- Y-Combinator alumni from the Winter 2016 batch
- Currently serves as COO of FlowCommand
- At FlowCommand, Jeff is responsible for all hardware design, firmware development, and operations
- Spent his early career as a Production Engineer and Lead Completion Engineer at Oxy USA
- Has been on Y-Combinator’s alumni application review committee for the last 3 cohorts, has raised three rounds of financing, and manages a staff of 17 engineers
- Member of Incubate USC, where he has assisted in writing undergraduate coursework on accelerators and incubators
Industrial product design and testing
- What makes a product an industrial device?
- Prototyping Industrial Hardware
- Arduino and development boards -> First Article
What kind of regulations and certifications do you have to work with?
- How to make it work when you are bootstrapping
- Class 1 Div 1/2 Testing/validation/documentation
- Explosion Proof Devices and Construction
- Intrinsically Safe Barriers
Visit our Public Slack Channel and join the conversation in between episodes!
Special thanks to whixr over at Tymkrs for the intro and outro!
About The Hosts
Parker Dillmann is MacroFab's Co-Founder, and Lead ECE with backgrounds in Embedded System Design, and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. He also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components. Parker graduated with a BS in Electrical and Computer Engineering from the University of Texas.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Host 2 00:01
Last week we launched an updated website. Go check it out. Our new site shows how engineers and purchasing teams can easily manufacture their electronics products. With macro fab. We have pages dedicated to the needs of web and purchasing teams with a brand new capability section that outlines all of our services from PCB assembly to full product assembly and testing. Lastly, we have a section that talks about our macro fab factory network, and how that allows you to get to market faster, visit macro lab.com to learn more.
Host 1 00:46
Welcome to the macro fab engineering podcast. I'm your guest, Jeff garoun.
Host 2 00:51
And we are your hosts Parker Dolman and Steven Craig. This is episode 189.
Host 3 00:58
Jeff karoun is a Y Combinator alumni from the winter 2016 batch and currently serves as CEO of flow command. At flow command. He is responsible for our all hardware design, firmware development and operations. Jeff spent his early career as a production engineer and lead completion engineer at Oxy USA. He has been on why combinators alumni application review committee for the last three cohorts has raised three rounds of financing and manages a staff of 17 engineers. Jeff is also a member of incubate USC, where he has assisted in writing undergraduate coursework on accelerators and incubators.
Host 2 01:42
Thank you, Jeff, for opening our podcast and coming on to speak today.
Host 1 01:48
Yeah, glad to be with you guys.
Host 2 01:51
So my first question, though, is what is Y Combinator for people that don't know what that is?
Host 1 01:57
Sure. So Y Combinator is, is a startup accelerator. That's essentially a three month long program, based in in San Francisco in the value that takes early stage companies and helps make them better. So you've probably heard of a few bigger name companies that have come through there like Stripe, Airbnb, Dropbox, for example. And it's a it's a fantastic program, that that I would encourage every startup to take a look at, if they haven't heard of it before.
Host 2 02:32
And so what is flow command?
Host 1 02:34
So flow command makes sensors and software for oil and gas companies. So essentially, these companies have a huge amount of assets that are kind of spread out in the middle of nowhere, and they're very difficult to access. And so we give them very lightweight sensors that are cloud connected, that allow them to see kind of everything that's happening on those assets from their phone or computer or or anywhere they have access to the internet.
Host 3 03:04
And so you've been you've led the development of all the the actual products at flow command, right?
Host 1 03:10
Correct. Yes, it's a, it's been quite a journey for sure, we still have some some dinosaur bones of our of our old products that we leave on the shelf. And so it's really fun to go back and look at those devices compared to the, you know, the types of things that were that we're putting on the market now and that we're developing for future use. But industrial hardware is is really hard. And so we try to have as much fun as we can doing it.
Host 2 03:42
So I guess that leads right into our conversation today, which is industrial product design and testing. So what makes a product and industrial device besides just saying it's industrial device? Yeah, I
Host 1 03:52
Think industrial is a descriptor that largely relates to a device's purpose. And so when you think about, you know, what I think about as kind of the quintessential consumer electronics as a smartwatch, because it's very lightweight, people are fairly familiar with them at this point. And so this is a consumer electronics that that takes a very different form factor than industrial devices, it has a very different purpose. And so it's in some ways, it's kind of the antithesis of what we think of as industrial devices. And so industrial devices generally serve a purpose that relates to manufacture or process things like this, that that are largely being sold for the most part to to other businesses instead of consumers. So the end use and the end buyer are very different types of profiles.
