong, which product that is failed on. So you have to have traceability. Gotcha. And so one of the things that I did, because it's actually one of the reasons why I chose the chip in the first place. Um, there were other reasons. I mean, I designed the network interface before, uh, I went with the LPC anyway. So it's like, it was, I was using, at the time I was using like a DSP chip from microchip. And then when we went, uh, with the second iteration, I ended up using that. But so yeah, what I ended up doing is, uh, every sensor that gets calibrated and goes through our system, we take that MICAD and we, um, pair that up to a serial number, like a more readable and more user friendly serial number, you know, like with a prefix, you know, PCB, something, something. Yeah.

Chris Gammell: Yep.

London: And then we were stored on a database. So I, I, I wrote the, I mean, for the products that I was doing for that company, I did every aspect. So I went from designing the electronics that you needed to actually doing the stuff that you needed for the production side of things. So I had the advantage that I could actually go and write the program, uh, that connects to the database. So I write a, write the program that sample the data from the sensors, check that they worked. If they did, then we'll log that against their Mac ID.

Chris Gammell: Uh, right, right. So that's part of your calibration. That's part of your traceability all the way back to the ATEX and stuff like that.

London: Exactly. Yeah. And now the issue is though that, yes, I've got this unique Mac ID, but now, now what do I do? Because the CAN IDs weren't, you know, they wouldn't, they don't easily pair up with the Mac ID, if that makes sense.

Chris Gammell: Right. Cause it might be, you might have 20 different sensors that all have their own unique ID and then 10 of them go on one system, 10 of them go on another system. Yeah. They might, the first 10 might have the ID zero through nine and the second ones might have the ID zero through nine as well. Right. Exactly. On the CAN ID, stuff like that.

London: I mean, we, we tried all sorts. I mean, at one point we were pre ID in them and then just literally writing the letter, the letters, so the numbers on the actual probes themselves. Yeah. So when they went out, uh, the customer can go ahead and say, okay, this is probe five. So therefore that probe is for the fifth compartment. Cause the thing is the sensors, right. We had to associate the compartments so that drivers know which sensor is actually in which compartment. So the ID system we went for in the end were to match the compartment numbers and so on. So we saved like the first 10, uh, CAN IDs for the first 10 compartments in our truck, for example.

Chris Gammell: Right. That works until truck manufacturers start naming their compartments, ABCD then.

London: Well, or Lucy or Ben.

Chris Gammell: Yeah, exactly. Right. Right. This is, this is the extra super plus.

London: It's just a, I, I, I'll come back to that in a second, but, uh, yeah. So, you know, to tell you what though, we, we tried doing that for a while. We tried, uh, putting ID numbers on the actual probes and send them out only to find that the customers went ahead and just put the probes wherever they wanted. So like we've got documentation here, which is completely traceable. It tells you exactly at this point that you should put pro five in compartment five. What else do you want us to do? So in the end, we just kind of went, okay, well, we'll, we'll pre ID them because we had to as part of the calibration because at that point I'm now, I was able to actually calibrate 10 sensors at a time kind of thing. So we'll send them out. But if you get this wrong, fine. There's a system you can follow that we're trying to rectify this. And you'd be surprised how often doesn't matter how, how many times we go out there and try our best to give, give the customers, uh, five probes with five unique IDs, they'll still go ahead and mix and match. And it was, oh, get the driver, get driver calls up and say, um, I'm filling up compartment five, but the display is showing

Chris Gammell: compartment one filling. Yeah. That doesn't surprise me at all. I mean, like that's, that's what it comes down to though. Like, you know, like unless you have a completely integrated, like, like tightly regulated system where you are doing the full, like the full, fully folded, vertically integrated, where you're doing the actual end user integration as well, then you're always going to have people use it however they want to. And so the best you can do is just try and notify like, like your system

London: did, you know, that's the best that you can really do. Yeah. I, I, I tell you one thing I knew I was in America, right. You know, that my product finally made it to America. When I went there, for, I think it was like a two week thing. And I sat on the side of the truck because usually when we fit the display, we fit it on the left side of the truck because it's just, you know, we drive on the right, on the correct side of the road. And so, which is supposed to be safer. And so when you get to America, the driver saying, well, we want the display on the right side because that's what's better for us. And I was like going there and going, oh, because I like, the way I laid out the display is that the compartments go from left to right. So compartment one is on left and compartment five is on the right. Yeah. Got to America. It's like, this

Chris Gammell: doesn't make sense anymore. Right. Exactly. And then do you change the firmware or do you change just how they interface with it? Do you have different versions, that kind of thing?

