MEP EP#209: A Bad User Experience

Podcast Transcript

Host 3     00:11

Welcome to the macro fab engineering podcast. We are your host, Stephen Gregg and Parker Dolman. This is episode 209. Oh, man 209, almost 210 It's almost 300. Just around the corner, actually, the The 100 The 100 episodes happen a lot faster than you think. Well, it's also depending on your timescale, if you think of, like, let's say, we're gonna do 1000 episodes. Yeah, it's pretty close.

Host 1     00:44

It's like that theory, that when you're a kid, days seem longer, because each day is a greater percentage of your life. Whereas when you're older, like, they they're such a small percentage of your life that they just seem to go faster.

Host 3     00:59

So these episodes are starting to go faster and faster. All right, what's up? Barger? Well, one thing I wish was going faster. The pennant tar project

Host 1     01:12

Seems to be going pretty quick. Yeah, it's gone pretty quick. It's just, I wanted to go faster. I want I wanted to get in production.

Host 3     01:19

But we have have revision to flipping a machine now up in Wisconsin, somewhere in the boonies of Wisconsin. And so it's flipping so that's good. And they been experimenting with a couple of different, like, a couple different software packages that to run the game on the computer. So it's very interesting what they've been coming up with up there. So I can't talk too much more about that part, the high end software part. But it's pretty exciting. Secrets. Yes. Secrets. But um, seems like mission pinball framework is also working pretty well. So that's the one I can talk about. And basically like the one that we're going to sell the platform with. So like, we'll sell when we sell the board. We're like, it works with MPF. That's like the software package we'll go in with. Yeah, cuz that touches the most people, right? Yeah. It's open source and all that good stuff where, but we're also like, the protocol to talk to the board is open. So if you want to design your own thing, you can, right. Yeah. Well, and so this is Rev. Two Rev. Two is working. But you actually have plans already for rev three, right? Yeah. We already have plans for rev three talked about a couple of podcasts ago. But they're all simple. Like, cosmetic changes. Yeah, just cosmetic changes gesture. Well, you were saying what, like 90 degree trace? Oh, yeah. Some traces that didn't didn't chamfer in.

Host 1     02:56

We are changing like one of the connectors a little bit. Right. Right. Right. You talked about a couple times. Cool. So you think rev three, and then you'll go to production.

Host 3     03:04

Yeah, that's what I'm thinking. Cool. And for that is you have to build all the like the supporting hardware now for the board. And so whites are the next thing. And so a couple months ago, I tested out the APA 102 C 260. New part number. And I did a reflow test. And it works perfectly fine in our lead free process. And that's a smart LED as well. So it works great. Because usually with those smart LEDs, you get issues with moisture and the plastic and or the plastic isn't designed for, you know, 250 plus Celsius temperatures. But these seem to work great. And I can buy them directly from the manufacturer, which is like was it neon world? And they come on a real right? Yeah, they come on real? Yeah, neon world there. That's the APA Electronic Code. I would highly recommend them. They're pretty like they let me just buy one real, which I was pretty happy with. And I think it was only 70 bucks or real. So it wasn't too bad. How many come on a real Awani 1000? No, it couldn't be that much. Probably a few 100 I would think. No, it was it was definitely more than that. I think it was 10,000 No, those things are much bigger than that. You get 10,000 resistors on a reel.

Host 1     04:38

Yeah, but that's like, you know, you still got what was it? Is it eight mil spacing.

Host 3     04:45

I mean, is this a big real or a small real

Host 1     04:48

Small real? Let me I'm actually checking my inventory at macro February.

Host 3     04:54

How many do I have, you know, speaking about reels. Just a funny story. We We got some pin headers made for us not that long ago, and totally got bit by a Chinese company, it really sucks. I shouldn't say bit like all of our headers are totally acceptable to us. But we bought a few different variations on pin headers. And this company that we worked with a one of our clients that actually used it, and they recommended them to us. So we got in contact, got samples, and everything was hunky dory and great. And then, so we asked for some drawings for these things. And they made some custom drawings just for us looks nice. And they put the wrong real size on the drawings. So we went out and we bought new magazines or feeders for our machine for that those real sizes, and then the headers came on, and they're not those sizes. Dammit. That's whatever I mean, we're still able to use them. But it's just like, now we have the wrong size. feeders. Yeah. Now you're going to be shopping for parts that fit those feeders? Well, like when it comes down to this kind of stuff, in my opinion, never really hurts to have extra feeders, because you never know what's going to happen. But these were like 44 millimeter. feeders are huge. Yeah, you know, so it's like, the only thing that comes on 44 millimeter. feeders are like ginormous connectors.

Host 1     06:20

Yeah. So they're 1000 are real. Yeah, that makes a lot more sense. Because I have 990 in inventory, I use 10 on the test the Auburn recommend them. So I have got the test boards almost built. And then I'm gonna get those all made up, I'm probably gonna build like 20 test. Board No, 64, we built 64 of them. So we have enough to do one whole game.