Host 3 04:49
Also think there's there's a bit of a concept behind the word industrial that that kind of relates to the word robustness, you know, industrial means like more bullet proof in a way,
Host 1 05:01
For sure, very much. So it's in the industrial hardware space, though it may be quite a bit less sexy than than a smartwatches. It's actually, you know, we're afforded some things that that the consumer space is not, we have space, which is really great when developing products, because it allows us to be able to QA QC thing, you know, to quality check things easier to assemble them easier. And talking about, like, the overall volume of the product. Yeah, exactly, exactly. So you know, the sexiness of the of the enclosure, for example, you know, matters quite a bit less in a beautiful consumer device versus, you know, just a generic off the shelf enclosure that might go into a manufacturing plant. And so, you know, the way that that design of hardware gets approached in industrial places, is very different in the sense that we largely start from the exterior, and say, you know, this is what our box needs to be to, to, for, you know, to fulfill some specific requirement. And then we packed basically, as much as we possibly can, you know, into those boxes, to make them as useful as possible. And so, you know, when you think of satellites, and other, you know, space related equipment that's going up now, you know, these are types of devices that that also take the industrial approach in the sense that they're allocated XY and Z dimensions, and they fit as much as they can into that box.
Host 3 06:38
Hmm, that's interesting. So when when developing a new product, or even coming up with a revision to a current product? Do you tend to start with a function list? And then develop around that? Or are you saying like, here's my main singular function, I have a box that fits that, how else can I fill this out?
Host 1 06:58
Yeah, for sure function takes the precedent there, for a huge part of the process, particularly early on in the development, and it changes a little bit maybe later on, as you're kind of refining things. But for the most part, when, particularly as an early stage startup, and developing industrial hardware, our only purpose is to fulfill some specific function or perform some specific function, whether that is something on the digital side, you know, communication related, that's one piece, whether it's a sensing piece, on the analog side, that's another kind of totally different category. But but very much, what we think about when it comes to success or performance metric is is more driven by function than it is, you know, making the device one inch smaller. Whereas, you know, if you think about an iPhone or something like that, you know, one millimeter is an enormous deal there. And for us, I don't particularly care about one meter, one millimeter in any direction. So it makes things you know, actually really nice and neat, to be able to have that flexibility.
Host 3 08:14
For sure, for sure. And you know, going back to the the kind of the word I use earlier robustness like that, can you also drive a significant portion of it, I mean, you can have in an industrial device, you it is possible to have a bit of open and blank space in there, if that means that you are achieving that robustness that are expected.
Host 1 08:35
In an ideal world, that would be the case. And when, you know, when I think about designing devices, and how our economics, you know, internally work, it becomes favorable to have devices that last basically, as long as possible. However, we all kind of know that there are plenty of businesses out there that that want you to be re buying equipment every X amount of years. And, you know, they're they're a good handful of very big companies. This is this is really prevalent in the airline industry, actually, where for the most part, the equipment, you know, the engines for an airplane are almost given away basically at cost or maybe at some small loss. And and most of the revenue of those of those companies is actually generated in servicing and repairs and things like that. So in an ideal case, yes, I would for sure agree, you know, that we build really robust things that last forever. And anything kind of that deviates from that plan is largely just driven by business decisions, not necessarily by the best hardware engineers and engineering that's out there.
Host 2 09:49
So you're going off the light, your function dictates your form, basically like how big devices and stuff like that. It kind of plays into Stephens current project, he's building a really fancy fermenter for beer brewing. Cool. And the old, it's a ginormous box. And the only thing driving the box size is what, Steven?
Host 3 10:13
So okay, well, it's two things, but they're both connected, I very much wanted all of the input outputs on this device to be on one side of a box. And with the size of all of the connectors, and everything in in a line, like that just forces the size of the box to be large in one direction. Well, all the boxes that are available with that size, in one in one axis are large, and the other axis is too. So yeah, that you know, but those are the decisions you make when designing these kinds of things, I wanted everything to be on one side of the box, if I was fine with it not being on one side of the box, I could probably make this half as large, you know,
Host 1 10:56
For sure. And, you know, this is a really interesting point, actually, because the the types of issues that I think people encounter in their home are hobbyist or, you know, hardware hacker related problems are almost exactly the same as the problems that we encounter, you know, at low volumes, particularly in industrial hardware in that for some cases, in prototyping, it's totally acceptable to go to 3d prints for certain things. But once you cross some specific volume threshold, then that doesn't really become very feasible, if you have to go outdoors, 3d printing, you know, gets a little bit less attractive. And if you actually have to perform some mechanical functions, it also becomes less attractive. And so you end up getting forced to this kind of next stage of prototyping, so to speak, when you go from like a homebrew project, you know, and transitioning that into an actual product on the market. And so the next step of that, to me is essentially going to a lot of these, you know, off the shelf manufacturers for certain things. And so, whereas we may have, you know, to keep using your example, Steven, whereas we may have, you know, put some two by fours together, or something like that, to make our initial enclosure, just to see kind of what the dimensions look like, the next step is, is to go and look at enclosure manufacturers, and see, you know, what those companies have off the shelf, and there are, you know, at least at this moment, four of them off the top of my head that are very large in the United States that do semi custom work, which means, you know, they have kind of a blank slate, and you can choose where you want your CNC holes, if you want countersinks in there, if you want screen printing, and so like this is very much the the next step of transitioning a product, you know, from kind of a homebrew no pun intended into something that actually hits the market is to basically go and look for these companies that have a huge amount of manufacturing capacity, that that, that you kind of don't have access to it at the scale that you know, sub 1000 unit kind of scale. And then we've had a huge amount of success, working with some of these semi custom suppliers and manufacturers.