London: I changed the firmware. The thing is, um, working for a small company, you, you end up doing everything. So I'm there with my laptop on the side of a truck, you know, not far from the, from the highway. I guess I think that's what you call it. So in one of the service stations, they're just coding away going, I want to turn this the other way around because I never thought we needed to turn this. Oh no, this has to go right to left. And then made it back to the UK. And then I got told, can you make this a little easier for us to configure and go, what else do you want? Right. Well, what about drivers who may have to display on the back of the track and going, well, you know what? I don't care. They can code it themselves. Damn it. Yeah. If they think this is so easy, what don't they do it themselves? Right. I was like, and that's how you end up losing a job. That is correct. Yes. No, but to be honest though, I tell you that the, the most fun I've ever had in a field is always when you have to do panic work, when you ever have to kind of quickly code something or like, that's fun. That's fun for

Chris Gammell: you. Usually that, that I wouldn't, I wouldn't think many listeners would agree with that.

London: I, I, I don't know. I, I, I see you learn a lot for sure. Like no, no doubt about that, but like fun. Yeah. I, I tell this to people cause the thing is though, I love what I do. I love electronics. I love working in bed. I love getting up in the morning and, you know, making something work, but I find it fun or, uh, I guess you could say, I sort of get my thrills when there's an issue that only you supposed to fix and you get it fixed. And I want to, I tell you one of my favorite situations. So I've got two stories if you don't mind, cause I think we're, we're, we're quite into this. Um, so two stories I'll tell you. So it's one of my favorite ones. And my friend's going to kill me for telling this. So the first one is, this is the, the first year into, um, I just started the job, uh, working at that company, my first weekend and my friend comes upstairs and kind of goes, I'm having this issue with this board I designed a while ago. And what it is, is that this is an RFID tag system. And what system? Sorry. Uh, RFID tag. Oh, okay. Uh, contactless, I think it's called. Yeah.

Chris Gammell: Yeah. Yeah. So like the, the, the, however many megahertz, like the little coils that

London: communicate with one. Yeah. Yeah. So he's designed this system and basically his board drives a relay and the relay drives the signal, which tells the my controller, a second my controller to let go of the handbrakes for the truck and make it safe. It's supposed to kind of, uh, prevent people from, I should point out. So what happens is with trucks, when you're delivering, you actually have to leave the engine on to actually, um, uh, power the pump. Yeah. Exactly. And so quite, quite often it's been the case where either the driver forget to put the handbrake on and the truck rolls down the hill.

Chris Gammell: Like, like idling or something like it just idles forward. Yeah. Yeah. And it's happened. You'd be