Host 3     06:54

And yeah, should be pretty good. And then on the brewery, I got the whole cart welded up. And I welded it with MIG steel. And it's a two inch box tubing, 14 gauge. It's pretty scoop. It's hefty. Yeah, it's pretty good. And I welded it all with all my you know, my welding table. So it's everything was flat, and everything's got 90 degrees. Like when I set it up, and I put it on my concrete floor in my garage. It didn't wobble. That's awesome. I was like, yeah, what's your, what's your method for holding a 90. So for welding holding a 90, where I do is I take whatever the first stick, right? So let's say it's two inch box tubing, and I clamp it to my welding table. And it's that kind of welding table. It's got holes all in it. And so you can use those clamps that kind of stick in the hole, and then you clamp it down. So you do that. And then I have a just a speed square, or speed trying speed squares triangle. I don't remember if it's a triangle shape, but it squares I think it's called as I actually bet I think that's called a square. Okay. Yeah, so I think the brand speed square and the Kleenex of squares, yes, speed. It is a speed square. That is what's called, that's what it was called. And it was I use that. And then I kind of, I get it all squared up. And then I clamp that second piece down. Check it again, it's always moves when you clamp it, right. And so then I take a little, little tiny hammer and kind of just tap it into into place. While it's like kind of clamped down, and then I clamp it even more, check the square again, make sure it's good. And if that's good, then attack all the way around it. So 1234 tax on it. Oh, yeah, that's right. You have to make sure your corner is off the table. Right? So is your corners floating so that way you can tack the bottom of it. So do you take the the external edge or the internal corner first? I usually just do the top. Okay, okay, just because of the polling like, you know, as it cools, well, I just do them all in succession. So I go, boom, boom, boom, boom. So they're kind of cooling, they're kind of cooling a little the same time, and then everything is clamped, right, so then basically, I tackle for and then I just unclamp it and then go on to the next one. Oh, then if I have to have that corner, and the trick is if you have to have that corner rest flat, you have to make sure to make sure you grind that that tack flatter because it will rise up but if you grind it flat, you still won't cut the tackle way through but you got to make sure that's still flat when you're, you know shifting your material around on your welding table. And that I mean that's how I always like to I'd like the welding clamps that are supposed to keep things 90. I've never had luck with those. I call them gimmick clamps. Yeah, those were those were actually pretty nice for screwing like two by fours together in a 90. Yes, way better than Well, I don't know why they suck for welding the design for welding now, that is not good. Yeah, I've watched a handful of YouTube videos on how to get like a good 90. And it seems like everyone has their own like trick. Yeah, they

Host 2     10:29

Mine is just the old school just speed square. Put it all together. It probably takes longer. So it's probably not if you had a fab shop. That's probably not the way to do it. Yeah, but you'd have a fixture probably if you had a fab shop. Right, right. Right. But yours just works. Yeah, it works great. It's just a little slower. We were talking about making some fixtures.

Host 3     10:51

I milled some. Yeah, I've got some really nice. Where's mine? 90s. I don't know. Why don't you get a mill and mill? I had to buy a 20k mill the make $100 worth of squares. 250k get it get it right. Okay. Yeah, yeah, I milled some Gosh, one inch aluminum? Well, four of them actually. My own speed squares that are there. Yeah. 90s. And they're super nice. But yours are designed to clamp to right, which would speed up my technique basically. Yeah, but I haven't actually used them yet. Because I haven't had any welding projects yet that need them. But yeah, the idea is, is to actually put four clamps on them and hold. Hold a really nice 90 And then just let it cool with the clamps on it. And it'll remain 90. Yeah, I mean, these things are really Skookum it's pretty hard to bend a triangle of one inch square solid aluminum. Yeah, yeah. Well, I need to use some aluminum then. Yeah, it took a long time. I mean, I was one off, so I wasn't trying to optimize it, it was just like cut triangle go. Put an inch. Although it was pretty fun because I used an eight millimeter. But there was the first time I use an eight millimeter bit on this. So it was an eight millimeter cut cutting bit that was running at 45,000 rpm. And it just, it just eats a loose haul. Oh, it just it just is a fun. So I got that done. And then I want to put a sock on casters now and it rolls really good. So the next step is to put the counter on it. And I'm going to use I have an old butcher block from IKEA. It's that old Workbench I had at macro fab that really big, long wooden one. Oh yeah. So no one's using that the fab anymore. It's actually in like the compressor room now, like mold taken apart. And so I'm just going to tomorrow, I'm actually going to go to work with my trailer loaded up on the trailer, take it home and cut it up to put it on the top of this brewery. It will live on. So I do that, and then the pots can go on. And then I have to mock up where I want the pumps. Because I want I'm going to cut some brackets out of steel mountain where they need to go. Well, the brackets all that sense of the pumps are where they need to go. And then I have to figure out how to build a hoist. Alright, great fun. Because the the mash tun is like waist level. So trying to pick up you know, probably about 100 pounds of, of grain grain wet grain is above your head is probably not the best thing to do. Yeah, that's a back workout. Yeah. And so we were talking about how we would do this. And so I could build a crane that's attached to the cart. A lot of people just to solve this problem a lot of people just have like a electric hoist or electric winch bolted to their ceiling, which would work but I was planning on doing it outside or under my back patio base that can brew wherever because on a cart now, which would be nice if it's a nice day, do it in the driveway. And it sucks in the garage, stuff like that. And so I the I liked the electric idea and wish I had it was like in one spot, because then I would just put one I just screw electric winch to the ceiling, and just call that good. But I'm thinking about a combination of all you had an idea of build like a little frame like a gantry crane. Yeah, basically like a basically a U shaped frame above everything. Yeah. So you just wheel it over it and you can lift it out. And then we'll get away, which would work. Or you could have that be a permanent part of the cart and just use, like a frame overtop. Yeah, I was always thinking but I need it to be the lift that Hi. I also need to be able to swing it away from the cart. The lowered again. Oh, yeah, yeah. So the gantry idea. Like, you could just wheel the gantry over it and pick it up. And then you can use the gantry for other things.

Host 1     15:27

Yeah. engine hoist engine. It doesn't always do not have to make it add like I beam then. Yeah.

Host 3     15:36

I wonder if I could I yeah, I could totally weld something that could lift the engine. Yeah, absolutely. Ton

Host 1     15:41

Lift? Yeah. Yeah. That's not hard. And we can test that by lifting the engine see.

