#426 – An Interview with Dean Pick

Dean Pick: This is The Amp Hour Podcast. Released January 20th, 2019. Episode 426. An interview with Dean Pick. Welcome to the Amp Hour. I'm Chris Gammell of Contextual Electronics. And I'm Dean Pick of Bi-Performance Development. Welcome, Dean. How are you doing? Doing well, thanks. Yourself? I'm good. I'm good. This is a nice follow-up to our robotics episode last week with Chris Osterwood. We're going to be talking a little bit about actuators, but more than that, too. Kind of motion. I mentioned mechatronics last week. So that's kind of what you do, it seems like. Your consulting work and the companies that you're working on, it seems like you're combining kind of electrical and mechanical. Could you tell us a little bit about that?

Chris Gammell: Definitely. So we are running a company called Kinetics Automation that is pioneering a new type of mechatronics actuator or a driver that uses shape memory alloy that basically acts in a straight line or linear motion but can generate pretty amazing force and precision. And there's a pretty detailed or rich history behind that that I can get into.

Dean Pick: Yeah, that's great. I mean, maybe even before that, could you tell us kind of the state of linear actuation in the first place? Like where the industry is right now and how most people do it?

Chris Gammell: There's a lot of classic technologies out there that a lot of people will be familiar with. You know, it started, I'll say, with hydraulics and pneumatics or air-driven parts that go back, you know, hundreds or even hundreds of years. And then with the advent of electronics and electronic drivers such as solenoids and AC and DC motors, you start to see sort of electrified direct driving. And more recently, we're seeing an emergence of things like voice coils and piezoelectric actuators, which really start to round out the sort of capabilities in terms of how much force or how much movement or travel you could get out of these different products. So they all kind of fall into their own special niche categories.

Dean Pick: And the voice coils and piezoelectric, how does that actually work then? Like is it, you're basically just sending electrical signal and it's driving a speaker, like a speaker type thing?

Chris Gammell: Yeah, the voice coils are exactly that. If you were to deconstruct a speaker and let's say take the cone off, you could connect that to basically anything, typically low, low force. But it's, you know, highly responsive, obviously, because it's got to react to the sound frequency range. So it's quite good at fast motion, precise motion, but low force. Whereas the piezos are very good at, you know, high forces up to a very, very small range. So fractions of a millimeter or thousandths of an inch.

Dean Pick: Okay. And that's different ranges for different applications. That's kind of the idea?

Chris Gammell: Well, the applications, yeah, it's a, it's a, you can look at it from one of two ways. If you're looking at an application with, you know, a specific set of requirements or specs, you're going to ultimately end up in a certain group of products that are going to be able to, to cater to that, that need. Whereas if you, you know, started from the actuator and built it forward, you'd end up, you know, inside its own capabilities without, and wouldn't really be able to go too far beyond that.

Dean Pick: And the other, another one I've thought about or not thought about, but I think I've seen before is it's like a, it's an actual motor that then just gets translated. The motion gets translated from rotational to linear. Does that have a name or is that just a general actuator? Or is that not even that?

Chris Gammell: Oh yeah, you're right. That's actually a whole class of actuators that I didn't touch on. Man, they are, arguably they're quite popular. So electric screw or electric ball screw are some pretty common names for those. And they're exactly, as you say, a rotary motor, usually DC or AC driven. And then there's usually some kind of a mechanism that, that will take that rotary action and turn it into a linear motion. So if you think of like turning a bolt with a motor, the, the nut that's riding along that bolt could be attached to anything and you would get a linear motion from that.

Dean Pick: Yeah. I kind of think about the steppers that you see on a lot of low cost 3d printers, you know, the rotational motion. It's got a lead screw or it's got a ball screw. And then that basically makes, makes the gantry move in whatever direction you want it to, to move. That's where I've seen a lot of it, but that's what I've seen in my house at least.

Chris Gammell: Oh yeah, totally. And they are, they're very cheap because they tend not to need closed loop control. You just count the steps and you tend to get pretty good precision out of those as long as you're not trying to move anything too, too heavy.

Dean Pick: And that, well, that kind of brings us to the too, too heavy kind of, it seems like the stuff that you're working on is, is somewhere in this sweet spot. And I was looking at a chart that you have on the, on the kinetics automation.com page. And it seems like you, you actually map this out of displacement and force. So what, first off, what is it? And then how does it actually hit into this sweet spot?

Chris Gammell: Right. So shape memory alloys are a metal, a metal alloy of at least two types of metal. And the ones that we've been working with primarily are based on nickel and titanium, roughly equal parts of both. And it was discovered several decades ago that with the right kind of heat treatment and let's say mechanical working, that this material would change shape when it undergoes temperature change. So similar to how, you know, water turns to ice, you know, a state change, but in this case it stays solid and the shape of it changes. So, you know, in our case, we, we've really focused on applying the wire product or form factor from this material, which basically goes from a long length to a short length when we apply electricity to it.

Dean Pick: And it's because you're dumping heat, basically you're just burning heat in like a coil or something. How are you actually using that electricity in this case?

Chris Gammell: Right. So for us, it's, it acts like a resistive element. You know, a good analogy would be like the, the resistor wire inside of a baseboard heater. So when you're connecting that wire across a voltage potential, its own resistance, you know, follows Ohm's law. Current flows and that current is turned into a heat inside the wire in a process called joule heating.

Dean Pick: That's interesting. I mean, and so that's seems like the, uh, does not seem very energy efficient. Is that fair to say?

Chris Gammell: You know, it really depends on the, the use case for it, uh, for, uh, things where you need to, let's say position something like an antenna or an airplane flap. Um, this is probably the most efficient technology to do that because once you bring it to a position, it takes, um, basically negligible amounts of energy to hold it. And then on the flip side of the coin, if you need to, let's say, move that flap up and up and down repeatedly and quickly, let's say a lot of cycling, then it has to go through this temperature change. And yeah, you start to lose some of those efficiencies.

Dean Pick: Yeah. And so it's almost, almost sounds like a hysteresis kind of effect where it's like you set it and it kind of stays in this and then it has, it requires an additional amount of energy in the other direction to make it go back. Is that, is that a good characterization of it?

Chris Gammell: Uh, there is definitely a hysteresis component to it for sure. Yeah. Uh, I like to think of it almost like, almost like a lattice, which is what the, um, the material creates. It's like a crystal lattice, uh, and you could, you know, play with like a mechatronics or, um, I forget the name of that childhood toy set, you know, with the wheels and the spindles, but, uh, connects or tinker toys. Tinker toys. Yeah, that's it. So, you know, by metering the, the power in or, uh, the electric, uh, amps in, you're essentially able to precisely control that, uh, tinker toy lattice. Uh, and you're, the, what we found is you can control it to a remarkable degree, uh, almost on a nanometer scale, which is kind of difficult to describe, you know, in terms of, uh, what you could compare that against.

Dean Pick: Yeah. And so can you maybe paint a picture too of, of like how the, the motion translates? Because I can imagine, I can imagine gating current at the input, right? So you put in 500 milliamps, when you get to 501 milliamps, it moves another, you know, you know, nanometer or whatever you said it was going to be the, the actual motion stuff. But like, how does the actual motion translate then though? So you said you add that extra milliamp, where does it actually go? Is it inflating like a balloon? Is it, uh, is it like, I don't know, heating like a sandwich?

Chris Gammell: Yeah. Right, right. So, so that's a great question. Um, maybe I'll start by saying what it definitely is not. And that's a pretty common misconception is, you know, people think of it as a solenoid or a motor where that current is related to the power output. So, you know, I put in, let's say an amp of power and I expect to get, uh, you know, one foot pounds of torque or a hundred, a hundred newtons of thrust, uh, shape memory alloys break that convention. Uh, so to answer your question, you put in, um, a milliamp or 500 milliamps of energy. And what that's doing is it's, uh, doing one of two things. It's either maintaining that material in its equilibrium. So holding it at its temperature, or some of that is actually going towards changing that, that lattice, uh, that crystal lattice, which takes some energy in itself. So if you're trying to hold something in place, you're basically just maintaining its, its, uh, heat loss to the environment. But if you're also trying to get some work out of it, then there's, it's a combination of those two things.

