- PCB serial numbering or other unique marking
- I have batches of boards and need to put some unique numeric identifier and a MAC address on each board.
- Besides stickers, laser etching, scratching or DIY stencil and paint, what are folks doing?
- PinoTaur REV 3 boards arrived
- REV 4 changes already 😉
- Ben Jordan of Autodesk will be a guest on the podcast
- Product manager and is responsible for the EAGLE product
- Get your Eagle questions in early!
- Adventures in Rpi 4
- Midi overkill
- Clock master
- Possibly be a usb midi device?
- Roland TR-606 drum matrix
- Looks like a possible ram IC issue. The unit functions and does things that only a processor could do like blink leds at a particular rate
- None of the buttons have memory – they do not “store” anything
- Upd444c obsolete – 6514 SRAM is a drop in replacement
- Roland TR-606 drum matrix
- The FX Dev board returns
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 1 00:10
Hello, and welcome to the macro fed engineering podcast. We're your host, Stephen Craig
Host 2 00:14
And Parker, Dolman.
Host 1 00:16
This is episode 252.
Host 2 00:19
So you know what episodes coming up soon
Host 1 00:21
Host 2 00:23
There's that that's next week. And that's actually with Ben Jordan of Autodesk. Who's going to be a guest on the podcast and he's the product manager and responsible for the eagle product. So on Slack channel, get your question your Eagle questions in. I don't know exactly what we're gonna be talking to Ben about, but it's probably gonna be about Autodesk and Eagle.
Host 1 00:44
I would assume so.
Host 2 00:47
The Star Wars Episode.
Host 1 00:48
That is coming up, isn't it? Yes.
Host 2 00:51
So that's coming up in 231234 weeks, four weeks?
Host 1 01:01
256. Yes. All right.
Host 2 01:05
We usually record that like the week of Christmas. Right. So what are we? What did we don't have a new movie? Oh, there is a new movie. The LEGO Star Wars Holiday Special.
Host 1 01:18
Honestly, I saw the trailer for that. It looks awesome.
Host 2 01:21
Yeah, so I think that might be the because we usually do a movie like the latest movie is like the topic of discussion, so we should probably do that. There's also Mandalorian
Host 1 01:30
I was just about to say the man goes out and this season is already awesome.
Host 2 01:36
Yeah, I'm pretty pretty hyped.
Host 1 01:40
If anyone out of Disney listens to this podcast, just make Mandalorian style movies just make the main movies like that and we'll all be happy
Host 2 01:52
Well, there is some controversy about the Star Wars Mandalorian but not many, not many. Not a lot of people
Host 1 01:59
Nowhere near as much as all the other stuff. Yes.
Host 2 02:03
Okay, let's get down to electronics so the pin guitar revision three PCBs landed
Host 1 02:14
What did they land today or had they been marinating on your desk
Host 2 02:20
And teriyaki sauce. Teriyaki flux No, I they finished up last Friday at pick them up on Monday I still have to solder all the through hole components on it.
Host 1 02:34
Oh, I was about to say that they're not already in games being played.
Host 2 02:38
No no, not yet. Not yet. My actual problem was my I was actually going to solder them all last night the through hole components. But the my soldering iron wasn't working my ex Tronic 4040
Host 1 02:52
Sheet or that thing's still still chugging along. Yeah, not the
Host 2 02:56
Hot air gun works great. The iron not so much. It apparently did not like being left on for some undisclosed number of hours or days or months. I had no idea how long I left it on for one time I came into my office here and I noticed that it was on Oh wow. That's no bueno and it wasn't hot.
Host 1 03:20
Okay, so you just cook the living bejesus out of it?
Host 2 03:23
Yeah, that that heater element you know, no longer functions as a heater element.
Host 1 03:30
I think the longest I've left an iron on is overnight. Yeah, and it's always like you get up in the morning and like
Host 2 03:35
Oh damn it Yeah, some smells a little bit off and in that room but I had no idea because I actually couldn't remember when I found it on I couldn't remember the last time I used it
Host 1 03:50
Yeah, that's know that that iron was there the day I started at macro fab
Host 2 03:55
That was our original iron. I actually brought that in from from home to solder boards at macro fed. Nice. So then I brought it back home once we got proper equipments. And yeah, that
Host 1 04:12
Was like five I mean, that thing was in the that thing was in the engineering lab for a while too.