Host 2 13:18
So speaking of prototyping, how do you start prototyping an industrial piece of hardware? I know that's like very broad, because industrial hardware is like, the scope of that is ginormous. But I guess let's talk more about like, industrial hardware in terms of IoT devices, which is what flow command does, yeah.
Host 1 13:38
So we generally start this process with a basic block diagram. And so we have to understand what the functional components are, that need to exist in the system to fulfill whatever requirements exist. And so we then take that and basically take each block of that functional diagram and figure out some very easily accessible off the shelf way to start doing something with it. Now, we know that whatever we're going to end up doing to to get started is not going to be the you know the final form, or chip or whatever we end up using. But But we've got to start doing something because there's only so much you can actually learn from looking at a datasheet of anything, right, you've got to get your hands dirty on it. And so going and prototyping and starting to do something as soon as possible is really critical. So generally, there are a lot of breakout boards and kind of Arduino related peripherals that exist and are very, you know, off the shelf you can buy them and ship them overnight, anywhere in the US. That is by far the easiest way to get started in transitioning a block diagram on a sheet of paper into a physical The thing that actually works. And so, you know, I always encourage people to go and you don't even really have to scour the internet to find a gyroscope breakout board that already has an off the shelf library from from a chip that you can buy from any of the major US distributors that will allow you to just kind of crank something out very quickly in the Arduino IDE and see how it works and be able to move it around and look at a Serial print on your computer. So that's very much step one is do something as quickly as possible and really utilize off the shelf resources kind of to their to their maximum extent, you do not want to, conversely, go and hire someone to draw out a custom schematic and a custom footprint for a custom circuit board and make custom silicone that you'll end up going and putting on that ship. You can go as overboard as you possibly could think about for something like this. But it's really important to, to iterate very quickly, you know, one of our in the startup community, you know, one of our biggest challenges in being a hardware company is the iteration cycle. If you're a software company, you can spin up code and push it almost instantaneously. And for us, you know, having to live in the physical space. It's our goal to try to emulate those quick iteration cycles as much as we can. And though it's almost impossible that we'll ever reach there, it's it's the right thing to try and strive for.
Host 2 16:37
Yeah, it's the same in software startups is move fast break things, right. But hardware, it's move fast break things and then soldered a lot of green wire.
Host 1 16:47
I can feel that big way. Yep. So
Host 2 16:51
Actually, the, it's funny how you bring up Arduino stuff, like when I'm looking for like a chip to try out. And I find the one I want, I'll actually will search that and said that weren't like like 18 Mega, let's just say example 18 Mega 328 P, and that's a microcontroller. But that and then plus like breakouts, gets less hits then like that chip number plus, like Arduino, like plugging that into eBay is probably the quickest way to find like a circuit board. That's already got it on there. Yep,
Host 1 17:22
For sure. And I think, you know, the Arduino community is extremely extensive and comprehensive and responsive. And anyone that is trying to learn any type of firmware development or hardware development, there's really no other place to start than in the Arduino community. You know, I'll tell you the first devices that we shipped to two big companies still had at Mega 2560s in them, and you cannot be afraid of that chip, because it is so well known. And so well defined, you know, the deficiencies and problems there. And frankly, most projects that that people are working on that ship is perfectly suitable or even you know, an Uno style chip is is totally suitable. I think the you know, in the industrial space, there is a tendency to, to probably move towards overcomplicated chips or rather, you know, get away from the Arduino community because it feels like it's very hobbyist. But frankly, by doing that, you can add so much more complexity that's unnecessary to your project. And frankly, you can do almost anything that any industrial company is going to, is going to have a need for in a very, very simple chipset that you've got boot loaders and libraries for off the shelf. And the more custom stuff you have to do in general, the slower you're going to be. And it's much much more important to get equipment into these customers hands to understand what's actually important to them, than for you to be sitting and spinning your wheels and trying to think about what you believe to be what's important to them. Market Fit. Exactly, yep, get it as quickly as possible. And so an MVP with an Arduino chip is great. It's a great place to start.
Host 3 19:22
Well okay, so that does bring up some interesting questions because I'm looking a little bit down on notes. I'm kind of cheating here. But I see that I see that some of your products have some pretty hefty certifications on them some stuff that brings up some nightmares from the past some class one div one stuff now I don't know the answer to this but if i My gut feel is if you ever showed up to to a testing house with a Arduino Uno inside of a box right no way that's going to pass any kind of testing so you know I agree like it's great to to get you know, kind of scratched down on paper get us I get A Sketch together, right? With uh with Arduino. That's fantastic. But eventually you do have to migrate that to an actual board that is designed to be able to pass tests, right?