Dean Pick: Interesting. Okay. And so, and so it is more correlated to temperature you're saying than it is to the, the current that's actually put in there. Uh, that's correct. Yes. That's yeah. That's, I think some of the problem here is that I just don't, I mean, like I took a material science class in college, but that's, you know, I mean, sure as heck didn't go over this stuff. Um, you know, so, so then the controller that you, you make is basically a temperature controller at that point.

Chris Gammell: We, uh, we, uh, we, uh, we don't actually sense temperature. Um, it's really hard to do. Uh, and we, we certainly try connecting certain things like thermocouples, but, uh, we quickly burnt those up because, well, they conduct electricity. So, um, what we found was there's a, a very nice, uh, correlation between the length of the wire and its, uh, temperature. so we can actually infer its temp. Really what our controller does is act like an interface between a power source, a larger control system, and then our actuator products.

Dean Pick: That's great. You said there's a wire going through this shape alloy. Is it in... I still don't quite understand what it looks like though. Is it like a pile of silly putty that then expands or is it like a wedge that kind of slides across two different materials slide across one another? What does the actual internals look like if I

Chris Gammell: cracked one open? The wire, the shape memory alloy wire itself probably would look like something like a solid core wire that you'd use in electronics. If you stripped off the insulation, you would have something that just looks like bare, boring wire. If you held those two wires...

Dean Pick: I've already actually redone my mental picture here. Sorry about that. The wire itself is changing shape. I was thinking that this was a wire that was imparting heat into like a block of metal. But you're saying that the wire itself is actually the thing that's... So you heat up the wire and the wire does the work. Is that right?

Chris Gammell: That's exactly right.

Dean Pick: Oh. How can a wire actually impart that much force though?

Chris Gammell: Um, that's a great question. It's... So this wire is not like soft copper. It's, uh, got a really, really high proof strength and it's, uh, able to self-generate. Uh, and we've actually tested this. It's able to generate enough force to break itself, which is pretty amazing.

Dean Pick: That's awesome. Yeah. Okay. That sounds like a really good joke to play on the intern on the first day. Be like, yeah, I go, you know, dump three watts through this thing and then I just have it like bust in half and look at their face as they come back.

Chris Gammell: Oh, I, uh, sadly that's not a, uh, that's not a, just a, in the future that's happened before.

Dean Pick: Oh, okay. Uh, so, so like what is the, so is it like the wire is kind of in a J shape and then it expands out? Is that kind of how, like then, so, okay, so now that I, I've imagined this wire, I'm looking at my finger actually and watching it kind of like bend and is it like as the wire heats up, the finger kind of extends to a full length and that's what actually makes a linear motion happen?

Chris Gammell: So, um, not exactly if, so if you kind of held, look down at your hands and imagine you got a piece of wire, if you got something lying around, pick it up. So you got these, this wire between your two fingers and it's in a straight line. If you were to pass that wire over a candle, for example, it would pull your two fingers together. So the wire itself changes length.

Dean Pick: Oh, oh, oh, oh, oh, you're saying that it's, I'm holding the wire on either end with two fingers on each end. Yes, yeah. Okay. Oh, interesting. So it's actually, so the crystal alignment is actually in that, I'm going to call it the X direction, obviously it's very arbitrary here, but you're saying that the, so the, the, there's a normal length and then when you heat it up, it shrinks and that's what actually imparts the motion. And then when you go the other direction, it grows again. That's right. Huh, wow. Wow. So how do you get it to actually do it in that line? Like, uh, it almost sounds like a, like a, it's not really a tendon because a tendon actually pulls, right? Like there's actually motion that translates, but this is actually just like a, I don't know. It's like a, it's like a string that just shrinks up into itself. It's almost like a scrunchie or something like that.

Chris Gammell: Oh, you know, a tendon's not a, not a bad analogy for it. Um, yeah. And, uh, you know, to try to give it some, some context in the, in the electrical world, you know, we talked about putting that wire across a flame. Well, you could also hook up both ends to let's say a nine volt battery, uh, you know, at, at your two, uh, at your two thumb and fingertips. And that same effect would happen. It would draw your two hands together. That's crazy.

Dean Pick: So is it, so then how does it translate? So is it like the longer the piece of wire, the, the less strength it has, or like what are, what are some of the constraints that you actually have to play with then?

Chris Gammell: Yeah. For us, we, our first application, which led to the, you know, the creation of the company was we needed more force than one single wire could generate. So when we need, you know, a certain amount of force, we have to multiply the number of wires and they, they all work in parallel. Um, and then to get a certain amount of travel, it's all about how long the wire is. So generally speaking, those wires can work, you know, let's say, um, if it's a 10 centimeter or 10 inch length of wire, I'll go with a hundred. It'll, it's a little easier for the numbers. If it's a hundred millimeters or a hundred inches, um, it can shrink between three and eight millimeters or three and eight, uh, three to eight inches of its total length.

Dean Pick: Interesting. So does that mean that you, so you're not going to be a super compact, um, overall unit size because it's, you only get the three to 8% of the overall length, but you get the other benefits of holding strength and everything else.

Chris Gammell: Yeah. So, you know, in terms of its aspect ratio, it can have a very small cross section, uh, because the wires have so much, uh, strength to them. You, you can generate a lot of force with not much cross section, but in order to do something of, you know, any real value, you tend to have a long actuator.

Dean Pick: So what do, uh, I guess that ultimately plays into what industries are going to be interested in this kind of thing, right?

Chris Gammell: Right. Well, um, I guess the industry that we spun it out of was, uh, was an automotive application where we were controlling a, uh, a motorcycle clutch. Um, but when we spun it off into kinetics and started, you know, talking with different industries, we saw that there was all kinds of, uh, applications that we hadn't really thought about in, um, aerospace, in robotics, uh, for grippers, uh, for clamping or workpiece holding and a whole just variety of very interesting applications.

Dean Pick: Yeah. I would think the, the thing that I think about is like the holding, the, was it the holding force that you need? Is that, is that the right term?

Chris Gammell: Um, it's, yeah, it's, you know, when I talk about the precision really depends what your feedback is. So you could have really precise precisioning, but you could also have very precise holding, uh, in terms of how much, how much you're clamping on something.

Dean Pick: Right. So like, uh, if you're like a robotic finger and you don't want to crush the egg that you're holding or something like that because of noise or whatever else.

Chris Gammell: Right. Exactly. It could hold an egg or if that egg was 2000 tons, it could hold that too without, you know, marring the surface. Yeah.

Dean Pick: Yeah. Cause when, and then you're also saying that the amount of power that it takes to hold is also much less cause that's the, that's one of the things I had never thought about before I started looking into this stuff was, you know, you say you spin up a, you know, you have a, uh, what did you call it? Like a ball screw linear actuator type thing. You spin it and you spin it and you spin it and you hold it at the top. You don't have to, you know, use the same amount of force as you used to spin it up to the top, but then you do have to impart some force. You can't just turn off the motor and have it stay there. You have to either lock it in place or do something else or keep applying current to the motor to hold it in place. And it sounds like for your application to hold the same thing in place, there's just less current because the, the, it's just like you said, maintaining, maintaining the heat in the wire.

Chris Gammell: Right. And, uh, when we first approached it, uh, we were, because we were coming from a background where we were doing clutch control, which is sort of a cyclical thing. We ended up spending a lot of time figuring out ways to suspend this material in different types of cooling fluids. Um, but then when we started approaching different applications for, you know, positioning, uh, that's when we realized that fluid, uh, actually doesn't help us. So we then returned to airfield and are now looking at different forms of insulation to really make this a green actuator for that subset of applications.