Host 2 04:17
Oh yeah, that too. Yeah. So yeah, instead of replacing because I can actually just go online and get a new heater element for it. Actually just went into like the junk closet at the fab and pulled out some soldering irons. I was just in there. So I'm using a nice thermal tronics now. Oh, wait for go. Yeah, I gotta wait for a tip to show up for it. But I have a nice thermal tronics unit I can use.
Host 1 04:47
We have we have met cows at WMD.
Host 2 04:50
And yeah, MacAllan. thermotropic is back cow
Host 1 04:54
Right there. They're super
Host 2 04:55
Nice. Yes. That was the curry heat. It Fact.
Host 1 05:00
Yeah, they heat up pretty damn fast.
Host 2 05:03
It's not just that but it's also they hold the heat. Way more. Way more effectively, I guess then a a resistor element style. I think because I can just quick, more quickly react to the thermal load. Anyways, yeah, rev threes done revision four coming soon, not because of like, it's already ish. Yeah, not because of anything bad. I already found some stuff that I just don't like about the layout. And I can go ahead and change it, and they will satisfy my own inner demons.
Host 1 05:44
Your OCD? Yeah. Well, not OCD. But just out of curiosity, what are the things that you already don't
Host 2 05:52
Like? It just like the how I had the solenoid power input, it's completely fine. I just don't like how the orientation of the components is sub optimal from a space layout. Even though I'm not going to change the spacing of the layout, I just wanted to make sure like, the I have a big I have about a square inch of the board that doesn't have a component it bothers me. So I can like rearrange the components so that the current flow or current path, I should say, will be better. And the components will be more spaced out instead of clumped in that area.
Host 1 06:31
Just Just do the hacker maker thing and put text there and put like a random like poem, or
Host 2 06:38
Had does have the pennant tar logo that Enoch did for us. Oh, wait a day. Yeah, it's on the back of the board.
Host 1 06:48
Let me see. Let me see. Oh, that's that's legit. You're gonna need to have a picture. Put that out.
Host 2 06:55
Yeah. Well, the idea is to I don't know if we're actually get to it is actually let like, the person who does like the pin the pinball art for the machine to do it on like the also let them do the art on the backside of the board. Like no one will ever see it unless you're servicing the machine. But still pretty cool.
Host 1 07:20
Actually, if you have spare space on the board, like now, nevermind, probably not a good idea. I was thinking, have it such that like repair log, or like signature or date or something? Could be Sharpie on to the board?
Host 2 07:32
Oh, yeah. Oh, yeah, there's gonna be a spot for a serial number thing in the middle of the board. Yeah. Because I do want to actually do like, tracking of where, what happens to boards and stuff like that. Yeah, all the other changes we did, like we made the pads larger on the on the MOSFETs. And so it's easier to do to do reflow on them, instead of making basically the pad the size of like, the part footprint or you know, a little bit bigger, actually made the pads larger and put solder mask over them. So actually, they're technically solder mask defined pads for MOSFET. Kind of weird, but that way they have more stiction to the FR for. So when you do when you do rework, especially on prototypes, where like, people tend to blow up MOSFETs because they wired machines up wrong. It'd be easier to do rework on them. And oh, yeah.
Host 1 08:27
Have you ever de lambed a Deepak footprint?
Host 2 08:30
I've never have I never have, but I've had reports people have done
Host 1 08:34
It. Wow, you have to go to town to D Lam one of those? Yes. Yeah.
Host 2 08:40
And then we added diodes to the backside of the board. So that's cute.
Host 1 08:45
Yeah, they're just protection diodes.
Host 2 08:48
Yeah, they're the reverse flyback that used to go on the coils. Oh, that's right. We
Host 1 08:52
Talked about that a few weeks ago.