Host 3     15:49

Yep, and lift an engine, it can lift beer. You love beer. Yeah. So I got some design decisions and process decisions, I wanted to figure out on that, before I go into it, I already bought some like Skookum, like J hinges, like, if I was going to weld it to the frame, I would use j hinges and that way, you can have the crane on it. And then when you're done, you could lift the crane up and then turn it sideways and like put it on the back of the cart. So because this thing is going to be like seven feet tall, and so wheeling it around, like for this huge crane on top. And like bumping stuff is something I'm not looking forward to. I mean, it might not be as pretty as you want. But you could just put, I don't know, just put some like circular.

Host 1     16:41

What's I don't even know the way to describe it. But basically, it'll just like, make the crane just an arm, that's a tube that slides into something on the back and it's able to just spin around. And then when you're done just like take it out.

Host 3     16:55

That's what I'm thinking except the two I was going to I have J hinges. Sure so so they write on a ball bearing at least that that but I am because I can write. But my I guess I need to run some math to like if you have this arm that picks it up, and then you just swung it out. Is it gonna flip the card over? Oh my gosh, she just started to talk about like, torques and moments and arms and things. Remember all that from school? statics? Were no, you're talking about dynamics dude. Well, that's dynamic. But you could treat it as a as a static or not. If it's moving.

Host 2     17:40

As you treat it, you treat it like, what you would do is you would the worst case scenario would be if it's, you know, perpendicular to the opposite direction holding the weight would flip the card over. And you can treat that as a static problem. Yeah, you could. Or what was it? Oh, it's like a dynamic, a dynamic dynamic problem is if you just took a static and just integrated it a lot.

Host 3     18:05

Yeah, that sounds about right. Yeah. Yeah. This is what happens when electrical engineers start to talk about mechanical stuff. Yes, I did take a Stax class. Unfortunately, in college, I, you know, I had the opportunity to either take a statics class or take one math class and get a math minor. So it was like screw statics. Oh, that's cool. What math classes you take waves in wavelets, which, which was an entire, it was awesome, because it was taught by a math professor, but the only people taking it were engineers. So she was very, she was like, I'm gonna make you do a handful of proofs, but it's not going to be all about proofs. It's going to be way more practical stuff. And so we did a lot of

Host 1     18:52

Fourier series stuff. So is a wavelet just a smaller wave, like a pig and Piglet?

Host 3     18:59

It was? Actually it was it was a lot of it was Oh, my God, it was tons of like, if something is periodic, then you can do XYZ to it. It like if you had to describe the class it was it was? Yeah, because I had I took a I took a proofs class and math, that sounds like it would have probably sucked. And most of the people in that class were math majors. I was the only engineer. And I don't know if I told the story before on the podcast or not. But basically, by like the fifth week, I figured out the pattern of how the class was structured. It was like, you basically had to do one proof. That was a pain in the ass. And then the next chapter was, here's the easy proof to do that one. And then you basically it's that repeat. Right? So what you found a pattern in a math class and a math class. Yes. And so what I did is I went I was like, I always read redhead in my textbooks. And so I did the proof, the easy way. And my teachers thought I was cheating. 100% said I was cheating, because I did the easy way because I read one chapter ahead.

Host 1     20:11

That's great. Yeah, I got it. I got a really bad math joke for you. You ready for this? Okay, so an engineer, a physicist and a mathematician are living in a house, and they all go to bed, right? So the house catches on fire in the middle of the night, the physicist wakes up, and he sees the fire and he, and he goes, and he calculates the exact amount of water that is necessary to put out the fire and he throws it on the fire and it goes out. So he goes back to bed. The engineer wakes up later, and the house is on fire again. So he devises this entire like sprinkler system and designs this entire thing. It sets it up and he efficiently puts out the fire. And he goes to bed. And then later in the night, the house catches on fire again. And the mathematician wakes up and he goes out and he looks at the fire and goes, there's a solution to this and goes back to bed.

Host 3     21:04

It must be one crappy house if it catches on three, actually an engineer messed up because he his system only worked once. That's true. Actually. One whoever, I guess actually, yeah, I would. A fire extinguishing system probably only functions once in a while in a building. Commercial building. Well, I it has to be reset, right? Yeah, you'd have to put new fittings because it's like, it works based on like the melting temperature of a pellet. Yeah. Like it like breaks. Yeah. Yeah. Hmm. random thoughts of the fire extinguishing

Host 1     21:47

The macro of engineering progress. Cool. Okay, Steven, what are you been up to? Well, I'll give you an update on the simulation of the week. So the last time

Host 3     22:00

And right here we'll have a sound effect or an intro for the simulation of the week.

Host 1     22:08

Are we a simulation? I don't think it matters. Okay. So what's the simulation? Well,

Host 1     22:15

They said the last time we did a simulation, it was for a low pass filter. It's a ceiling keyed, two pole low pass filter. So for this simulation, I built upon that and made it a bandpass. So we have a high pass and low pass cascaded seal and key filter. But I went above and beyond that. And we've there's also a Monte Carlo tolerance simulation and a Gaussian tolerance simulation in this. So basically, there's three examples in the simulation, which you can go to my GitHub, which is analog, and analog E and G, on GitHub. And you can download this and check it out if you want. So the first example is just a bandpass. So it's like the actual calculation is what you will get. The second example is Monte Carlo, which is a built in function that's in LP spice. So I actually have it simulating, I think it's either 100 or 200 times I can't remember. But it simulates 200 times, and each time it picks a random value for a variable x. And then it plugs that in, and it spreads the tolerances of the components around by that random value and then re simulates. And then you can plot it. And you can see the variations of the cutoff and the cue and things like that across 200 simulations. So with the Monte Carlo, it's entirely random, it's your tolerance is just going to be wherever the random value between plus and minus 10% is what I have set for all of the components on there. So you can sort of see your like window of opportunity of what the filter is going to look like. Worst case, absolute worst case, right? And 200 simulations is probably plenty enough to get absolute worst case. But it doesn't necessarily guarantee that you'll get worst case, because it's all random. But it's probably pretty good. If you wanted to make that 1000 runs and your probability will go up, I guess,