Host 1 20:11
Hopefully, your company grows to the point where that will make economic sense for sure. That's, that's definitely the goal. And one of the nice things about kind of building off of this platform is that that transition is actually generally a very seamless one. You know, with respect to bringing very simple, you know, off the shelf chipsets into a certification environment, it kind of depends on what you are trying to get certified or what you need to get certified. I take that as a bit of a challenge, I actually would be curious if I could get an Arduino Uno certified, inside some kind of of enclosure, that'd be a really interesting challenge to go after. And I bet that it could be done. It would be almost definitely a first for any of the of the certifying agencies. However,
Host 3 21:06
I bet you they have seen it, though, I bet you they have,
Host 1 21:09
I would bet so and I bet that a substantial, substantially larger volume of you know, some of the vacuum cleaners and other you know, consumer grade devices that that we use every day, do contain AVR pic related type microcontrollers in them. And, you know, these, all of these functions started with, with a simple board with a simple board and a breakout board for the most part. And they're very accessible for sure. And there are certified I should say,
Host 2 21:45
I've actually written an article about like, if you're prototyping on Arduino, and then move into a product like what parts of the Arduino you actually need? And how do you transfer that over to your, to your new schematic and layout? So
Host 1 21:58
Yep, for sure. Yeah, I
Host 2 22:01
Think Stephens thinking like when you build your actual industrial device, like the Arduino Uno, so to speak, is still in there. Whereas usually, you do embed Mike the microcontroller onto the PCB.
Host 3 22:14
Well, and that no, that's what I was getting that, you know, eventually, you have to get away from the UNO form factor, because you would want to eventually get your product certified. And I really doubt that that, you know, inside of a box would get certified. I don't know the answer to that, like I said, but But you know, and it may be something where you put an exact 328 P down on on your board, and run it effectively, like an Arduino and there's nothing particularly wrong with that. But you also might port to something else that is, you know, cheaper or is a little bit more stripped down and directly related to what you're going for, you know, there's a lot of options in that. I guess I was just kind of going at the idea that well, let me get let me bring up an interesting story real quick. I actually visited a satellite manufacturer here in Colorado, and I swear to God, there was an Arduino Uno, you know, screwed down into the chassis with, with connectors on it. And and it did get sent to space, and it did its thing, you know. So like, it's not like impossible. And the funny thing is all the designers in there were all like, they were the kind of guys that you would expect would fit the senior engineer, quality level. And when I put all of them were like, we needed something fast. We needed something that would work. It didn't have to be like, it's not ripping at a gigahertz. So it fit the bill, you know,
Host 1 23:39
Right, for sure. i It's funny, you mentioned that because this first device of RSR, you know, Gen zero, which is maybe even a generous term device that we that we put into the field, I actually remember about two days leading up to Demo Day at Y Combinator sitting on the floor of YC headquarters mocking up in SolidWorks our enclosure which is an off the shelf enclosure that I ended up spray painting about 50 of so I'm very good at spray painting now. But mocking up you know the interior of that enclosure, which was essentially a metal plate that also came off the shelf from from the enclosure company and on that metal plate where some rubber feet and screw down into that was a mega with a board sandwiched on top of it. Then we had a modem wired together you know with just basic screw terminals kind of right next to it. And so, you know, this is kind of the step in between, you know, using a just strictly a mega or you know, and going with a fully custom board is you know, you can you can start by just using, you know, some maybe larger enclosures and basically just wiring these things together and and making them feel like kind of more of a permanent connection between them. And so those were the devices that that we put into the field. First were just, you know, a handful of, you know, Maga and a few breakout boards basically just slapped onto a metal plate, latch, the enclosure closed, and that was it. Now, the longevity of those wasn't great, but, but they did work for enough time that that we could learn, you know, where they started to fail, and move from there
Host 2 25:28
At the company previous to macro fab at Dynamic reception where I worked. Or one of the first products was a it's a motion control for cameras. And it was a consumer device. But it had an Arduino Uno in it with a shield and an enclosure. And we did 1000s of those a year. And it I mean, they got dropped in swamps and still worked.
Host 1 25:52
That's amazing. happens much more often than then people think, yeah, it's just, it's just too easy not to, you know, you kind of have to find a reason to move away from them.
Host 2 26:03
Yeah, the only reason we ended up moving away was it's, it ends up being a packaging problem, because it's because you have two PCBs now, basically, and there's the header headers that are separated and everything. And so one of my first jobs was to take that product and make it to a single board solution, so you can make it thinner. And it also drops the bomb cost by 25 bucks. So you don't have to buy an Arduino anymore. And you can charge the same amount, because it's new and improved.