Host 2 08:54
Yeah. Cool. Revision 14, mainly, it's mainly layout changes, there's not going to be like will change the connector for the RGB lights, because right now it's a big 3.96 millimeter pitch Molex part, we're going to switch that to a, a two millimeter. What's it Yeah, a two millimeter one, which is the connector we're using on all the RGB boards. So this way, it just keeps the same connector family throughout the whole chain and not have to have a special starter cable for that whole RGB serial light chain. Just makes it simpler on the bottom and ordering cables. So you only have to order like one type of cable instead of like, two types of cables.
Host 1 09:45
Hey, speaking of cables, have you man, have you gone out and gotten quotes for your cables yet?
Host 2 09:50
I actually just finished up changing all the stuff that we talked about last week. I finished all those changes on Monday this week. And so they went out today. You Oh, very cool. Yeah. So we'll definitely know, in two weeks, because next week is is guest podcast. Or if you follow me on Twitter, I'll probably be posting about it too. No, cool. Exciting stuff. It's getting there.
Host 1 10:21
But revision five is right around the corner.
Host 2 10:24
I hope not.
Host 1 10:26
You think four is going to be it?
Host 2 10:29
I think so because the revision threes is production ready. We're probably not going to be changing any components like part numbers and stuff. So it will be basically probably read for the build a couple just to make sure I didn't actually completely mess something up with the power circuitry and then just, you know, print. Right. So, you know, for you might actually just be I just even be ordering a blank board and just printing it out and making sure I didn't short something when I changed all that circuitry.
Host 1 11:01
Yeah, yeah. cheaper and easier. Yeah. Because
Host 2 11:04
We don't need to, you know, we act or just Pipelight even just the MOSFET section and power input and just make sure you know, hey, this actually does work.
Host 1 11:14
So in your, for your experience. Do you think three to four revisions is kind of like the sweet spot
Host 2 11:23
For pinball controllers? Yes. Because that's it's really funny because pin hack, pin HEC rev four was the first production pin heck. So for people who are new to the podcast pin heck was the original Pinball Controller designed like seven years ago. And actually right around the start of macro fab. I designed that Pinball Controller, pin. Oh, TARS the new one. And yeah, rev four pin heck was the first production version. And then the LTS or long term? Was it long term support? lts was Red Five of the pendant. Panek. So
Host 1 12:05
How many pin hex were have been made?
Host 2 12:09
Like 1000 1500 ish. Somewhere around there. I got one right there. That's actually was the last revision ever made. That was a rev eight board.
Host 1 12:18
Oh, all the secrets are in that one.
Host 2 12:21
We have we made we did have rev six, which we never actually made a release of. And then we did a rough seven. And then this is a rev. Eight. And rev seven and eight never went into machines. Rev eight has an onboard Raspberry Pi compute module.
Host 1 12:38
Oh, that's right. You spent a lot of time on that. Yeah, that
Host 2 12:41
Was a lot of time for no no gain. Learned a lot though. A lot about it. A lot of what we learned about pin hack went into penetrator. Like it's funny though, is like we basically started 100% over though,
Host 1 12:55
Like Yeah, but you stripped down a ton of stuff. Yes. Yeah.
Host 2 12:59
I mean, it went from a 14 inch by seven inch board to 11 by five was it loved my four, five as measuring it with calibers.
Host 1 13:14
Because that's what we have available right at our desk. Within within reach of my right hand. I also have my
Host 2 13:21
You have the X style, like Viking style, which is like
Host 1 13:27
The big ones. I have the Harbor Freight cheap style is the one that I have read.
Host 2 13:33
Well, me too, but mine is only the what six inch model.
Host 1 13:38
Oh, yeah, I bought mine for doing my CNC work. So this is the LED 12 inch or whatever. They have 11 inch,
Host 2 13:45
I think. Colbert who's the operations manager at macro fab. He's got like, the Mondo like 24 inch one.
Host 1 13:57
Yeah. Your panels. They are ginormous.