Host 3     24:19

Do 10,000 Because that's a real. Yeah, there we go. That would take forever to simulate. The cool thing is, though, is if you look at the output of the Monte Carlo filter, and the output of the like, just non toleranced filter, you can see the line of what nominal would be and then you can see the spread around that, which is kind of cool. And then there's a third example in this simulation that's assumes a Gaussian distribution of values for each component. So it's a little bit more real world where you can see

Host 1     24:54

You know, with a Gaussian distribution, it's not equally weighted. For all tolerance values, everything is going to be a little bit more towards the center of or the nominal value. And so you can get a little bit better idea of where you're cut off and cue and everything is going to sit. So check out the simulation of the week, I think it's called bandpass. I don't remember what I named it, but it's up there. So I've also been working a bit more on the fermentation controller. The last time we talked about it, I was looking for some parts, mainly the switches, and the kind of the user experience on the front. So

Host 3     25:37

Ooh, I like your term user experience. You know, we started using that at work, instead of like the word front panel. It's called user experience. It's a little bit cheesy, but I like it. I like it a lot, too. Yeah. So or it's better than user interface, in my opinion, because it's just like, What are you experience Are You Experienced? So check this out, I'm showing Parker some pictures, I got like a handful of switches and button pads and stuff. I went, I went on Amazon and just bought a bunch of like the cheapo stuff just to get a feel for like a handful of things.

Host 1     26:16

But before we get into that, I ended up deciding to go with 40, amp, solid state relays on that, just because I want them to be robust. And you know, we're talking about running a fridge on this, which is going to be less than 15 amps. So these are just beefy, Solid State Relays. And I kind of liked the look of these ones. They they're not the like, if you go on Amazon or look at Solid State Relays, you got a bunch of like the tan ones, but I got these black ones that have a plastic cover, they go on top of it is I don't know, they just look a little nicer to me. So I got a eight of those relays in. So that's enough for both of our boxes. And the part number for those is B E M dash 14840 D A, in case you wanted to go and look at these relays yourself. The nice thing about these is that they interface directly with the processor, like you can have a three volt comparable, well, they're three two, what is it? Three to 32 volt, DC, so any anywhere in there will drive it on. Gotcha. So I can plug it directly into my 3.3 volt processor and go to town on it.

Host 3     27:31

Do they? Do they have a rating for like, what you're like? What's on the input? Or that?

Host 1     27:38

It's a they it's basically like an input diode? So they do have an input currently you have to be able to supply but it's minimal. Okay, yeah, it's pretty, pretty easy. So I've actually never thought about that before. They're they're very, very light to drive. Let's put it that way. I've driven them from processors before. And for the for the most part, I should only have like one or two of them on at a time. That because I mean, there's nothing really magical about this box. It's just powering some outlets. So the switches that I've been going for or looking at is mainly for the user experience on the front. And I've been looking at tactile switches and membrane switches to kind of get a feel for what I want it to end up being like.

Host 3     28:30

So the is what sort of driving this is like, what, what inputs is the user going to have to actually press because I could set it up where there's arrow keys, like an enter and a back or something like that. But I've kind of I don't know I don't really like that because I don't like the idea of having to scroll through temperatures with arrow keys. That's just a little bit of a pain in the ass. You know, it's just having to press a button over and over and over and over again. Yeah, so hold it down it scrolls past what you need. Well Oh my god, I absolutely hate that when you have to hold the button you wait a second and then it starts to go but then it like goes like Lightspeed past your number. Well, that's the thing is acceleration is yeah, the acceleration because the because I've designed gooeys that do that on my controls, but I always have a progressive acceleration. Whereas most time you get the slow, slow, slow lightspeed. Yeah, you don't have a slow, slow, so a little bit faster, a little bit faster, and then fast. Yeah, like you need to ramp into it. Because that's what I don't know. I think that's what humans expect. But so when you just go 1234 The

Host 1     29:44

Yeah, actually, the stereo that I have on my TV. I actually inherited it from Chris church at McAfee. It's great. It sounds awesome. It does everything great but the volume control Will is absolutely horrendous on it. Here's the thing. So every upper down, click on the remote is point one dB on it. So like it's stupid accurate, you don't need that resolution on there. So the remote is awful at detecting a button press. And so you'd like to get like a reasonable jump in volume, you have to press it like 20 times, which you'd never want to do. So you have to press and hold it. But if you press and hold it, it goes point one, and then it goes up by one decibel, like at 20 hertz, you know, so it's just like, it's like point one. And then look, it goes up to like, unbelievably loud. And so I've gotten to the point where I've like, I've got a feel for it. I know how, like how to press it just right to go up or down, like by three decibels. You know, it's really, really stupid like that is that's an example of a bad user experience right there. Yes.

Host 3     31:02

Okay, so back to like. So I don't like the idea of just having switches that are like up, down, left, right enter to be able to enter in a temperature, I kind of want a keypad with it. And I've got so many pins available on this processor, I chose that that's not a problem. So I've actually got three different keypads here, I'm showing them to Parker, I got one that's that's got these big, like rectangular or square, chunky, chunky tactile switches. This board is kind of ridiculous, though. On the back,

Host 1     31:38

You can probably barely see it. But there's like, it says the number of the button and then a three digit or four digit number next to it. This, this whole board, which is a four by four matrix of switches. It's just a ladder resistors. So you apply voltage to it. And then whatever button you press you, you send the you send the output off to a DAC or an ADC and then you just read it and you know what button it is. And that's fine. And all that's really simplistic, but that means things get really screwy if you accidentally press two buttons, right, because if you press two buttons, then the value is like all over the damn place. So even though I really liked this one, I liked the feel of it. And they're big, chunky buttons, I don't think I'm gonna go with this.