Host 1 26:32
So this is this is actually a really interesting point that you bring up that we very much actually felt at our company, too, which is you had your hands forced into basically moving to a custom board. In other words, what you are trying to accomplish, you you basically could not do if you stayed with that same setup, where you had, you know, off the shelf, you know, boards that you were slapping together, the same exact thing happened for us just with kind of a different failure mode. And so what we had, you know, going off of that same device where we had a few breakout boards that were just wired together, we had vibration problems. And so we were going in installing this on equipment that was vibrating to an insane degree. And what would end up happening is, wires would just pop out all over the place. And so we've got, you know, a full inbox of angry customers basically saying, you know, Hey, guys, nothing's working, what are we going to do here? And you know, that that was the forcing function for us that that made us say, Okay, we've got to go and figure out how to draw up a custom circuit board, which we had never done before. Basically, everything was completely off the shelf up until that moment, I want to
Host 2 27:43
Just use a bunch of hot glue. I wonder if hot glue is class one div one compliant?
Host 1 27:53
As long as you meet your IP, but you could do it. Yeah.
Host 2 27:57
So speaking about that is so the regulations and certifications for industrial devices? Yeah. How do you make that work when you are first starting out? Because I mean, you started out and it's just like you and a couple people. If that, and you got to get these tests and certifications done. And they they do cost quite a bit of money.
Host 1 28:19
Yeah, this is a really hard problem. And actually, I talk to companies on a fairly regular basis that say, When should I do this? How should I do this? What do I need. And the range of devices that exist out there and the types of certifications is enormous, there's no way a single person could possibly know, you know, even in a single category, kind of what's needed for all of the all of the stuff everywhere. To compound that the regulations are different in different countries, if you want to be in Canada versus the US, you have different regulations in Europe, and in Asia, you have different regulations as well, which are which are mostly the same and some of the standards kind of share most of their principles, but you may have different language requirements, different font and text requirements, you know, things like that. When when we think about how this process started, and when I talk to companies about this, I generally tell them that what they should do is have someone forced them into this process. And what I mean by that is, if you think that your customer is going to require something that doesn't necessarily mean that that customer will require something and so if you get started in oil and gas, you know, hazardous location explosive gas environment with some with some device, and go and sell it on the market. You know, if you don't even bother with the certification, you may actually find out that most of your customers want your sensor device in non hazardous locations, and you will have wasted a huge amount of resources. Basically down the drain for nothing. And so when you go and you know, as an early stage company, and have conversations with big industries, it's really, really important to let those customers drive your decisions. And so if they say to you, I will not install your equipment, unless you are hazardous location certified, it's probably time to go and do that. If it turns out that they say, Hey, we really like your stuff, and actually, we have some, you know, general application stuff, you know, purpose with with your, your equipment, that's much, much better for you, because you won't even have to go down this road. Now, going off of the assumption that you will be forced into needing certifications at some point. This is an industry that tends to be generally geared towards bigger companies with lots of resources, and largely favors specific testing labs that have tended to either write the standards historically, or have the marketing dollars to make it seem like they have basically, you know, kind of own the market. It depends on what you want to do. So the big categories here are electrical, hazardous location, and radio related FCC kind of related things. Now, there are plenty of other ones that I'm sure I'm not familiar with. But those are the ones that most hobbyist hardware hacker type companies are going to be encountering. And what what I think is the is the best approach here is to really lean heavily on some of the experts in the space. And so what we did was was twofold. The first thing was, we found a lab that we really liked working with, that we had really good communication with, that was willing to be a little bit generous with us with respect to giving us some guidance that many labs are not willing to necessarily give. So that's one piece is is find, find a lab that's going to be a good relationship partner for you. And the other is, you almost definitely have to find someone that has done this before. Because unfortunately, the standards that are out there are not put in a form that you can read and understand and say, Okay, here's what my footprint and component values need to be someone that's done it before. They've got the calculations, they know what the distances need to be between components, they can do this kind of thing very quickly. And so your freelance marketplaces online are an excellent source for finding people. If you don't have anyone in your community that can that can help with those topics.
Host 2 32:51
So finding a base a, like engineering consultant, that is in this field is your best bet, then,
Host 1 32:58
Yeah, I mean, it's it's the same as you know, getting a legal document done, you know, your one option is to go and learn it from scratch and go to law school. The other is to, you know, you may pay a little bit more. But if you can hire someone for a couple hours, at a reasonable fee, we have found great success with going down that route. And and often the types of freelancers that are that are assisting and consulting on those projects tend to get really interested in and, you know, will often put in a bit of extra work to kind of learn about you and your company,
Host 2 33:29
Or come work for you. Yeah, exactly.
Host 1 33:31
If you've got the workload for it, I would highly recommend it. It's a it's a it's a very opaque trade, for sure.
Host 2 33:41
So I have a question, actually, yes, I was in the oilfield a lot. Right after college. And so what is class one, Div one slash two? What does that mean?