Host 2 14:00
Yeah. But yeah, it's like a freaking tomahawk. It looks like
Host 1 14:06
Cool. Yep. So a speaking of Raspberry Pi from yours to my projects. Actually, I don't know if I'm going to do this on Raspberry Pi. I'm still figuring this out. But but on my adventures in Raspberry Pi section. i A few weeks ago, I said I wanted to come up with a project that uses a Raspberry Pi. And I think I have one. And I kind of love it because it's so so ridiculously overkill to have a whole operating system to do this one function. It's so overkill that it feels appropriate for a Raspberry Pi. Because like so much about the Raspberry Pi is overkill especially for what they end up going into. So I ran into a situation last night where I was playing some music on my Korg Ms. 2000 which is a keyboard synthesizer And this thing is really fun. It has like, it has all the features you could possibly want minus the one that I do want, it doesn't have a trigger out, it doesn't have a sync pulse. So it's, it's really great to play by itself, but connecting it to other instruments. It doesn't play super nice. So in the world of synthesizers, usually you have some kind of sync pulse, that one master item in your whole array of gear can send a sync pulse out such that everyone is working off of the same time base. It's like a metronome. Yeah, it's a metronome. Effectively, everyone has the same clock, right. So in the world of modular synthesizers, that's everything. Everyone has to have a master clock that that they're all talking off of. But in the world of like keyboards, synthesizers, everything is contained within the box. So they don't necessarily need to communicate with other things. So like, the thing was my Korg, Ms. 2000, has great sounds, but it doesn't have great drums. And I wanted to interface my ms 2000, with my modular drum rack, so I can have a drum beat that I can jam along with, well, I need to bridge the world between a keyboard synthesizer and a modular synthesizer. And of course, I could just buy something that does this. Or I could make it right. So I'm trying to I'm trying to come up with a project that's like, I can start it and finish it in an afternoon, you can, like make a really simple kind of thing. So my thought was, could I take a Raspberry Pi and turn it into a MIDI device that would receive a MIDI clock signal from the Korg, Ms. 2000, because the MS 2000 does have a MIDI interface. So receive the clock commands from the MS 2000 and convert that into a voltage pulse that gets sent off to the modular synthesizer, which can read in voltage pulses, and instead of running on their internal clock just stepped forward in time based off of the clock, the MIDI clock that comes out? And the answer is yes, I mean, that's certainly a possibility. This could also be probably easier done with an Arduino, right, and like a little MIDI shield or whatnot. So I that's what I what I was meaning earlier was like, I might not do it on a Raspberry Pi, I might do it with an Arduino. But I'm still figuring out which one I'd rather do. And frankly, I'd rather my computer send the MIDI information. So I would love to send MIDI from my computer's audio interface to the Korg, Ms 2000, and then use the MIDI through on that and patch that back again to a MIDI to voltage converter, that then goes to my modular rig such that my computer is the one that has the master clock that tells both the MS 2000 and the rig to march in time. So I don't know, like I still have to figure out which method I want to do. Like do I need this to be tied to my computer? Probably because it's not like I'm not one of those cool guys that goes out and plays music anywhere I'm that I'm that like, basement, troglodyte that plays music down here, you know. So that's sort of my, I don't know. So I gotta figure out which which way I want to do it. There's, there's, of course, there's one wrinkle to the whole thing on there that I'm not sure if I want to, if I want to implement or not. So
Host 1 18:45
You could have a clock that just always clocks, right? That's always just at whatever tempo at whatever BPM is just spitting out voltage spikes, and MIDI CC commands that say clock. Now clock now kind of thing. And that's fine, though, the one thing that is not great about that is you can you each clock pulse doesn't differentiate from any other clock pulse. So if you have it going to your drum rig, your drum rig will just keep playing drums like you can't. In other words, you can't reset your drum rig. You can't say, Okay, I pressed a key on my keyboard. Now begin the drum clock, you need a little bit more complexity to that. And I'm not sure if I really care about that. Because what I could just do is wait to play my music until the drums you know, loop back around because in most cases, I'm only talking about one bar or two bar or whatever.
Host 2 19:39
Well you could if you using the Raspberry Pi you can put the smarts in it.