Host 3     32:24

I mean, I, the funny thing is like the user experience is for me and for you. Like I'm not designing this for mass production. But still, I just, I don't know, I don't like how fast we're gonna be pressing buttons. And they're like, there's no desire to press more than one at the same time. So I could use this, but I don't know, just, I want a seven and a four at the same time. Yeah, I know I'm gonna give it to Parker and Parker is gonna break it. And not like physically, but he's gonna find some way that it doesn't work.

Host 2     32:53

I have an issue pull requests on GitHub.

Host 1     32:59

So I've got another keypad I think this is this is one that's all over Amazon. I think it's also like an add a fruit or a SparkFun one. And it's like an old telephone keypad thing. And I just, it's mushy, I don't like the feel of it. And it's like a really crappy single sided board thing. And I don't really feel like I want to go with that. So the last one I've got here is actually a really cheap or inexpensive membrane button that has sticky back on it and it has a little flex cable that can just plug into the board. It feels kind of industrial, and I liked the idea that it would be waterproof. So all I have to do is mill a slot in the board slide this ribbon cable through and then there's an entire membrane button thing and on the membrane button, it's a four by four grid, it has arrow keys, and a whole number pad thing and a function pad and start and stop. So I could do this all with this membrane switch and the thing about it is it it's good just runs on a key scan algorithm thing. So no problems with pressing more than one at the same time. And the little box I got came with two of these so I already have enough to do both of our projects. So I might use this. I like it has a lot of good features behind it and like I said it's sort of feels a little bit industrial Lee in that sense. So I like that. I did buy 10 of some easy switches from the company easy switch. The part number is 700 SP seven m one QE a P two BLK BLK in case you want to look it up and these are I really like II switches. They're big like beefy switches and they're not cheap. But these are kind of going to be like the ones that will last for sure the longest. So I bought these because they got like a rectangular shroud around them. And the button itself is rectangular and it's really like clicky tactile. And these are kind of like the most like chunky switch out there. So I've got those on the way they haven't arrived yet, but I'm thinking about potentially using those in combination with the membrane switch. So the membrane switch can be for entering numbers because it has a full numeric keypad. And then using these things for like scrolling around menus, just I don't know for the chunky clicky tactilely thing.

Host 3     35:36

So in case anyone wants to check out the membrane switch on Amazon there you top y o t o p,

Host 1     35:48

That's a brand new top 16 Key matrix membrane switch key pad keyboard for Arduino is what it's called. They had it all that SEO. Oh, absolutely. And then the four by four, tactile that runs on the ADC input is by robot dine button keypad four by four module. So they don't actually have part numbers go figure. But these are all like Arduino special stuff.

Host 3     36:17

Other one Oh, is it? That's a yo top? Oh, yo, top? My Yeah. I don't know, I kind of like it. The funny thing was just out of just for fun. I gave all of these switches to my wife. And I was like, if you had a, like, if you had to interface with something, this is the kind of conversations I had with my wife. If you had to interface with something which one of these would you like? And she didn't like the membrane one. She's like, it feels cheap. Does she say interfacing with you? I stuck the membranes on your forehead, my forehead. If you wanted to interface with me, one of these. She's a good sport. But by the next podcast, I'll have all have all the switches and all and I'll have one picked out. Yeah. Yeah. I think it would be I think that'd be actually good test is to actually do the software side for all of them and set them up and be like, Oh, he's holding up the case right now. With the membrane switch on. It looks cool. It looks really cool. That case, I keep forgetting how ginormous that thing. Oh, it's huge. It's it's yeah, it's monster. It's it's like the size of my torso. It looks like a secret service. Briefcase. Yeah. And with this keypad on, it looks like a safe. Oh, yeah. Yeah, a little bit. Yeah. We're going over Killman of Gil always do overkill. Yeah. Cool. Under the RFO RFO. Time. So this week, we got, we got a request on our Slack channel from Magic Wolfire. Or Wolfie, or I'm not sure exactly how to pronounce that name. But they asked to talk about PCBA cleaning to remove excess flux, how's it done in the factory? And maybe some ideas of what to do at home? This is quite painful if you have to do more than one board. And yes, it is absolutely painful if you have to do more than one board. So I guess we can talk about how, let's say how McWrap does it? And then we can talk about how WMV does it? And then we can talk about how we individually do at home? Yeah, I like that. I think one thing that one thing that we can start with is just say that, you know, there's three kinds of flux out there. There's more than that. But there's three like major kinds of flux. Yeah, there's no clean, there's water, clean flux, and then there's rosin bass fluxes. Yeah, I would agree. Those are the three major types. Yeah. And they all have different ways that you can deal with them or not. Or not. Yes. Right. Yeah. So how does that how does macro fab deal with flux on board so at McAfee lab, we have we use two different kinds of fluxes we have a no clean, so all surface mount is no clean. And through hole is done waterbased. The special thing we have is are no clean is designed to be washed and not get ruined, because that's the big thing with no clean is.

Host 2     39:40

It's called no clean flux because it's designed to not be activated at room temperatures, normal temperatures that you interface with PCBs. It is not activated so it won't corrode. That's actually the step farther back is the reason why you would clean flux is so that it won't corrode your body ord because the whole, what the flux does is when you apply it to your pads is it's actually eating the oxide layer off your legs of your chips, it's eating the oxide off the pads, it's eating the oxide off your solder as well, all at the same time, so that when you get that nice Philips on your solder joint, you don't have any cracks, or what's called a cold solder joint. And a cold solder joint is when there's an oxide layer. It can be many things, but most of the time, it's an oxide layer that's in between, like the joints in the joints. And so yeah, put simply, the flux cleans it. Yeah, the flux cleans everything and gets rid of the oxide.