Host 1 33:52
So the two different divisions relate to essentially how explosive and how frequent explosive gases present in a specific environment. And so if you think about the gas that comes out of an air compressor, at you know, you're filling up your tires at a gas station, there's a bunch of gas that's right, kind of near that thing, you know, you can, you can, you can, you can smell the fumes, you know, when you're right by there. If you move 10 feet away, you're probably not smelling any fumes, right. And so the same type of calculation is essentially done in the oilfield, where we say if we're nearby, some equipment that might be emanating fumes, or might be in a plant that has some particular piece of equipment that might have gases that flow through it or something like that. Depending on how frequently and at what concentration you may be classified in, in a class one or class two type environment in a tux, there's a class zero as well. which is which is equivalent to class one div one in the US type scale. And so in the oilfield specifically, if you're within five feet of a wellhead or an open port in a tank, things like this, you generally are in a in a class one division one environment. If you move further away, you're going to be classified in a division two type environment which I believe is for the subsequent 15 feet. And then beyond that your your kind of general application which means you don't necessarily need a hazardous location type certification.
Host 2 35:39
So dumb question, do they? Is there a class one div one like cigarette lighter?
Host 1 35:46
There is class one div one, almost everything. It's shocking how much stuff that's that's out there. If you go and look at a class one div one power drill, something you could buy from you know, your big boxes here for maybe 40 bucks or less probably now, depending on which big box you go to. The class one div one power drills are probably about 100 to 1,000x, the price of those things, and the amount of electronics that they're actually packing in, there's almost nothing different depending on their, their, their suppression method. But yeah, that I don't know that they're specifically, you know, lighters or anything like that. But I would tend to doubt it.
Host 3 36:31
You know, at a previous company I worked at we did a lot of class one div one testing and design work. And basically any product that was class one div one you could purchase as a class one div one. Or you could just purchase it as a non mark one, the exact same unit, we just lasered on the label and added, you know, 110% cost adder on there, and you're just paying for the fact that you would that we put the label on there?
Host 2 36:58
Well, it's not just the label, you're also like ensuring that it's been tested. So it's like if Well, all of them were tested. But yes, yeah. But I'm saying is if your device caused a problem, it's on you. That's why you have that markup. Right. Right. It's your insurance.
Host 3 37:17
You know, also above and beyond the the class one div one stuff, there's actually also groups that are in there. And the groups beyond that define what kind of atmosphere they're in and what kind of gas they can, like, acetylene or hydrogen or something like that.
Host 2 37:33
Taco Bell bathroom.
Host 1 37:37
I don't know if there's a specific stringent testing amazing. Yep, so there's a you have, you have a group on top of that as as well as, you know, your your key class, which is effectively, you know, temperature related. And you can have multiple certifications on a specific device, almost every device has multiple certifications, you can have multiple groups on devices. And so you know, often we'll see, class one div one Groups A through D. And you know, these are very common things. And when you go and send things off for certification, you generally want to package them all in one, one certification pass, because it can be very expensive to go and recertify things.
Host 2 38:20
So speaking of Taco Bell bathrooms, explosion proof devices. So what what is an explosion proof device? Besides the fact that it won't, it won't explode? So like, my cup here is an explosion proof device, right? Yeah.
Host 1 38:38
Yeah, I don't know that. It would be interesting to, to, to make a whole Taco Bell bathroom into an explosion proof device, that would be that would fit some of your most recent themes of you know, useless devices, maybe actually, that could be very, very useful. Anyway, so, the way to think about this is kind of from the outside and what I mean by that is, you know, you have to decide firstly, what regulation is required. And so, in our case, you know, this is environmental related. And so, so, if you are in a hazardous location, you need to fulfill a specific set of requirements. You know, that that is they're almost definitely written by UL and those requirements essentially say that you will not do something in some specific conditions. And so when it comes to hazardous location, it's largely ignition related. For the most part, though, you know, there are other provisions in there that you have to fill including Ingress Protection. In order to fulfill anything related to class one division one, you do need Ingress Protection as well. And so, there are several methods that you can use to fulfill their requirements of division one or division two requirements. And so one of those methods is explosion proofing. And so when you see, you know, in big industrial devices, you know, they have these very robust, you know, thick metal enclosures, those generally are using explosion proofing, as the method to fulfill those requirements. Explosion proofing is physical containment, it means that actually a spark can ignite. But that spark and ignition must be contained physically within the enclosure. So, that's one big category that gets used, it's a very popular method, there are, there's an abundance of off the shelf explosion proof devices that exist, that anyone can buy, they're not terribly expensive. But in general, you know, these are, these are really simple ways. For low cost that you can, you can try to fulfill the requirements of division one, another method that can be used as is electrical containment is basically, you know, energy containment. And what that means is, you have designed a circuit, so that any lightning strike, for example, you know, any surge that comes into your electronics, will basically be moved to a location that will not cause a spark to ignite at all, or the spark to exist at all. And so this is called intrinsically safe, which means kind of, it doesn't really matter very much what happens on the outside of the device, any of that can be contained electrically based on the design of the device. And so the off the shelf components that usually fulfill this are zener barriers, you can look for intrinsically safe barriers that exist, you know, largely, you know, for DIN rails and things like that, the circuits themselves, you can, you can usually find a basic schematic online, it's about five components that you can place as an SMD on your circuit board. And that will do it. So you're talking design work, plus maybe $1 worth of SMD components. And, and you can fulfill intrinsically safe, you know, in most circumstances. And so those are, you know, two ways that you can fulfill those requirements. There are a few others that are that are far less common. But you know, pick the one generally that that suits you best. One is more expensive, on a per unit basis, and has a lower upfront cost. That's the explosionproof. explosion proof is more upfront costs for design. But you know, over the long run, if you expect to do anything with any kind of reasonable volume, intrinsically safe is really the favorite method, in my opinion.