Host 1 19:42
That's you see, that's what I'm thinking. Well, I mean, you could do that we're
Host 2 19:46
Gonna read Oh, only trigger the drum clock when the certain key gets pressed
Host 1 19:53
And effectively what it is is my my Ms. 2000 is played by me that's an actual keyboard. So Like, I control the reset. But my my modular rack is not actually played by me, it's just receiving information from whatever it's receiving. So really it it has two signals, it has just the traditional clock signal, but it also has the RESET signal. And the RESET signal is basically just says, go back to the beginning of your pattern and start from there and just listen to the clock from that point. I just don't know, I think I think what I'm going to start with is just the clock configuration and the translation between MIDI and voltage and start from there and just play with it. And if I really need to add the reset pulse, then I'll figure out how to do that then. So I don't know, I think that's kind of funny to get eight gigabytes of RAM. And like a full full keyboard system to be able to listen to a MIDI command and send out one voltage spike. I love that. So that's my adventures in Raspberry Pi four.
Host 2 21:03
I got it. You can you can play Doom on the Raspberry Pi with your MIDI keyboard. With your synthesizer.
Host 1 21:14
Oh my god, that'd be incredible. Yeah, you totally could. I mean, the thing about MIDI that really, really sucks is it is super slow. Like just it's brutal flow. So I mean, I you could play it, but it would be awful to play. Like key inputs would be pretty laggy. Someone Someone has to have a Raspberry Pi poor to do like that has to be a thing, right?
Host 2 21:44
Oh, yeah. I mean, there's a whole website dedicated to things that run doom.
Host 1 21:50
Yeah. What is that website? Is it will it run doom? Could be because I swear I saw one the other day where someone got Doom running on a pregnancy test.
Host 2 22:06
I think they were just using the screen. I don't think they were actually running it on the circuitry in there. Well, I'm sure
Host 1 22:12
That's the case in in with most of these things right.
Host 2 22:15
Now, most of them they actually get it running on it. I could be your
Host 1 22:21
Only Gosh, this toaster. Do. All right, we're getting off in the weeds here. Yep. I don't remember what the website is. But I think it's something like Will it run Doom or whatever?
Host 2 22:36
Run doom in Task Manager. That's great. Okay, yes. Um, okay, so what I want to talk about is the PCB serialization. This was a topic a couple of weeks ago, we just haven't gotten around to it, to talk about it, because we just keep coming up with amazing, interesting topics that kind of supersede this. But it came up in the in the chat Slack channel a couple of weeks ago, is PCB serialization and other unique markings for controlling, I guess, inventory, or tracking your PCBs for testing and programming? So how do you do that at work? If y'all do that?
Host 1 23:26
We actually do not see your allies early on in the process. We do it way later in the process. And one of the reasons excuse me, while we do that, is it allows us to break apart runs if necessary. Throughout the process, let's say we have 100, PCBs and five fail in testing, we can actually break those apart and send them off to engineering and then pass the other 95 that do pass and serialize those so we serialize late in the process, it just works for us because our quantities are lower.
Host 2 24:02
Oh no, that's a completely valid method. That's what we mostly do here at the fab two. It's the main the main thing is is we do it kind of like a QA control quality control is making sure your everything gets you know, looked at tested, programmed, etc, etc. Whatever post assembly you need to do. I kind of look at his like post maybe post aeoi is what we typically do, but not a lot just let's see. Go beyond actually like what do we use it for? But like how do you do it? Stickers markings. I think we were talking about in the Slack channel. Like using silkscreen to change it or stickers. I personally like stickers. The best so Lock screen changing the silkscreen tends to be very, very expensive, man because you got to think of the process of they're actually silk screening with a screen that that marking on it. So they have to change that screen for every single panel. Right?
Host 1 25:17
It doesn't make sense to do it that way. No.
Host 2 25:22
Ordering it, I saw some people do it and copper as well. And I'm like, Oh my God, how expensive that is. It's basically every single. Every single PCB is different,
Host 1 25:32
Right? Yeah. New Gerber's for each serial.
Host 2 25:36
Yeah. If you're still on number. Yeah.
Host 1 25:39
I remember at the Feb, we, we had found polyimide stickers that could be printed on and they were high temp. So you they could actually be applied to a panel before it even goes in the reflow. Oven.