Host 3     40:44

And so a lot of factories nowadays use no clean for everything. Because you get to eliminate an entire step. And your whole process, you don't have to clean your boards after you're done. That's the whole point of going no clean. So are at McWrap, we have it's kind of a special process because we have a pace jetter that you can only get no clean flux for.

Host 2     41:08

So you can't get a water. If it was up to me, we would have water based everything.

Host 3     41:14

And that is mainly because water base flux is really, really easy to clean because you just have to clean it with warm, distilled water. And it comes right off the board and you get really nice good looking boards. Because no clean surface mount stuff, you don't get a lot of splatter or anything. And so it's kind of localized to s&t pads, so you don't see it. But when you do no clean through hole, that stuff gets all over your board, especially if you're running like a selective solder or wave solder, like it will leave residue all over your board. But it's fine because it's no clean. So it's not going to harm anything, but it doesn't look good. And so that's why we use waterbased for through hole. And so that gets into the weird dynamic of not everyone can do that though. The no clean that we use in our pace gender for our surface mount, when it most most no clean, when you wash it, it gets rid of part of it, and not all of it. And actually, when you do that, because it leaves that white crust behind that stuff is awful. And that stuff is actually corrosive. Now, yeah. And so if you want water wash, no clean, most no clean, it will leave this crust behind and that crosses corrosive, and it will eat your board. So So you have to be careful how you're no clean reacts to being washed. And so the great thing is with this luck with the pace jet, the pace that we have to use for it, is it actually forms the no clean. It's actually like a hard shell. Yeah, and it's clear, right? It's, it's kind of caramel colored. Actually, when it's when it's thin, it looks clear, but when does like but like when it goes on top of like the on top of the joints, it does look like a little Carmeli. When you wash it, it, it rejects water, it will not allow any because it makes like a hard shell, it's kind of weird. And so you can so you can water wash it, no problem. And it won't, it won't make the no clean, corrosive. And so that way you get kind of the best of both worlds. If you're just doing surface mount board parts, you get to not have to wash them. But if you have to do through hole, you can use water based and then eliminate all the weird splotches, I guess. But if it was up to me, and if we could, I would use waterbase through the whole process, just because the also opens up so that you can do like RF stuff. Because those guys want their boards, you know, sparkling clean, they don't want any residue anywhere.

Host 1     44:02

You know, I would I would say RF, but also AF at the same time. Oh, audio. Yeah, well, and here's the thing, in a lot of the circuits that I work with, we we deal with current forces that are based often the collector of of BJT transistors, and you have really high impedance there. It's such that it makes a really nice current source. And if if you have any flux residue around those transistors, you will get issues and offsets and things. And so in oscillator circuits and things like that, I make 100% certain there's not a lick of flux, even if it's no clean flux, because that flux adds parasitic parts effectively to your oscillators and you'll get incorrect readings or you'll get like like on on a saw wave or like a linear ramp. It'll turn turn into like a pseudo sine wave, you know, it'll curve, because you have this added impedance that you were not expecting on there. It's it's especially bad like, like I said on high impedance transistor circuits, but also film caps and things. They do not like flux being around. And in general, so So a WD we also use no clean flux, we actually switched over last year to entirely using no clean flux. And that works out well. And we've we've tested actually, the the no clean flux we have can be washed and it washes off entirely. Yeah.

Host 2     45:34

So do you all use like a subpoena fire?

Host 3     45:37

You know, and that's, that's the interesting thing is we don't actually use a saponify. Or we found a no clean flux that will water wash if you want to. Oh, that's cool. But we don't need to most of that. Yeah, that's one thing is if you're if you're wondering what the client with usually the manufacturer will tell you, the flux will tell you what it is, right. You're gonna have to tell me what that no clean flux. That's kind of exciting. That's true. Yeah. I mean, the thing is we actually so we have a aqueous technologies Trident, board washer. And it's basically a glorified dishwasher that you put Yeah, that's what we have. Yeah. And those things are, they're great, but we stopped using it because we didn't need to. So we don't actually wash reports. But when we did some tests, previously, they came out. Clean, super clean with without, I mean, with the no clean on it. Yeah. So you so your boards have no clean all the way through, then

Host 1     46:34

That's right. Yeah, we do. We do no clean all the way through. But if we have like specific requirements, like really sensitive circuits, we can wash it if we need to. And that's actually an important thing to note. Like, if if you are engineering a product, and you know, you can't have flux on something, you can specify that to your manufacturer, and they can do specific cleaning processes. But if not,

Host 3     46:59

Contact your manufacturer and figure out what you're going to get because some some will wash your your things, and then some will use no clean without you asking, you know, you'll have to find out what paste they're actually using in there. Yeah, I think it's it's something like 80 plus percent of, of contract manufacturers use no clean now. Yeah, cuz it's easier. Yes, faster? Well, it's a whole process that most time you don't need anymore.