Dean Pick: Hmm. Okay. And then, so, so we were talking about the, we first started talking about like, if you part, put current through it, it's going to contract. You do have as much control then on the other direction? Like, can you actually impart force then when you release or is it, is it only in that, that contracting direction?

Chris Gammell: Uh, we tend to think of it as a, a one way, like a unidirectional device, uh, device, kind of like a solenoid. Um, we do use a bias spring again, like many solenoids to return it and that, so you can get some, uh, some push in the opposite direction. Uh, but it's not, uh, it's not the same amount as it's, you know, forward direction.

Dean Pick: So that actually would push the, so I'm thinking now of like a tall box that has a plunger hanging out of it by that three to eight inches you talked about. You turn this thing on, the plunger can go all the way down. It can actually pull something down, but then if you release that power, then it's just going to pop up at whatever the, the spring internally is. Uh, yeah, that's right. So then you could have a stronger spring in there, but then it also takes more energy to overcome that spring force, right?

Chris Gammell: Um, it's not so much, uh, energy as you're basically going to sacrifice how much you can lift or push, you know, push or pull. So, you know, that spring is basically cutting into the available, uh, power available at, you know, to do the work that you're trying to do. Okay.

Dean Pick: So then, so I, and like I said, I did reference this chart, but I didn't actually come back to it. The, uh, the, the different force and displacement stuff. So it's, it's kind of showing, you're showing it kind of in the middle there, but what, why are those other things kind of gaps? Why, why was there that gap that you, you're starting to fill with the kinetics, uh, product?

Chris Gammell: Uh, it was really just a kind of a discovery after we looked at, you know, what else is out there, you know, our, for us, the, the technology falls under this small displacement range. So something kind of akin to what a solenoid can do, but at a much higher degree of force. And there really is no other technology that falls into that, that space. So it's, it's not that it was kind of a gap that people had missed. It's just that the science, uh, or the available tech that was out there couldn't really easily service that. So what we saw was a lot of different technologies were, you know, trying to overreach a little bit or do things in, you know, a variety of kind of fancy mechanical ways to, to enter that space. And by, by that, I mean, the piezoelectric actuators start to look at inchers so that they could amplify how far they could go by taking small steps and like rationing its way in. Uh, whereas let's say the electric ball screws, they started getting into, you know, very fine, finely machined parts and very fine pitches so they could get into, uh, super fine resolution. Um, so these are sort of ways that other tech sort of is flying into service space. But, um, you know, looking at the strengths of the material, we think that, you know, there's lots of areas where this can outperform.

Dean Pick: Yeah, that's great. Well, let's go back to the, uh, the, the way you kind of came across is I, I, you know, we had obviously been introduced by former guest John Davis. Uh, and, uh, and I had expected you to be like, uh, I, not expected you, but I, I thought this was all going to be about material science and stuff like that. But I, uh, you're a, you know, you're basically a consultant who kind of came into the solution. So how did you actually, and I'm not saying you don't do material science, but how did you start to discover that this was a potential solution? Like, what was the process to get to the point where you're like, Oh, not only is the solution for our job here, but also it might be a product.

Chris Gammell: Right. Yeah. And you're definitely not offending me by, uh, not calling me a material scientist. I'm definitely not by, uh, by any stretch of the imagination. Uh, really how we got to kinetics and got on shape memory alloys was, um, at by performance, uh, the real purpose of that company was to create transmission controls for motorcycles. So completely different field of work, completely different, uh, different focus. And, uh, we had this, uh, overall system design in mind that would combine, um, gear control and engine control. And, you know, more importantly for this, uh, topic, uh, clutch control together to give a motorcyclist, uh, either push button shifting, uh, or automatic shifting. And it was this, uh, this clutch actuator that really stumped us for quite a long time. And what I mean by that is, um, you know, a motorcycle on, uh, sorry, a motorcycle clutch is a little bit like a automotive clutch on a manual car where you have, uh, in the case of a car, you know, you're pushing a lever into the floor on a motorcycle. You're pulling a lever into a handlebar, but really that, that lever, uh, is acting either on a cable or in the case of our, uh, our target bike, it's a, it's a hydraulic system. Um, and in this case, it's like a master slave arrangement where you're, you know, you're basically the master and you're pushing fluid down a pipe and it's pushing on, uh, on a slave cylinder, which is mounted on the engine, which is working the clutch.

Dean Pick: And it's still doing the same, the same clutch action. Like, uh, I've, I've seen animations of clutches before. It's basically just like disengaging the, oh boy, I've got, I should say, Dean, I've gotten myself in so much trouble in the past where I started to try and talk mechanical stuff. I'm like, oh yeah, the cam shaft, blah, blah, blah. Yeah. I don't know. Uh, but I've seen, maybe I can find an animation of a clutch happening. Could you quickly explain what it actually, even what a wired clutch does?

Chris Gammell: Yeah, no problem. So, uh, you know, super simple clutches, not all that different from what you're going to find in, in vehicles. You know, you imagine two, two discs that are pushed together. So they're kind of transferring torque, right? And what's pushing them together are usually just a bunch of springs. So in order to get a clutch to separate, you know, in order to, to get it to slip or to separate, you basically have to overcome the force of those springs. And that's what you're doing with the hand or your foot, depending on whether you're riding a motorcycle or, uh, driving a standard car.

Dean Pick: Okay. Yeah. That's pretty, that's pretty simple. That's good. Oh yeah. Deadly simple. I'm just going to defer to all things mechanical for you, you know?

Chris Gammell: So in our case, um, what's happening at the hand is you're basically controlling a master cylinder, which another word for that is like a piston pump. So you got, you know, got a chamber and it's a gut, you got a piston and you're forcing that piston in and there's something on the opposite end doing kind of the opposite effect. So like, uh, you know, for every reaction, there's a, for every action, there's a equal and opposite reaction. Um, down at the slave, you're, you know, it's basically pushing a rod that's separating those two discs, uh, in the, uh, in the engine. So what we found was, you know, we needed something to push this piston, you know, to more or less replicate this master cylinder and, you know, solenoids definitely couldn't do it for a variety of reasons. Um, too weak and they had no positioning control. Uh, you can, you know, like the

Dean Pick: feedback component you're saying, right? Like if it's, it's usually like, uh, it goes, but it can't, you just can't tell where it is.

Chris Gammell: Yeah. This, this idea of like a bang, bang, like two position definitely wouldn't work for us. Um, motorcyclists are very attached to their vehicles. They kind of feel everything that's happening. So you need a very high level of control. Otherwise it's just not going to fly. Um, right. So that, or you might fly off the bike. So there is that also bad. Yes. So PISOs were definitely out and that kind of led us to take a serious look at the, uh, the motor driven, uh, actuators, like the ball screw actuators where you're taking, taking a motor and turning that into something that's going in a straight line. And, you know, I always had a fundamental problem with that. It kind of drove me insane that, uh, we had to start with rotary in order to get linear, but I put those, you know, I put those sort of gut, uh, gut feelings aside. And we really started to try to develop, uh, this piston pump using traditional actuators. But, uh, we, no matter how many different, uh, variations we tried, it always ended up, uh, too big and too bulky and usually too expensive as well.

Dean Pick: Yeah. Cause I imagine you have to do it pretty fast. Like what were the constraints on the timing you had to get to?

Chris Gammell: Uh, kind of in like, like big round numbers, a quarter second was kind of our max for how quickly we want to slip the clutch or disengage it. And then, you know, we needed very fine control. So it's not that, you know, we're bringing it to just one position. We needed that kind of, uh, response time across a range of different travels and pressures. And then it also would need to lock up as well, uh, at a pretty quick rate. So you didn't, uh, burn up your clutch. So what does that mean when someone says, Oh, I burnt my clutch? Well, if they, you know, if you let those plates slip too long at, uh, speeds that are, let's say a little bit too high, you're really just burning up the material that's meant to be gripping itself. Um, so you can really do some damage by, uh, by letting a clutch slip too long.