Host 2 25:51
Yes, yeah. And that's how we track panels is the polyimide sticker. And you have to get a you don't have to get a special printer for but you have to use a a printer, a thermal, instead of like a direct thermal, which is like what most people's when you print labels off of shipping and that kind of stuff. That's a direct thermal. But you have to use a thermal transfer tape or film. They call it film transfer film, which is a basically the film has a higher temperature. Ink, I guess I don't know the exact science behind it. And then it transfers onto the polymide label. And then that material, the polymer label won't discolor because it got hot, because it's not direct thermal. I don't know what the exact chemical reaction is or not chemical reaction. It's probably a lot like toner. I'm guessing yet Stephens shaking his head is an adhesive. And an anchor b&e, he's
Host 1 26:58
Have Yeah, yeah. And and he's have an ink that after cures is not susceptible to high temperature. Yeah, could be
Host 2 27:07
Seems it makes someone knows someone will correct us. Without anyways, you use a pie my label, yeah, that's got a thermal film transfer, print on it. And if you have like a Zebra printer or whatever, just make sure it has the ability to use a transfer film.
Host 1 27:26
And that's nice. If you are wanting to serialize the moment your your item hits the door, you can print it out, stick it on, put it in inventory, and then you're serialized from day one.
Host 2 27:37
Yeah, we use those when we're doing all post QA too, because in case you had to go back and do rework, you don't have to reprint a sticker. The sticker is going to live the reflow live through rework. But kind of like the example I will give is the process I like to do accessory for mid volume stuff is to basically once it comes out of reflow and so on panel is sticker everything then. And I generally will just print it as a serial number, it doesn't matter, right, and sticker the entire panel up that way, then deep analyze, and that way you can also so if you have a panel number, like here at the FAB, you can associate the PCB with the panel ID. And so then you kind of have an idea of like, you use those as like mini lots. And so you can track failure rates that way. But that way, throughout your whole process is you scan that barcode from the get go instead of like coming up with a serial number ahead of time, and then applying it to like that, that I think it head times the wrong word. Instead of like reading an ID off the like let's say you had a serial number that you get the the chip can generate or something like that. A unique identifier, or I think this was a MAC address, right? Instead of reading that and then printing a sticker. Like, you don't really have to do that you can just put a unique number on the board. And then later down the road associate that unique number with that MAC address in a database somewhere. That tends to be a lot easier from a process standpoint. Hopefully that made sense.
Host 1 29:35
Well, yeah, from your system level, as opposed to pre configuring all of those numbers and then having to associate everything you can associate at the time that you're breaking everything apart.
Host 2 29:47
Yeah. I don't know if I yeah, this is a couple projects ago. But yeah, we had that problem with we were doing the XOR badges one year and And each badge had a unique identifier, right? Well, we had, it was like baked into the firmware. So each board had its own firmware which had that unique code in it right? Well, all those got flashed, and then we had a whole sticker list. And so you had to go find, you take the badge and then get the unique identifier from it, and then have to go search the sticker sheet.
Host 1 30:26
Oh, that's it was like, were they were they at least sequential on the security? Are they random?
Host 2 30:34
It was, when we started out, they were sequential.
Host 1 30:41
That sounds terrible.
Host 2 30:42
And it was pretty rough. And so going the opposite way is the best way basically, print your sticker sheet, and then just randomly assign those two boards, because doesn't really matter. It's just a unique number doesn't have to be an order, and then you assign anything that's unique about it to that number later. Makes it a lot easier. You see,
Host 1 31:05
We have it even easier at work, what we'll do is we get boards all the way through assembly through testing. And when it comes time to pack them into their retail ready boxes, that's when we say oh, you know, this order was for 198 of them got through testing. So print me serial numbers, x through x plus 98. And it does that and assigns them to those. And then it prints a sticker for both the unit and a sticker for the box that it goes in. And we we stick it on both things, and then you're done. And so those other two units that may be failed testing or assembly or whatever, those go to engineering, and they get pushed into the mix later on. And we always know what two numbers are to attach to what other or where they go. Because we do it almost at the very end.
Host 2 31:57
Yeah, it really depends on what you really care about tracking, right? Do you care about that? That PCV you know, it's life from the moment it hits your dock, like you, you individually label each PCB in the panel before you even assemble. Some people do it that way? Because they need that level of tracking for doing well can't remember the standard but there's a it's not it's not the any of the Nitro standards. No, it's not for ITAR either, but it's it's not the ISO, it's not ISO 9000 But it is another one. That is you basically got the you have to track the the material all the way through for for material reasons, okay. It's actually it's like actually, like, where the material comes from is what they care about.