Host 1     47:25

Well, and we've switched over 100% to using no clean for rework to so we're no clean, top to bottom, um, we do have some water wash, in case we need to. But yeah, just in general, we didn't know clean. So so that's, you know, at your manufacturer, the best thing and like we, we've said this 1000 times about a bunch of different things is just always ask. Always ask always asking if you need to put it on your fabrication drawing if if you have a fab drawing of what you want to use. And that's important

Host 3     47:56

Too, if you're, especially with parts that are sensitive to being washed, like men's microphones, and any kind of Anymote a lot of like open air sensors, stuff like that. Anything with a hole in the top with a hole on top. Yeah, is let your manufacturer know that you have a part like that. Right? And so that they're not stuck, or you're not stuck with 100 boards or the 1000 boards, that that part is ruined because they got washed. Exactly, yeah, no, that's that's really big. And that is that is both the responsibility of the manufacturer and you to chair and designer and yeah, you to check but also like it's not, especially at macro fed, it's hard to check every single part single part that comes in Yeah, I remember multiple times, we had to redo orders there because we washed a part that wasn't supposed to be washed, you know, there so at the same time, anything that is supposed to be like if you have like jacks or something like that they can be washed, but if you have potentiometers most of the time they can't be washed. And when like if you want to play it safe, just assume that it can't be washed, you know? Oh, that's another one. wire wound inductors Yeah, well, they will they corrode, they corrode. So do not wash them. Yeah, that was one of the hardest I guess like failure analysis I've had over like the past six and a half years now was trying to figure out why some a batch of boards wasn't working. And on this one circuit and it ended up being a through hole wire around resistor is inductor is the flux got in there the water clean flux got in there because it was because some boards were built by hand and then some were built by the selector solder. Well, it's like the solder shoots the the like with force, the flux Yeah. And so it actually shot through though it would shoot through To the holes and up into the inductor, and that would and when you would wash it, the wall, the water wasn't getting up into the inductor to clean it out. So basically we had to go is go from, we went from using water with that product to a no clean for through hole. And of course the customers at that point, it's like all this flux is leftover. And I'm like, You want Corona conductors? Or do you want a little bit of flight residue? No, no clean flux residue. It's like God pick your battles. Right? That's that. Yeah, that's a tough one. That's really hard. And that's on the manufacturer on that. You know, for that? Yeah. On that one. Yeah. And that's one of the things it's like, like Colbert, who has run, you know, how long has he been doing? Oh, 20. Several years. Yeah. And he's never seen that before. Yeah, I remember that one. That one was crazy. Yeah. So when it comes to rosin flux, I say get rid of it, no matter what like, and a lot of times like for hobbyist do and stuff at home, a lot of times you end up with rosin flux, because you have like an old spool of solder, or that's what I use at home. Yeah. And it's great cleaned like, like crazy, but it's it also leaves like, goopy earwax all over your board. Yeah, I would say that's probably the best thing is, the call is earwax. Yeah. And it doesn't smell great. And and it doesn't age well. Just get rid doesn't corrode your stuff, though. It just sits around looks nasty. Yeah. And smells terrible. So yeah, get rid of the get rid of rosin flux as soon as you can. So at the home, how do you clean off rosin flux,

Host 1     51:46

You know, so a while ago, I went out. And actually, pro tip, pro hobby home tip, I guess, target sells alcohol for really cheap. I can't remember how much it was, but it was really cheap. And they sold, they were selling like 95% Alcohol. And so I bought like a whole bunch of it. And I filled a bucket full of that stuff. And, and I'm not talking like don't go buy nail polish remover or the alcohol that has a green tint on it. Don't buy the crap for taking makeup off, buy like as pure alcohol as you can. And I keep a bucket of that around all the time, such that I can just scoop some out, put it in a tray and then throw my boards in there. If if I know I have a bunch of parts that just don't like are fine with being submerged, like ICs are fine with being submerged. Resistors cap for the most part are fine with being submerged, I just drop it in alcohol and let it sit for a while. And then I'll either take a nice, like soft bristle brush or a toothbrush or something and I'll just go over it. And I like to spray it off with compressed air. And I will do it until the boards don't feel sticky anymore. That's that's the biggest thing like, no, no, I mean, like flux leaves a sticky residue keep going until that until they're clean and like not sticky at all. And that's the biggest thing. Because I've learned certainly in my own work, but also at work. If a board feels sticky, they're still incorrect impedance there. And and I've, I've, I've traced it before where I wash a board, I look at an oscillator on the scope. And I see that it's not linear. And then I wash it and I end it's a little more linear and I wash it. And then I just let it sit for four hours and alcohol spray it off and it's perfectly linear. So just let a whole bunch of alcohol fix it for you.

Host 3     53:40

Yeah. So I at home, I do the alcohol thing, except that I usually buy. I'll wait for my local liquor store to have Everclear on sale. That's, that's always cheaper than isopropyl. Yeah. And you could drink it. Not after the flux. After the flux. No, but you so yeah, I basically buy the cheapest Everclear you can buy. And that works really well. Because that's like straight up grain alcohol. 98 95% No, no, it's more than that. I can't remember, there's a certain amount that the atmosphere will allow, you know, like, the atmosphere. No, seriously, like, you can't have 100% alcohol, right? It's because it'll just evaporate or like, oh, there's a word. There's a word for that. Yeah, it'll drink moisture to be like at equilibrium. I can't remember what that's called. Yeah, there's a certain word for that. But you can go higher, but you have to add stuff like benzene to it. Right? Right. Yeah. Don't don't drink that stuff. Well, and I think I think benzene also tastes like but right so they put it in there so that you don't drink it. That like IPA? is I don't think that's the reason why they put it in there. But oh, sure. I think I think they like it is a terrible flavor. And, you know, I mean, I've never tasted it before. So I'd neither have I but I've heard. So don't go drinking IPA. It's not good for you. There's plenty of other alcohols you can drink. Yeah, anyways, I use I usually use, you know, ethanol, ethanol, grain alcohol or Everclear from the store, because that's usually cheaper than I like buying like 95% or higher IPA is generally very expensive. Yeah. And the stuff for human consumption is way cheaper. I'm telling you, maybe I just got lucky. Maybe it was on sale, but I went over to target and it was it was dirt cheap. So I bought like the whole shelf or the belco. Yeah. Yeah. So either do that or I wait for goes on sale to at specs. So which is our local, you can get your hands on actual like PCB IPA. That's, it's more expensive. And it's harder to get for a home gamer, but it's better. It's more pure. Because I mean, even the stuff you buy it, like the drugstore is still diluted somewhat.