Dean Pick: So is that almost like two, like two sanding plates, kind of like two sanding, uh, discs kind of going against each other. Is that kind of the idea or what does that actually look like? The, it is a consumable you're saying?

Chris Gammell: Yeah, it is a consumable. It's a, think of it as almost like one sanding disc working on one steel plate.

Dean Pick: And then you just sand it for too long and one's, one, one's going to win, huh?

Chris Gammell: Well, yeah, you'll lose your sand paper and you have to toss out, toss that out and slap a new sheet on. And while that's, it's a little bit easier set and a little bit more expensive than just a, you know, hardware store grade, uh, sand paper.

Dean Pick: Right. Right. Well, and you said, um, you said you need to be really precisely controlled. Is that because, because you're going at different speeds and you're disengaging at different amounts of torque on it? Like what, what actually, why, why does it have to be so precisely controlled?

Chris Gammell: Yeah. So great question. Um, we set out to develop basically something that could control the torque from between the engine and the rear wheel of the motorcycle. So it wasn't that we were just trying to push, push plates apart. We were trying to develop a system that could control how much torque is getting delivered. And it wasn't just an on off or a, you know, a zero and a one. We really wanted fine, fine control of just how much torque would go through. And that, that, uh, turned into a transfer function or an equation that relates, uh, you know, torque to the speed of the engine to the, the amount of force or pressure we're applying on the, on the, uh, clutch plates. And there's even a travel element to it as well. So, um, it's a, it's a complex system and we needed, you know, very fine graduation in order to tailor, uh, the gear shift, if you will, to a, you know, the specific riding conditions. And it's, there's just so many elements in play that we kind of needed a high resolution device as a result.

Dean Pick: Is this, uh, is, is this something that you would have patented as like anti anti-wheely control or what, why, why, why did you, why, why did you decide to go on torque versus just like speed or other things?

Chris Gammell: Yeah. So the, the clutch is in essence, a torque control device. Um, and you know, the same could be said, let's say for like a brake disc on a car or on a motorcycle, it's a torque control device. You know, in the case of the brake, you're trying to take, take energy out of the vehicle and you're applying braking torque to slow things down. In the case of the clutch, it's usually locked. So you kind of want a one-to-one, but in those special cases, when it's slipping, uh, you need to start getting into some different ratios of how much, how much torque is going. And it's not always going from the engine to the rear wheel. Sometimes it's going the other way too. And it's these dynamics that can really upset the bike, upset the chassis and upset the rider.

Dean Pick: And just you're saying, cause of vibrations and just how, how it actually brings the force back into the system.

Chris Gammell: Well, you definitely want to be very smooth with the clutch control. Yeah. You don't want to send any torque spikes through it. Cause again, it comes back to, uh, to rider safety.

Dean Pick: So what did this ultimately, I mean, so this is a product that's still out there, right? I mean, you're still making the shift FX. Is that right?

Chris Gammell: Yeah. Shift FX has a quite a few products in the aftermarket. Um, but this product that uses the shape memory alloy specifically is really designed to cater to the, the factories that would install it as a factory option because there's just such a, such a number of, uh, aspects that we're touching on the bike that it's not really something that, uh, that we can sell today as a bolt on product.

Dean Pick: Oh, Oh, you're saying because of the level of integration that you need to have basically.

Chris Gammell: Yes. Yeah, definitely. And particularly around not so much the clutch control as the engine control, uh, mostly for speed matching.

Dean Pick: Yeah. How much, I mean, I, we've had people talking about ECUs on the show before, but in automobiles, engine control units are just really the processing in there. How much of that's happening in the motorcycle world these days?

Chris Gammell: Um, in some cases we're lagging and other cases we're leading. Um, okay. Uh, like a modern, uh, motorcycle has a, has a whole suite of typically can bus connected modules. Uh, so you could have an ECU, which is doing the, uh, the fueling and the ignition. Well, it's talking to a, a fly by wire throttle device, which is controlling, you know, the throttle or the amount of air that's going into the engine, uh, which is talking to an ABS module, which is controlling the, the anti-lock brake system, which is talking to a suspension module, which is controlling active suspension.

Dean Pick: And, uh, and all this is fitting between someone's legs too. So there's, there's also that it's not like enveloping them. Right. Right. Crazy.

Chris Gammell: Right. So it's packaging is uber important. Luckily these, these parts are small. And, um, one of the more intriguing devices, which we kind of started dabbling in, uh, recently was the, the inertial measurement unit, which I'd say in a lot of, you know, from a lot of aspects, the motorcycle industry is leading. So, uh, we're talking about a six axis, uh, IMU is the, the acronym or inertial measurement unit. It essentially gives these different, uh, CAN bus enabled devices, uh, an insight into exactly what the motorcycle is doing, how far it's leaned over, if it's going up or down a hill, how it's accelerating, if it's skidding, sliding, wheeling, um, just a really amazing window. And with that kind of window, you can really improve how, how the bike responds to your inputs.

Dean Pick: Really? Okay. I didn't realize there was that kind of, I mean, like to be completely honest, I, my, my knowledge of bikes is very low, but I always thought that they were very much just a old school engine, you know, with a chain going to a rear wheel. So I'll have to apologize about that. So I didn't realize there was this much level of control. And also, is this for like racing bikes, the street bikes, or, you know, what, what are we talking about here? Right.

Chris Gammell: So, you know, all this tech that I'm talking about, you're going to find it on the high end, uh, sport bikes. You're going to find that on the high end touring bikes, but you know, just like, um, electronic fuel injection, it, it trickles down even into, um, you know, uh, low entry commuter bikes, for example. So it's really just a matter of time. Uh, and this stuff will become more and more pervasive. Still no airbags though, huh? Well, I think Honda tried to do that and actually has those things installed on their, uh, Honda Goldwing. Um, really?

Dean Pick: Oh, that's the really, really big one, right? Like the, like the cruising bike.

Chris Gammell: Yeah, that's a, it's a big, big bike, but, uh, maybe more interestingly, um, a lot of the, uh, suit designers. So we're talking like leather jackets, uh, they're building basically the equivalent of an airbag into those jackets and they have really interesting sensors, basically like IMUs to tell them when the rider has a, um, well, we call it a get off event, but it's a fall. Um, and it, it, it triggers that airbag so that, uh, before you hit the ground, you know, it's supporting collarbone, neck, those kinds of areas. Neck support. Right, right, right. So you're basically strapping an airbag to you.

Dean Pick: That's, that's great. I mean, that's great. Um, and I think that's kind of like a really good use of technology here. The interesting thing to, that we were talking about before the show too, is that like the fact that this could be a completely automotive, uh, a completely, um, automatic, uh, like, like just like an automatic car. So I could hop on a bike and potentially be able to not have to throttle, uh, not have to use the clutch at all and shift at all. So it kind of makes bikes more accessible.

Chris Gammell: That's right. Yeah. Essentially you get the convenience of a scooter with all the, all the benefits of a motorcycle. So easy to learn, easy to operate. Uh, but it also gives you a, can give you a progression path so that, you know, when you're ready to start, uh, you know, experimenting with your shifting, you can flip it off of automatic and actually start picking your shift points. And, you know, that's sort of what we would like to see from any entry level rider. We're, you know, we're, we're, our philosophy is not to take control away, but to, um, to, you know, give them a way to get on a motorcycle, but also get the most out of it. And, you know, for us, that means eventually trying, trying it out yourself.

Dean Pick: Right. Yeah. I think that's, that's a really good point too, because you think about all the things you have to learn when you're getting on a motorcycle and you have to learn how to shift while you're also balancing, you're also turning and signaling and all the other things. I'm sure from a safety perspective, that really helps a lot actually.

Chris Gammell: Yeah. It can definitely help with rider concentration and, um, it really can help with balance too, because, uh, you know, while the clutch is on the hand, the shifting is done at the foot. So you're constantly repositioning your body and changing your weight to do the shift. And that can be pretty unnerving for, uh, for new riders, particularly those that have never, never driven anything, but their parents' automatic car, which is becoming kind of a reality.