Host 1 32:53
So have you have you heard of magic PCB? I'm sure you have I think magic PCB is a is a Kleenex brand name of this kind of technology. But you can you can purchase PCBs with an RFID tag. buried inside the F
Host 2 33:10
Oh, yeah. Okay. I, I've seen I'm looking at the website now for magic PCB. Yeah, I've seen stuff like this. And yeah, you like router, a pocket and the side of the board of the panel? Well, yeah,
Host 1 33:24
In your board files, you just define an area where you say it's acceptable for it to go over here. And then and then the PCB manufacturer just does that themselves. And then you always have tracking capability of the PCB, and you have, effectively a serial number installed in the FR for itself. And that's, that's kind of, I don't know, we talked about that a while ago at at, at the fab a long time ago. I mean, it it significantly raises the cost of the PCB, but it's pretty cool to be able to just, you know, scan and say like, oh, this PCB is this and then, you know, if every station workstation has a scan capability, you can get lock tracking history of everything. It's pretty, I don't know, it's pretty neat.
Host 2 34:09
Yeah, it's one of those. You can do the same thing with a sticker. This is more of your also you basically going beyond the sticker now because it's getting done at the PCB fab instead of at your dock. Yeah,
Host 1 34:23
Actually, so a friend of mine, got a job a few months ago. This is Colorado. So Colorado has two main industries in terms of tech, and that's medical and military. And this guy got a job at a military place doing electronics assembly. And he's not allowed to talk about what he does, but he did talk about the process, which is kind of interesting. So they have people who walk around the assembly floor. And if you have a tool that's been lying on your desk for 30 minutes and that person as they walk around has seen you not use it in 30 minutes, they will come and they'll put it up for you, they'll put it up on the pegboard for you. Like, there's like predefined areas where every tool goes, they always go there, at the end of the day kind of thing the NDS shift is you clean up and you put all your tools back in place. But every tool that this company purchases, they embed a tracking device in it, and they can actually ping. So if you lose a screwdriver, they can ping it. And the system will say where it is in the building like they can they can triangulate effectively where your tools are in the
Host 2 35:34
Game we've been playing. Yeah. We've been playing a game called GTFO, which is a four player Co Op, hor sneaky game. And there's a system where you actually use terminals, that's like a DOS prompt. And you can ping items you're looking for, and they pop up on your hunt. And they it basically just like that, it's effectively
Host 1 35:59
This. Yeah, which I don't know, that's like that it's way, way way. Next Level manufacturing that is like, the enemy cannot know that you use the number two screwdriver to screw this thing together.
Host 2 36:13
That is $500 toilet seat level manufacturer that's $5,000
Host 1 36:17
Toilet seat. But honestly, that would be super, super awesome to have in a manufacturing floor to just if you could be at your computer and be like, Oh, where is XYZ sub assembly? And you can be like, Oh, it's on Timmies desk over here, you know, oh, man, that would be awesome.
Host 2 36:38
No, I think we're in a consensus is stickers are the way to go.
Host 1 36:42
You know, one of the one of the suggestion that I've done in the past on on small runs, say 15 or less boards. I'll just, I'll just put a big square or big rectangle of silkscreen on my board, white silkscreen, and Sharpie shows up really great on that you can just write the serial number, right on that. And that makes that's works really well for low volume stuff.
Host 2 37:07
And I would say, Well, I'm gonna do with the pin, the pin guitars is I have three of Rev three. And I'm going to label them 123. And so that way, I know which one goes where and then what happens to it.
Host 1 37:23
I would have never guessed he would have done 1230. Yeah, ABC.
Host 2 37:28
I should just pick random numbers.
Host 1 37:31
Host 2 37:32
Do you have board? 83629342?
Host 1 37:36
Out of three. Yeah. That's awful. I'm sure there's there's standards around how to actually do that. Frankly, I've never run into him. No,
Host 2 37:50
No, you got to do laser etching. I've done that before. laser etching like tops of ICs. Or because that's how they put those markings on there to begin with is a laser them. So you can laser another spot on like a big IC. You can just laser a QR code on there. That's pretty cool.