Host 2     56:00

Yeah. And so back when I was renting a house, I had to build a whole bunch of pin hacks. Myself, this is like when I was first doing them. And so it was a lot of through hole. So I actually did I use waterbased flux. And I random my dishwasher.

Host 3     56:17

Didn't you also put some a simple green in there? Yes, Simple Green Works pretty well at cleaning boards. Yeah, and that worked great. So if you have a wife or a girlfriend, don't do that, because I'll probably yell at you. But it's completely fine. I've done it too. And what's great about it is they come out, like they're done when they when they come out. And they're like, sparkling clean. Yeah, you put a little bit of that little dish additive into their comes out perfect, sparkling clean. Here's, here's another thing, scour your data sheets, for if your parts can be washed. A lot of times, it's actually in there. It'll say this part can or can't be washed. And be careful. I even remember at McAfee up multiple times looking at data sheets, and they're like, footnotes, they're like tiny text written, there's like this part cannot be washed. Don't get it near, you know, a mile from water, you know? Yeah. And keep your eyes out for that. Because well, okay, so most of the time, if it can be washed, they don't say it can be but if it can't be, it'll be in there. Yes, it will be in there. Yeah. Cool. Hopefully that answers some questions. You know, actually, it, the whole thing about this is painful. If it's more than a few boards, get a bunch of alcohol, put it in a tub and throw all your boards in there. And batch process. That yep, that's the answer I have for that. Yeah, that's, that's really the only way to do it. I haven't really experimented for because usually it's either I use a dishwasher when I had a batch do at home. And I used waterbase though to make sure that that was possible. Whereas it depends on if it sounds like magic wolfies already got a whole bunch of boards and he needs to clean so he might be stuck with it was rosin or no clean. If it's no clean, they don't clean them. Don't worry about it. Unless you're doing something that requires like your impedance to be really specific, but I doubt you are. That's kind of specific. Cool, oh, oh wait, wait before we go on, do not use an ultrasonic cleaner. Yeah, NASA actually came out with an article about that. Yeah, whereas an ultrasonic cleaner works really good at clean bare boards or stencils or anything else but like it will actually destroy the inner workings inside ICs and stuff NASA has their their research found that it can it can damage the end caps on capacitors and resistors to yeah for SMT parts. So it led for them it led to late and failures. Yep, so don't do it does a damn good job at cleaning. Yeah, that's actually um, Robin i i was looking for gone another big tangent right

Host 2     59:18

Here. I was looking for better carburetor cleaner to soak parts into because like the stuff that you get now is not as good as it was like a decade ago. Like the stuff a decade ago, you'd put it in there. And then like a couple hours later you pulled out and it would have stripped everything off the metal. Now you pull it out a day later a day later, not just couple hours and you still got scrub stuff off, dip it back in for a couple hours and keep scrubbing so it's just not as good anymore. And so I started going online and figure out how do auto shops deal with this because they can't buy the quote good stuff anymore, right?

Host 3     1:00:00

Ultrasonic cleaners. Yeah. They like an ultrasonic cleaner. Throw some water, water and soap, like dawn. Yeah, throw the parts in there and then turn that sucker on and it works on like, okay, that's I got to get an ultrasonic cleaner just to clean auto parts and apparently cleans it just as well as the old chemicals did just without the nastiness of the chemicals and and get a good one that has a heater built in. Yes, those they say the heaters were really good. Yeah, yeah. It just rips through grease. Yep.

Host 1     1:00:33

And all those, you know, those booger looking things that get in your carburetors? Is Gable Yeah, yeah.

Host 3     1:00:40

Yeah. Ethanol goop. So I found a fun little clickbait article that I thought we would we just touch on real quick. And whenever I see these kinds of things, I'm like, really? Like, is this just SEO content? Or is this like, is someone like out here reading this? So II web had just the five best Arduino projects article in the world in the world? And I'm sitting here thinking like, is Blinky going to be one of these? You know? What? No, I'd say it's actually like, they've got a whole article with somebody wrote up, like, a fairly significant amount of content about five different projects that are out here. And and, you know, with this kind of stuff, I'm I guess I'm a little bit confused, because there's like it. Yeah, they're projects, but they're like projects that other people have done. They're not necessarily projects that are like, geared towards you being able to do them. So it just feels really SCO II, you know, a little bit yeah, that's what looks like to me, like the first one is a robot arm. Yeah. Cool. But the best Arduino alarm system and they using the keypad that you have. Yeah, it's almost identical to it. In fact, yeah, it is. It's the same keypad just different. text on it. Text. Yeah. Oh, there's home automation. Scary. But it's very high level, like, you know, but it's interesting how it's written because if some of that stuff is really high level, and then it dives into the part numbers. Right, but it doesn't even give like a link to where you could see the projects. It's just like, yeah, ideas. Has engineering devolved to SEO? I think so. Always has been ever since man had to sell another device. Any Seo. Yes. Okay. I don't think I have anything more to say about that. I think we're done for this episode. Okay. Well, that was the macro lab engineering podcast. We were your host, Stephen, Greg. And Parker Dolman take it easy

Host 3     1:03:08

Thank you. Yes, you our listener for downloading our show. If you have a cool idea project or topic or Arduino project that should make a top five best list. Let Stephen and I know Tweet us at Mac fab at Longhorn engineer or at analog EMG or email us at podcasts at Mac fab.com. Also check out our Slack channel which will totally be discussing our top five best Arduino projects in the world. And if you're not subscribed to that podcast yet, click that subscribe button in the podcast app you're using. That way the latest episode gets right to right when it comes out. And please review us on iTunes or on Google or wherever you listen as it helps the show stay visible and helps new listeners find us

Transcribed by https://otter.ai