Dean Pick: Count, count me in for that one. Uh, and driving less and less these days too. I mean, honestly, um, I think, and I've, I've thought about like motorcycles from, in terms of, you know, efficiency, you know, you can get high MPGs or whatever kilometers per, uh, whatever the non-imperial units of efficiency is. I'm with you. Yeah. Yes. Yeah. Um, uh, but, and you know, obviously it was just the, the enjoyment factor of it, but, uh, it just seemed, it seemed like really out of reach to be honest.

Chris Gammell: Yeah. And, um, I think what, uh, what we'll see, hopefully to get more people interested in it, it really is an exciting, uh, uh, pursuit either for commuters or, uh, just for, you know, enthusiasts is we're going to see more and more electronic bicycles, which are going to start putting motors between your legs. And those bikes are very easy to jump on. You know, there's no, there's no real perceived fear factor. We're talking about sub kilowatt motors, but, um, I have a lot of feeling it's going to be a gateway drug.

Dean Pick: Oh yeah, I think so. I mean, I, I, I had a chance to ride an e-bike out in Seattle when I was out there and you know, those hills are still, they're not easy, but, uh, they are a heck of a lot easier and, and you can just zip around town really. So that's, uh, it was, it was really cool. Um, how did you get into this in the first place? I mean, are you a rider? Like why did you start going down the path of motorcycle design in the first place?

Chris Gammell: Uh, I can trace that back to about 2002 where, uh, I was involved in a university, um, varsity challenge, like an inter-university competition called formula SAE that pits, um, basically student designed race cars against each other. Um, but because of the rules of the competition, most teams ended up using a motorcycle engine. Um, and, um, back in our day, we were shifting that engine using really crude and rudimentary paddles and cables. And as a driver of that car, um, it made me, you know, really irritated at how clunky it was. Um, but as a, uh, you know, a student learning engineering, it, you know, gave me an opportunity to get involved in that car, turn it into, I guess you could kind of call it a thesis project to turn that motorcycle engine into a tiptronic or push button style, uh, uh, race car. And that's, that's exactly what I did. And, uh, after graduation, um, we realized, or I realized that, uh, there wasn't a lot of products or features like that in the motorcycle market. So it kind of led me into, to that line of work.

Dean Pick: And I mean, when, when you have all this stuff too, I mean, there, it seems like there's always kind of this underlying level of control there as well. So like, what does that, what does that actually look like when you're designing a new product? And I guess this is just more broadly for your consulting company too. Um, what, what are you actually doing under the hood in terms of control? Like what kind of micros are you doing? Are you doing Linux? You're doing bare metal. Are you doing RTOS? What are you actually putting inside the controllers? Right.

Chris Gammell: So we are starting from bare metal. Uh, we are, we have developed our own, uh, distributed operating system that can work, you know, either on the same, uh, circuit board between chips that monitor each other or can communicate to a third party device that, you know, was within our suite of products. So, um, yeah, it was a ground up effort to develop our, our operating system.

Dean Pick: I mean, why did, why did you decide to do your own versus doing it off the shelf RTOS or similar?

Chris Gammell: Um, that's a great question. Uh, maybe easy answer is stubbornness.

Dean Pick: Um, but a common answer to it. Yeah.

Chris Gammell: I, I can't tell you, I mean, outside of the context of this question, how many times, you know, we've tried to incorporate third party software or firmware solutions on a whole variety of applications. And, you know, for some reason we always end up in a trap where we're spending more time training and learning and then realizing deficiencies. Rewriting their code. Yeah. Well, usually these things are locked, um, or prohibited from, from touching to maintain warranty or what, uh, what have you for the license. Um, and it's when you hit these barriers or you identify, you know, things that you can't do, but could do if you had your own foundational OS that, uh, that just keep reinforcing this, this need for crowned up development.

Dean Pick: The best part is you get to write your own drivers. The worst part is you have to write your own drivers.

Chris Gammell: Yes, it's true. The, the interfaces is probably the worst

Dean Pick: part of the job, but it also sounds like, I mean, like if, if you stuck with it, I mean, I think it'd be problematic if you're like every year, like, all right, we're starting over again. But if you're, you know, maintaining it and you're developing as a, you know, IP effectively for years and years and years, and that can be really valuable because then the next time you start a project too, you also have, you have this knowledge base that you've created. That's also proprietary that you can use over and over again.

Chris Gammell: Yeah, it's, it is a good fallback. And, um, I mean, perhaps the most rewarding aspect of it is, uh, if I can relate it back to the motorcycle and the, the electronic shift transmission is, you know, we are constantly riding and, uh, testing when we find that there's an aspect or a, uh, a behavior that we're not completely happy with, um, we can go in and actually, you know, go straight to the, the heart root of the problem. And it could be quite abstract or complex, but, you know, with that sort of broad understanding of the whole system, um, it, it makes the, a change happen quite rapidly so that we can go back out and validate or, you know, test internally validate and then go to the, uh, to the street testing. But, um, I don't think, uh, don't think it would be too, too easy to do that. If, you know, we didn't have, uh, you know, a very deep level understanding of, uh, of how it's all put together.

Dean Pick: And so like, what is the relative power of like processing that's on board? Is it like M zero level and fours? Is it like, uh, you know, higher level, like a Linux, like a IMX six or seven? What, what, what kind of levels are we talking about here?

Chris Gammell: Yeah, it's, uh, we're able to do most of the functionality with just eight and 16 bit processors. So even below M four, uh, but I see us transferring to yeah, 32 bits, uh, potentially floating point math because, um, we have a variety of physics engines that are constantly running and updating inside that chip. And we see a path where we need to, you know, develop even more advanced physics engines that are, you know, need to execute in real time. So, uh, I see us, yeah, getting, going there eventually.

Dean Pick: Yeah. Well, and then you get more, you get more access to better peripheral stuff like that. You get the can buses, you get the lin buses, whatever, whatever else is thrown in with the kid, they get everything in the kitchen sink that they throw into chips these days, but it's always a painful switch over. I'm sure.

Chris Gammell: Yeah, definitely. It probably will depend on, on the level of the level of target bike and the price points before, uh, before we make the jump, it'd have to be a pretty premium product. Uh, we can do a lot with the, uh, the eight and 16 bit systems. Um, but, uh, you know, to really push the state of the art. Yeah. You know, we are, you know, when we talk about running physics engines, you know, we're, we're essentially predicting what engines will do, what clutches will do, what bike, uh, chassis will do. So, um, the, uh, the math can get quite complex, especially as you start to throw in the, the feedback from the IMU and, uh, and now are becoming sort of aware of the whole state of the motorcycle. Uh, all the forces, all the accelerations that are acting on it.

Dean Pick: And so like, what is it, does it, is it like, uh, can you explain a little more? Like, so if the bike's on its, say it's in a bank and it's like turning on its side or it's leaning to one, one side, will that actually change the nature of the motor or is it just like a, is it like a big state machine that's internal to it? Like how, how does that actually end up impacting the, the control of the engine?

Chris Gammell: Right. So the engine power is not too affected by the, the lean angle. Um, but certainly the traction, um, the traction or let's say available traction on the tires goes down as a result. And the reason why is because some of the, uh, the available grip is going towards keeping, keeping the bike from sliding because it's got to react to that centrifugal action from the, the turn itself. Um, so, you know, that's where torque spikes because of, you know, events on the, on the motorcycle have to be carefully managed so that you're never, you know, breaking grip in a way that might cause, uh, cause an issue.

Dean Pick: Oh, I see. Okay. So you might even, so if, so just to keep going with this example, right? So if the bike is leaning and you're saying, okay, we definitely can't have a torque spike, then you're just, that's an input then to your, your clutch program program because you can, you can't, you can't bear as much slippage or whatever happens internal to the clutch then.