London: surprised how often that has happened. Um, you know what it's like, it's something goes wrong often enough. Someone's going to make an industry out of it. Yep. Exactly. Um, and, or you could actually have somebody who just breaks into the truck, it just opens the door and just starts driving. Yeah. You know, just drives off and you know, that could happen. And so he designed this board that basically automatically apply the handbrake once the truck is parked up, even though it's, uh, idling. And so he won't let go of the handbrake until you hit the tag with it. So also act as a security system to prevent people from just driving off. But he was using a relay. Um, I don't think it was his choice. I think it was the RFID tag, whoever designed that aspect, uh, it was using a relay to drive the signal. And so what would happen is that if you knock it the right way, you can force the contact, uh, of the relay and cause, and causing a, a false signal. And so basically the issue was that he was reading from that signal directly. And I'm pretty certain a lot of your viewers, uh, would basically be saying, they'd be shouting, the bounce of signal, the bounce of signal. I'm certain they'll be saying that. And so it was like about a week or however long they were trying to figure out how to solve the problem. And at this point they were thinking about redesigning the board. And, you know, I only just started working there and they came upstairs and said, Oh no, you mind just coming down? Uh, I know this can be a bit much, but you mind just having a look? And I came downstairs and they described the issue and I said, depends on me. I, I kid you not both, both, both, uh, long time friend and long time boss friend as well. Uh, both kind of looked and went, Oh yeah. Okay. This is what I'm here for. Right. Yes. But it was just a software deep bounce. You had to do it. All you had to do is just measure the signal, treat it like a switch and then you're fine. And so that's what, that was one of my favorite one. So the panic one, uh, so like moment of pressure. And this is when I realized that I like being under pressure because you know, I ended up kind of, I like after the pressure's over, obviously that's the bit that you realize I've succeeded and so on. So we were out in the field, same guy and we're on the field and I left, I left one of the crucial boards that we, I went out to fit and all it was, it was just an optocoupler in there. Right. And basically converted 24 volts to something TTL. I think it was five volts at the time. Uh, possibly three, I think actually 3.3 volts. Anyway, it doesn't matter. And I left that board in Leeds and we drove, uh, over 200 miles to the customer to fit it. And, you know, I was panicking, Oh, what am I going to do? What are we going to do? What are we going to do? Well, guess what? There's a mappings nearby. And so I kind of went, okay, I think I can whip something together in no time at all. So I, I bear in mind, I didn't have a sword and iron. I didn't have anything. I didn't even have components. I just had the company credit card, which is quite dangerous. Okay. And so we're into the mappings and my friend was there going, okay, I can't find an optocoupler. I can't find an optocoupler. I can't find anything. And we're like going, dude, this only needs to work for about a couple of weeks and then the service guys can come back and actually refix everything. Anyway, it's going to, uh, remount new boards and all that. Anyway, so this is just a prototype. It goes, yeah, but I need to find the right optocoupler. And I just turned around and went, do we need an optocoupler? And said, all right, can we not just use a five volt regulator?

Chris Gammell: So that could take 24 volts. Right. Cause it'll be low, low enough current and stuff like that. Right.

London: Yeah. It's, it's, yeah. The only thing that we have to make sure is put a couple of extra passives in there just to make sure that it doesn't have any issues just being part, you know, being turned on and off randomly. Can we not just do that? And he goes, no, that's not going to work. Cause you know, it's 24 volts. It needs to be probably saving it. Then I was going, look, we can buy a sword and iron from over there. We can get the variable from over there. We can get this five volt regulator here. Two weeks time. We'll be fine. We'll be back to it and we'll replace it. Okay. I did that. It worked. It was great. It didn't get replaced until about two years later.

Chris Gammell: See, okay. So I think I know what this is here. I think you have MacGyver syndrome. I think that's what you know. Yeah. Yeah. That's what it is.

London: Yeah. Hey, it's a, it's a hero, you know.

Chris Gammell: Yeah. Right, right, right.

London: But I tell you that it's, um, it, it was just the fact that I was just standing there going, what can we do? Cause I don't want to drive back another 200 miles to get this sorted out. Yeah. And, uh, yeah, that was one of the, to be honest, there was plenty of other failures that I'm going to conveniently not mentioned.

Chris Gammell: Right. Of course. Of course. Well, that's good. Now, uh, so that, those were some great stories, man. And, uh, I appreciate you being on the show to talk about them a little bit. Uh, obviously, uh, you know, at the end, we should mention, you know, that you are a part of, uh, contextual electronics now, so that's cool. And I'm glad you're doing that. And, uh, you and I are talking about maybe doing a weekly, uh, video thing. So we'll have more details about that in the future. Uh, but weekly, weekly.

London: If anybody wants to listen to us.

Chris Gammell: Yeah. Well, we'd be doing video too. So.

London: Oh yeah.

Chris Gammell: Yeah. So, uh, yeah. Uh, thanks for being on and thanks for telling us your stories. It was, uh, it was, it was good stuff.

London: Thanks for having me.

Chris Gammell: Uh, where can people find you online?

London: Uh, you can mostly find me on Twitter. So my username is optical worm. Uh, I do have a company Twitter account, but to be honest, most of the fun is on my optical worm really. So that's the one to find me.

Chris Gammell: Cool. Okay, cool. Well, we'll have people fly there and, uh, yeah, we'll talk to you soon. Yeah. Bye.

Speaker ?: Bye. Thank you.