Host 3 42:51
Yeah, also, with intrinsically safe, you will and ATEX. And a lot of the other bodies will also have a two really significant numbers that apply to whatever your industry is, or whatever the application is, and that is the maximum amount of capacitance that the whole circuit will have. So you can't just make some random design with a bazillion micro ferrets on the front end. But they also will dictate what's the maximum energy that can be dissipated at any point in time. And you have to prove that your circuit, you know, if it were to discharge violently, what is what is the total joules that would be released by your circuit at any one point in time. And if it's beneath their threshold level, then it it passes intrinsically safe. A fun little exercise you can do even just like a little game in your head and take any active circuit, or any active element in your circuit, processor, regulator, whatever. And just pretend like it stopped being whatever it was and turns into a giant blob of solder. What would your circuit do? Like if your processor connected every single pin to every other single pin on the processor? What would the circuit as a whole do? If it dissipates beneath their threshold of energy, then you're fine.
Host 1 44:15
Yep. And this is more or less the type of analysis that the the testing labs go through, you know what, what, there's kind of a secondary branch in in intrinsically safe if people go that route that Stephen you kind of mentioned. One is that the the labs will certify you specifically to whatever you're connecting to. And so that certification will say This fulfills the requirements. If these two specific things are connected, you know, for the most part that other thing is already certified. The other is, is called entity parameters, which is kind of what you were mentioning as well, which is where the certification more or less says, so long as you are Within the specific ranges of resistance, capacitance inductance, then you're good to hook up basically, whatever you want, so long as that other thing is, is also certified. And so you know, if you need that kind of flexibility that entity parameters is, is a really nice route to go. If you've got, you know, a simple product that's going to be hooked up to the same kind of thing and high volume, I think that the fixed route is actually is a simpler process when it comes to certification. You know, all of these kind of three options, explosion proof, and the two intrinsically safe options, the entity parameter intrinsically safe option is definitely the most sensitive when it comes to certification, and the one that is probably most likely to result in, in revisions that are needed in order to fulfill and pass. But, you know, our business dictated it. And, and so that's the route we've gotten kind of across the board. And that's largely related to, you know, needing a whole bunch of different types of sensors, that we would need to hook up here. And so we basically said, you know, we're going to make this investment kind of upfront that will allow us to make life much, much easier down the line. And, and not be bogged down and worried about being able to get new customers and make sure that our certifications are all in line, we kind of said let's, let's net this as quickly as possible and go the entity parameter route,
Host 3 46:27
For sure. So how about intrinsically safe barriers, and, and putting those in your installation?
Host 1 46:35
You know, this is a size dependent type, type setup, I think, in the oil field, in particular, and in a lot of industrial applications, you know, these are, you know, the kind of off the shelf intrinsically safe barriers are a really nice and simple way to go. You can buy them, you know, with, you know, your your big electronics distributors in the US. And, you know, if you, if you need to do something kind of at low volume, I think it's a really good option, compared to what it would actually take to go and design your own and get it, you know, built into a custom PCB, I would imagine your breakeven is, is pretty low, because I think that the off the shelf intrinsically, intrinsically safe barriers are pretty expensive, you're probably talking about maybe a 20 unit payback kind of thing. But, but certainly for getting something in your hands and testing it and making sure that you kind of understand how the electricity is flowing, and how everything's actually working there. Those Those things are, you know, are a good starting point, I think,
Host 3 47:43
For sure, for sure. And anytime you see, you know, I product that has, you know, 1520 marks on it, the mark of the beast, you can you can just immediately say, Well, this is going to be a lot this is going to cost me
Host 1 47:58
Yep, for sure. For sure. Certifications are hard, they're expensive. Do update them if you need to, that's that's really what I would say is, you know, we're only now you know, starting to go back and and you know, as we're iterating on our device, to make changes, you know, as particularly there are new battery chemistry is that are coming out, you know, every single thing that you change once you've certified a device. If assuming that it is a critical component, anything you change at all, warrants a recertification. And so you know, you're you're in the four figures, basically, at a minimum, anytime you want to touch something like that.
Host 3 48:41
Yeah, you have to be really, really confident that, first of all your product is where it's at when you've certified it. And you also have to be extremely confident if you're ever going to make a change, because it doesn't happen fast. And it cost a lot of money.