Chris Gammell: Uh, right. So it might create like a, an inhibit, uh, event, uh, that might, you know, feed into, we do use state machines quite heavily. Um, so that kind of information can be used to change sequences or inhibit certain actions. Right.

Dean Pick: You have to be clear of some flag in order to actually engage an actual shift event or whatever, whatever the actual state machine is. Uh, yeah, that makes sense.

Chris Gammell: Yeah. I mean, it's, uh, it's not just used by potentially us, uh, for ABS systems, you know, lean angle is also critically important because if somebody's, you know, reefing on the, uh, reefing on the brakes because they're in a turn, they don't like where they're heading. Uh, that module has to be very careful on how it, how it, uh, controls the braking. Cause if, you know, it may decide not to, not to touch the brakes at all. And that can cause a problem, right? And it's going to push the rider too far forward, but if it locks, if it, yeah, the alternative though, is it locks and then the bike, you go straight down to the ground at that point. So it's that happy medium, which is oh, so, you know, oh, so critically important, but, uh, challenging to obtain.

Dean Pick: And so, and you're mentioning like third-party modules are just generally modules outside of one thing. So this is all talking over a canvas, but is there like standardized messaging? Like, how does it know, how do you know that you're, you know, you're this, this shifting unit and there's also an ABS unit. Do you have to talk the same language or like, is there like standard, standard headers that you send to one another? How do you actually standardize that communication?

Chris Gammell: Uh, yeah, it tends to follow the canvas protocols. So, you know, every message will carry an ID, which identifies it. And then, um, you know, that, that message might, might matter to some modules and it might not to others. So it's, it's really up to the, that point, that node to, to say, yeah, I'm interested in that information and take the message in, process it. And if there's an expectation for a reply, then that can go out or, or not.

Dean Pick: Um, so some of the protocols you're talking about though, like, again, you have to excuse my ignorance here. The canvas protocol is actually saying like, it's like a published mechanism for the braking system in general. Like, Hey, I'm a braking system. I'm about to fire or something like that. Is that, is that kind of how it works?

Chris Gammell: Um, yeah, I would say it's not standardized. No. So it has to be kind of established when, you know, by the people who are designing that canvas network, you know, these are the IDs, you know, we're going to do, let's say this style of messaging, it's going to go out on a fixed, like beat pulse or frequency, or this is going to be event driven, uh, messaging. Um, it's flexible essentially. So it's not really, uh, stipulated. Uh, it could be stipulated by a manufacturer or, you know, a higher, higher organization. Um, but it's not, uh, it's not that our TCU is message ID 99 and it'll always will be 99 and it'll always transmit.

Dean Pick: There's no lookup table to see. To see the, the RX, uh, whatever or shift effects version or something. Right. That's right. So then if shifted, the shift of X want to work, wants to work with different things, you have to have program in which ones it works with. Is that kind of the idea?

Chris Gammell: Yeah. And that again, kind of goes to why we're, we're approaching the, the factories is you can't just drop, drop in a canvas module into the, into the bus and expect everything to work. It's, you know, these devices have to be set up to either provide the information we need, or we need to send the information they need in order for these, these features and functions to actually be, um, be working and useful for the writer.

Dean Pick: Yeah. I mean, yeah. Otherwise you're just sending, you're basically sending mail to, to something without the right address, right? You're just sending mail to the central location. You're like, yeah, well, hope it gets there. Yeah.

Chris Gammell: Just a bunch of spam. Um, right. You know, on the flip side though, you know, there are third party tools that can listen to the network and that can, you know, I'm, that can give you diagnostic information. So those can be quite useful as well.

Dean Pick: Right. We've had Erlon from, uh, Machina before when they do an open source canvas protocol analyzer thingy. That's where I think I remember hearing some of this stuff, but it's still, it's amazing to me that there's still no, like, it still seems, uh, not abstract enough, but like, uh, it's so it's, there's not much cooperation between the companies. Then that's, what's amazing to me. It's like, so it's like you have to talk the specific language of each, each module or whoever's it is. The assumption that you're always working with the same company's equipment or you can talk to it.

Chris Gammell: Yeah. I think, uh, you make a really good point there. It's for some, it seems to be like, uh, barriers are thrown up for the sake of just putting barriers, uh, in place. Uh, they're not always driven by the right, you know, the right mindset in my opinion.

Dean Pick: Yeah. The motivation is money like usual, but that, uh, that is a consistent thing in our, in our world, uh, regardless of profession. That seems like speaking of money, uh, you guys also do other like consulting stuff and that's kind of how it seems like you got into this. What else, what else does, uh, sorry, I keep forgetting the name of the, the by it's by performance, by performance, by performance.ca. Uh, that's your website. Uh, what do you, what else do you guys kind of focus on? Is it mostly in the automotive motorcycle realm or is it outside of that as well?

Chris Gammell: Uh, so yeah, our products division, whether it's robotics or motorcycle products are, I'd say pretty far removed from our consulting work. Uh, we take more of a regional focus to our consulting work and, uh, in my home province of British Columbia that, uh, you know, tends to lead us down the road where we're working with power generation, uh, forestry, marine industries here. So, uh, we've worked on a whole variety of kind of interesting projects. Um, so over the years I've developed kind of a side, side set of expertise around, um, power generation from biomass, which is a fancy word for like wood waste.

Dean Pick: I was going to say burning, burning trash. Yeah. But making it into electricity, which is actually really cool. And that's, uh, isn't that the one where it wasn't like Sweden or Finland like ran out of trash. I remember some Nordic country, they they're like, yeah, we're out of trash now. We need things to burn so that we can, we can, uh, make electricity. It's like, oh, nice problem to have. Too bad shipping is so expensive. You know? Yeah.

Chris Gammell: Well in British Columbia, you know, our forestry industry is, is, uh, a huge part of our, uh, economy and it generates, you know, a lot of wood waste. So that, you know, for us, that's, it's not a, let's say like a municipal trash stream where you've got all kinds of contaminants in it. It's actually a pretty clean stream. It's really just a question of how wet it is when it comes to us.

Dean Pick: Interesting. So then is that literally just like using sawdust to boil water kind of thing and then make steam or what, what, what is it, what does a bio biomass plant look like?

Chris Gammell: A lot of, uh, a lot of forestry facilities, particularly like pulp and paper plants will have things called power boilers, which take wood and burn it. And the ways that it burns that wood are pretty, pretty varied. You can have something as simple as a pinhole grate. So you're just, you know, throwing wood on a, on a plate that, that has air coming underneath it and you get a big fireball and that's, that's used to, to generate steam, you know, boil off water inside tubes. Um, but, uh, you also have kind of really interesting and unique ways of, uh, of burning wood. Uh, one of the projects I recently worked on use something called a fluidized bed, which was, um, a giant industrial sandbox that was, uh, aerated from underneath with a 1200 horsepower fan. So the whole sandbox would lift up and you drive wood into it and you light it off. And that thing essentially becomes like Dante's Inferno. Um, just a giant, giant swimming pool of fire. Yeah. Pretty amazing.

Dean Pick: Giant swimming pool. I like this. Yeah. Yeah. That's not, and it sounds like, uh, because of the air injection too, it's kind of like when you superheat a kiln or not a, but a forge, right? Don't you usually inject air to superheat like coal or Coke to do like smelting?

Chris Gammell: Yeah, definitely similar to that. Um, so the reason why they use sand in this case is, um, it, uh, it actually envelops the, the, the solid wood as it gets introduced into the boiler through some kind of a feed mechanism and it helps, uh, burn wet wood, which is typically what you're getting. So it'll quickly, you know, evaporate that water and then start, um, breaking down that wood into its, uh, you know, molecular or atomic constituents. And then that's when you can start generating combustion and getting heat back from that fuel.

Dean Pick: And this is, this is for like localized generation at a plant, or is this more broadly for like powering homes in the area?

Chris Gammell: Uh, it's a bit of a mix. Um, you know, if it's a pulp and paper plant itself, it tends to have a massive power demand. So it'll consume its power. Um, whereas, uh, this last project, uh, with the fluidized bed, it was a standalone power plant. So wood came to the plant and it exported its power to, you know, keep the lights on in the, in the houses in the area.

Dean Pick: Well, that's cool.

Chris Gammell: Yeah, that's really cool.

Dean Pick: What are some of the, uh, the other constraints in that system? Like, is that like, or like what are the parts that you're actually designing in this, this setup?

Chris Gammell: Kind of the, one of the more interesting aspects for me as, you know, from a control engineering point of view was I had a heat and mass balance for this plant and it had, I want to say about a hundred different nodes. Um, so you're, you've got this open cycle or open loop cycle on the, the combustion side where air is coming in, fuels coming in, and then you have, you know, everything coming out the stack like, uh, CO2 and, um, nitrogen, et cetera. Lots of, uh, lots of H2O. Um, cause you're basically just steaming that off the fuel, fuel. Right.

Dean Pick: Right.

Chris Gammell: But that, that whole process, which includes, you know, lots of thermal equations to figure out what kind of energy you're going to get out of it has to intertwine with this, uh, closed loop water system, which is very, very, uh, complicated. And, you know, it's, uh, amazing how many different times you need to take water at one state or steam at one state away and reintroduce it here and pull it out there. And, um, you know, really at the end of the day though, uh, we're just passing it through a turbine and it's going back to the boiler and back and forth and back and forth.

Dean Pick: Right. Right, right, right.

Chris Gammell: The details are, um, yeah, amazingly complex and the, the kind of control equipment you need, uh, to modulate all these things.

Dean Pick: Right. Well, I mean, you don't want to just, you don't just open a steam valve with your hand or something, you know, you want to use machines and, and there's a lot of pressure and temperature, like you're saying, and pretty crazy.

Chris Gammell: Right. And we try to cut out the manual control as much as possible so that, you know, somebody in an operating room, um, can sit back and have all those, you know, levers and handles more or less at the click of a button. And yes, what was quite cool on that project is, um, and I will take no credit for it, but, um, we put the whole control plant on a, uh, on a tablet that, um, an operator could then take anywhere in the plant. So you essentially can have a control room with 20 screens fit on a tablet and you can sit anywhere you want and control it and actually see the, see the events happening right in front of your eyes.

Dean Pick: That's yeah, that's pretty cool. But usually that also means that, isn't that also like SCADA systems that are then like opening themselves up to like port, like open ports in the internet kind of thing where like there's municipal water stations that are like just open with like no password kind of thing.

Chris Gammell: Uh, this, uh, this was a whole closed, um, I'm not sure if intranet's the right word. All right, closed system. Yeah, closed system. Yeah, yeah. Sorry, no back doors that I'm going to reveal in this call.

Dean Pick: No, that's good. I, that's what, I mean like, honestly, that's what I want. And it's like, I feel like the, the modernization, like some of them, it sounds like that's actually a really useful thing. The tablet, like you're talking about to have that you're on site, you're like, all right, now I'm going to actuate this valve and I'm standing in front of it and I see that it turned red and you know, the light turned on and I can hear it going like that's actually really valuable instead of, you know, radioing between yourself and the control room or whatever. Um, so that's, that's pretty cool. But I feel like in, in the same stroke, they're also selling like, and you can do that from your home. It's like, well, maybe you don't need to do that. You know?

Chris Gammell: Yeah. I kind of like the philosophy of, um, like, yeah, closed, uh, closed intranet around the plant and particularly all the controls with, let's say a one way firewall that allows telemetry out because there are decision makers that aren't in the plant that need that information. Um, but, uh, you know, the idea of a, um, a manager remotely shutting a valve is, uh, is terrifying.

Dean Pick: Yeah. Yeah. I agree. I agree. Uh, yeah. And I mean like, and this sounds like it's big stuff. I mean, like what kind of levels of power generation we're talking about here too?

Chris Gammell: Uh, so that last power plant was, uh, 36 megawatts and, um, I want to say it was burning, uh, let me see about 60,000 pounds an hour of wood waste.

Dean Pick: Wow. And so wood waste is like coming off like, is this like branches that are trimmed off things that are getting turned into planks or like what is wood waste or is it like the pulp like you're talking about, like the offshoot from the pulp plant or?

Chris Gammell: So it was mostly a sawmill residuals. So, um, you know, the trees are coming in with, uh, with bark and branches. Um, sometimes, uh, with tops. Um, there's a lot of slash that's still left in the bush, but sometimes because it now has value as a, as a fuel that can be drug, drug in, uh, by truck. And then the sawmilling produces quite a lot of sawdust, uh, itself. And, um, you know, you get thin rough cut lumber that has to get, uh, planed into finished lumber. So that planing creates a lot of chips and then you've got a lot of kind of odds and sods that don't meet grade that kind of end up in the, in this pile. And, you know, depending on where you're situated, all that stuff has value. So these plants are the power plants, plants fight with, um, other plants like pulp and paper that can actually turn that fiber into paper or they're fighting with, um, uh, there's quite a few plants in British Columbia and Canada and the U S at large that produce pellets that tend to get exported.

Dean Pick: Yeah.

Chris Gammell: So that, that fuel is, even though it may be just, uh, adjacent to the plant, there's somebody who's willing to take it, uh, dry it, pelletize it and, uh, you know, ship it to Europe for, uh, for use in their, uh, their biomass systems.

Dean Pick: Yeah. That's, I mean, that's like a carbon economy that's coming from above the ground below the, instead of below the ground, right? It's, it's pretty crazy.

Chris Gammell: It is renewable power. Yeah. It's a carbon economy. It's renewable. Um, it's green. I think there's ways to make it greener.

Dean Pick: Well, it sounds like if you're, if you're, I mean, like you're literally using a hundred percent of the tree too, it sounds like. So that's amazing. Um, but that's, that's really cool. That's really cool that that happens.

Chris Gammell: Yeah. And compared 20, 30 years ago, we had, um, quite a few things called beehive burners in this province where the wood was a problem. It truly was waste. So they would just pile burn, uh, in giant things that look like beehives. Um, and they just had a conveyor belt throwing it on the pile. And really the, the beehive was just a screen to keep the sparks from starting forest fires, no emissions control, nothing like that.

Dean Pick: Wow. Yeah. Well, that's great. That's great that it's, it's changing too. Definitely. Uh, what is this? So, so what is, you also sent me a video and I, I've been looking at it, like trying to figure out what I'm looking at here. This is the, uh, the screen cleaner. Is this, is this the same thing? Is this for the, uh, for this, this power plant project or something else?

Chris Gammell: Uh, so you are looking at, uh, a different project, but it is a power plant unto itself. Um, so that was a project we did for, uh, for a different company. Uh, they call it the Coanda screen cleaner. Um, I affectionately refer to it as the damn brush and really, uh, what this is, is it's like a smaller, a micro hydro plant. So rather than creating a, a big dam and forcing the water to, to go down a pipe, uh, to, to spin a turbine that's at the bottom of the dam, this is called run of river. Uh, so what you're looking at is an intake screen for water and, uh, that water's then sent down a big pipe called a pen stock way down the riverbed. And then it, uh, it goes through a turbine down there, which generates electricity and then spits it out at the bottom of that, uh, at that river. Uh, so because it's operating in nature, that screen gets, gets dirty. Uh, so essentially we need to run a giant toothbrush back and forth over it to, uh, to keep it clean.

Dean Pick: It's just a, it's a damn brush. It's a damn brush. That's great. That's great. And, and like, like what would actually get caught in there was just like branches and, and some other things like that would actually like jam it up.

Chris Gammell: Usually it's like the little bits of leaves that, that accumulate at various times a year. Like a lot of this stuff will just flow over, but there's just certain, certain geometries that just like to hang around. So brushing off little particles of leaf, uh, can make a big impact on how much power you can, uh, pull out of the, uh, the hydro plant. Uh, plant.

Dean Pick: Huh. And what's, what's the relative power of that kind of, uh, hydro plant?