Host 2 48:56
I would imagine making sure your supply chain is pretty robust for however long you're going to sell that product for. Because if you have to do a part substitute that sounds pretty awful.
Host 1 49:06
Yeah, for sure. I mean, you get that end of life notification on one of your critical components. And that's your strength. Timer. Yep. Yep, for sure. That's true. Yeah. And, you know, the other thing that I'll mention on this front is, there are I think there's over 500 laboratories that are that are out there in the world of which, you know, there there are a lot in the United States. You know, when we went and started to bid this out, you know, we went to you know, your your big name brands, and we went to some of the smaller ones as well. And, you know, there are some industries that that generally really favor some of the big name labs. And when I say favor in this context, really what I mean is, it's very difficult to sell products, unless you have what you know A two letter kind of sticker on your product. And in cases like that, you know, you've gotta like sales drive, right? So you basically have to go to those labs and, and kind of suck it up, so to speak. However, you know, we found so much more flexibility than we expected in, in our decision to use, you know, one of the smaller labs, that we've been really pleasantly surprised. And so, you know, it, I think it's good to understand, you know, kind of the baseline for people that are out there that are thinking about going for hazardous location certifications, or, or any kind of certification to go and get your baseline, you know, get your feet in the water with, with bids, you know, at some of the bigger labs, and then, you know, go a little bit further and take a look at some of the other labs that are that are out there in the US, it's really nice to find something locally. And so, you know, I've done a handful of visits to our lab that we use primarily, which is just outside of Atlanta. But if you've got something locally, try and find them, go visit them go sit down and talk to them and get a tour of the shop. Because you know, you'll find, as with anything face to face is a lot easier than then, you know, just kind of sticking to email communication. And there are, there are a lot of things, you know, behind the scenes that those companies can do to set you in the right direction. And so, you know, on top of the technical piece, the cost pieces is a big thing, too. And so, when you're looking at your big names versus your your smaller names, you know, you're not just talking about five to 10% differences and certification costs, you may be looking at kind of a two or 3x kind of thing. So it really is a big difference between the cost there and for what you get. Some companies may not find that it's worth it. So it's a worthy exercise to take a look out there.
Host 3 51:58
For sure. And if you know you need that, or if you already have customers ringing the bell saying, I'll buy your thing, if it has whatever stamp on it, do that early, do it early in the process.
Host 1 52:10
Yep, for sure. This is this goes back to you know, get stuff in front of customers ASAP, because you have no idea before you sit in that room, if they're gonna say, I need this lab, or if you can go to a smaller lab, or if you need a lab. And so you know, just speed up that iteration cycle as much as possible, talk to customers as much as possible. You know, we drill this in all the time. But, you know, what we talked about internally, you know, at flow command is, you know, how can we make as few decisions as possible. And really, what that means is, you know, how can we let customers drive what we're doing to the maximum possible extent. And so we try to kind of lean on them as as much as as is reasonable to kind of tell us and give us feedback on where we should be going, not only from a product standpoint, but from, you know, an r&d standpoint, from an operation standpoint, a customer service standpoint, you know, those are the people that are going to give you probably the best and most honest feedback versus, you know, just kind of sitting in an echo chamber inside your company.
Host 2 53:17
So, thank you, Jeff, for coming onto our podcast. Yeah, absolutely. Steven, do you have any other questions?
Host 3 53:23
No, no, I think that was really fun. Really appreciate you coming on, Jeff. And this isn't something that we normally talk about. We don't we don't have a lot of guests that have had a lot of flavor in this world. So appreciate you coming in and enlightening us about it. Yeah, absolutely. Glad to be here.
Host 2 53:42
And so Jeff, where can people find you and find out more about flow command?
Host 1 53:48
Yep, so our website is flow command.com FLOWCOM M A N D? My email is Jeff at full command je ff the American spelling. And yeah, I'm easily accessible in both of those places.
Host 2 54:05
Everyone just sent tons of emails Jeff saying thank you for being on the podcast.
Host 1 54:11
It's good. I'm not an inbox zero person. So it's, it doesn't affect me very much.
Host 3 54:20
Great. Would that would you like to sign us
Host 1 54:22
Out? Absolutely. That was the macro fab engineering podcast. I was your guest Jeff garoun.
Host 2 54:29
And we're your hosts Parker Dolman and Steven Craig. Later everyone take it easy
Host 2 54:42
Thank you, yes, you our listener for downloading our show. If you have a cool idea, project or topic let Stephen and I know Tweet us at Mac fat at Longhorn engineer or at analog E and G or email us at podcast at Mac fab.com Also check out our Slack channel you get there by going to macro app.com/ slack it will redirect to the invite page bam you're in. If you're not subscribed to the podcast yet, click that subscribe button. That way you get the latest episodes right when it releases and please review us wherever you listen as a helps the show stay visible and helps new listeners find us
Transcribed by https://otter.ai