Chris Gammell: Uh, that one I think is, uh, also producing like 40 megawatts, uh, which puts it at the high end of the range for run of river. So run of rivers at a totally different scale than, you know, a typical hydroelectric project, which we'd be talking like gigawatts of power. Uh, but it also has a much near

Dean Pick: Niagara Falls. So I, I've seen, I've seen some of the big ones. Yeah.

Chris Gammell: Right. So the, at least with this one, uh, it's got a much smaller, uh, footprint so you can put it in remote locations and it's, uh, you don't necessarily take all the water out of the river. So the, you can actually maintain sometimes, uh, the river ecosystem, um, you know, without.

Dean Pick: Yeah. For like fish getting back upstream, upstream, that kind of thing. That's right. Yeah. That's really great. So, and then like, but it seems like in both these cases, what are some of the constraints around, uh, designing like, so this one's an outdoor system and then, uh, obviously the, uh, you know, setting up, uh, something that has a, a large, uh, you know, fire breathing thing, you know, like what, what are, what are the, um, reliability constraints that you're dealing with?

Chris Gammell: Uh, the one in the, the, the run of river project is, is interesting because of their, they're very remote. You know, we can't bring the river to, to the maintenance yard and it's usually way out there in the mountains. Um, so things need to be able to run autonomously, uh, and they need to be very reliable. Um, not to say that the power plant in a case is not, not reliable. In fact, it, it has its own weird set of criteria in that, that plant has to run uninterrupted for 50 weeks a year. So it, when it starts, you know, you really want it to run. No joke, 24 hours a day, seven days a week, you do not want it to turn off, uh, until it's, you know, scheduled maintenance cycle. So I can't tell you how many different ways that plant can turn off. Uh, so.

Dean Pick: Can, can turn off your site? Can turn off. Just because of safety mechanism? Yeah. Yeah.

Chris Gammell: Various trips in the system will, you know, by design, take it down and, uh, as safe a way as possible. But that has, you know, can have very tough consequences on equipment because it can be very hard on the equipment itself, but it's also very time consuming to take a plant down and bring it back up.

Dean Pick: Right. I just think about the heat cycling of, you know, if you have something that's roaring like the fires of hell, like you're talking about, and then it, first off it drops really quick. I'm sure that there's some thermal events that happen. And then also you have to heat it back up. You got to make sure everything's in place when you, when you turn it all back on.

Chris Gammell: Right. Yeah. So that fluidized bed, the big sandbox, it cools down rapidly. And if it goes to, gets too cool, you can't just throw fuel on it. It won't burn. So you have this sort of window of time for that, that system. You know, the turbine is in some ways even more sensitive because it's, you know, locked into the grid at 60 Hertz. When it trips, it immediately has to open the breaker and pull disconnect from the grid. And that's a big hit of, of a generation that drops out for, you know, typically smaller remote communities. Yeah. So, you know, going back to that whole kind of telemetry case where you got to get the information out of the plant, that information becomes very important to try to provide some, some clues and hints to the maintenance crew as to what needs to get looked at, you know, before it becomes a, a major issue.

Dean Pick: Yep. Yep. That's a, that's a bad call to get at two in the morning. Is that the, you know, it's, it's, it's bad if you get a call at the two in the morning, it's even worse if you didn't have a, here's what it probably is, you know, kind of message going with it either from the inside of the plant or whatever.

Chris Gammell: Well, the emphasis is, yeah, I'd say really on a predictive maintenance as opposed to like firefighting where there's now something's gone very wrong and you have to pick up the pieces and, you know, you hear a lot of talk about, you know, AI, artificial intelligence, machine learning, big data. I don't see a lot of people talking about, you know, real use cases, but you know, these power plants and these large industrial facilities are great examples of big data. These, you have, you know, a 500, a thousand sometimes of nodes, you know, that are kind of IOT or at least sending out data that's being recorded. And so you're accumulating all this information. But the analytics that you need to run to do this pattern recognition are extremely tedious and time consuming. So yeah, I see opportunities down the road for people to get into that, that line of work for predictive maintenance using big data and AI.

Dean Pick: So this is like a distributed control system you're talking about, right? Like the having a range of sensors, that kind of thing?

Chris Gammell: Yes. Distributed control system. Well, that's an interesting question. Usually it's consolidated inside of a single plant control system, but you have all these distributed devices that are, that are, you know, in the field. We try in this case, and I understand why we try to restrict the number of distributed control systems. You do want kind of redundancy or backup of your main control system, but try to keep it all in one room or one building.

Dean Pick: That's cool. Yeah. And I've been, you're totally right though about the generating of that much data. It's just, you could look at it, but, you know, if you always just said, you know, when this sensor goes below three, then we got to turn the machine off. That's, that's, that's as almost as expensive as just having a, you know, every third month we turn it off or whatever it is, you know, like it's, it's better if you can actually, if you can hold out until the, the known point of, okay, well, you know, in the past 40 times we ran this machine, when it got to this operating point, that's when the optimal maintenance was. And we're definitely going to do it. Then you can start to optimize for that then and, and decide to take the machine down as needed, you know, so that you're not breaking anything.

Chris Gammell: Right. And I'd like to think that I, you know, an AI bot or a program can, can start to do that for us because it's, it's a lot to ask, you know, individuals in the plant to, to do that. There's so many different aspects. So nobody can, can review, you know, 2000 data points over five years to, to figure out certain, you know, very subtle patterns that, that actually can lead to, to breakdowns or, you know, bad events. Right.

Dean Pick: Right. And expensive events too. Yeah, definitely. Well, Dean, we've, we have gotten through quite a lot of, we've gone, we've gone from linear actuators to, to biomass generators. So there's a, there's a, there's a long distance in there and that's pretty cool. You're doing some very cool stuff. You had mentioned though, that you are actually looking to hire someone as well.

Chris Gammell: Trying to find a, an embedded systems guru that can come into our team and work with us as we develop a next generation of, you know, transmission controllers for the motorcycle industry. So, you know, looking for a talented individual who, you know, has an interest in, you know, actually applying, you know, their skills to some real practical problems. And, you know, that's, that's essentially it. If anyone's interested, send, send a resume to HR at byperformance.ca. And we will definitely take a look at it.

Dean Pick: Yeah, that's awesome. And you said remote might be okay for the right person too. I think that's always important if, to say if that's, if that's, you know. For sure.

Chris Gammell: We, we're looking, you know, we're happy to do a broad net reach. Yeah. We're looking for talent.

Dean Pick: Yeah. That's really great. I mean, the things you've talked about so far too, you've, you've done some very, very interesting things. So, um, I'm sure there's, there's someone out there looking to put some processors onto bikes and then maybe, you know, maybe once in a while you got to go out and test them. I don't know. You know, maybe you got to go ride around for a little while.

Chris Gammell: I won't say, uh, I won't say you have to come with a motorcycle license, but it, uh, it would help.

Dean Pick: All right. That's cool. Uh, so yeah, where, where can people find out more information about you and the companies and, you know, where they might be able to see all this stuff we've talked about today? Obviously we'll have show notes, but you know, if they want to go with directly to the, to the websites.

Chris Gammell: People want to, uh, look at the motorcycle, um, products and the transmission tech, go to shiftfx.ca. So shiftfx.com. And then the, uh, that's an important thing. And then the, uh, the linear actuator tech can be found at Kinetics Automation and Kinetics is spelled, uh, K-I-N-I-T-I-C-S automation.com.

Dean Pick: Great. Yeah. And, uh, byperformance.ca is the one you mentioned before too. So that's good. I'm sure you're going to get some, some interesting people applying there. Well, Dean, thanks so much for, for talking here today. I really appreciate it. And, uh, I look forward to hearing about the next thing that you're putting together.

Chris Gammell: Definitely. Thanks for having me on the show, Chris. I appreciate it. All right. We'll talk to you soon. Sounds good. Bye.

Speaker ?: Bye.