Host 1 38:12
We used to do chemical etching at my first job. So we had these cylindrical devices that were made of 304 stainless. And we had a printer that would print out what are they called the film's for it. So you could you could type in your serial number, it would print out the film, you'd stick it to it and then you'd paint it on with some really nasty acid. And it would it would etch into the stainless and it's there forever. Yeah. And then and then they bought a laser and that took care of all that
Host 2 38:44
A little bit faster and easier,
Host 1 38:45
Faster, easier, safer. Because it was it was kind of funny just see people like with some pots of really nasty acid just painting and honor these things and like it was safe, but it was also still like near there's gotta be a better way to do this old school way. That is old school. Yeah, yeah.
Host 2 39:04
We always go back to a punch set.
Host 1 39:10
You know, we use we use one of those at work actually. If we sew something we sell numbers on it, we sell replacement panels. So if somebody gets like if their panel gets bent or scratched or something like that, they can purchase a new one. Or if they have an old version of ours that silver and they want to purchase one of our new ones, it's black, they can purchase one of those. But just for quality tracking purposes, we actually flip all the panels over and we hammer in a punch of the letter R onto it for replacement. So if we ever get one back for an RMA, we know that they took it all apart, the customer did and that voids their warranty. Ah, gotcha, interesting. And we're punching into 5000 series aluminum like you could I mean, you could basically flick it with your finger and Punch the letter into their software.
Host 2 40:05
When I was analyzing some aluminum, when was that? Last winter? I think punching anodized aluminum is pretty tough. Is a 6000 series aluminum anodized. Trying to do a good clean punch to
Host 1 40:22
Plow through the anodized layer. That's still pretty hard.
Host 2 40:25
It's pretty tough. Yeah. Especially, I was surprised how tough it was, especially since I did it in my backyard.
Host 1 40:34
So nice. So last week I was I was talking about doing some repairs. And on Thursday, was it Thursday, I think I think it was Thursday, I posted up a handful of pictures of the Roland TR 606 Drum matrix that that was messing with on there. And really, I posted a bunch of pictures just to show people that like, It's impressive how much they shoved into a small envelope. Given that it's all through hole. And someone even commented, apologize, I don't remember the name, someone commented where it's like, they, they didn't think that it would all fit in there. It's like no, no, like knit, it all just cram into this little tiny space. And it's like there's cutouts in the boards that fit through pockets and all kinds of stuff. It's I don't know, it's ridiculous to what went on with it. But effectively, so this drum, this drum matrix has a row of buttons that you can basically you select what drum you want, say kick drum or whatnot, and you have 16 buttons in a row. And you can select at which buttons you want that to be active. So as time goes across, whichever buttons are active, it'll trigger the drum. So your drum, you can create your beats based off of those 16 buttons. And with this unit, everything functions, but it every single step triggers every single drum. So it's like somebody with sticks, just hitting everything. Every time
Host 2 42:02
The clock goes, Yeah, someone's just kicking the entire drum set off the stage. Yeah, yeah,
Host 1 42:07
Exactly. So obviously, there's something wrong with it there. So one of the things I noticed with with the device, every single analog section works like you turn up the volume on every drum module, they all come together and function properly. And a lot of the functions that seem to be processor related are also working. So in other words, so when you press a button to trigger a drum, the the light actually flashes. So what that tells me is, I mean, I know under the hood, there's not some kind of like timing circuit that could cause an individual LED to blink. So if something is flashing at a particular rate, that means the processor is thinking about flashing it, which means the processor is actually running in some capacity. So it doesn't seem like the issue was in the processor, and all the analog circuits were working. And all the analog circuits get their triggers directly from the processor. And since everything is triggering all at once, that means that the processor for each step is sending out triggers for everything. The one thing that just doesn't flat out work is that when you, you when you select buttons to actually set up a sequence, it doesn't accept the sequence. In other words, it doesn't remember what you press. So I started looking at the RAM. And that's sort of where I've pointed my finger. I don't I don't necessarily have anything to prove that at the moment. But it just seems like everything in the whole circuit is working except for storage and memory. Because you can't write anything. You can't recall anything. And it doesn't remember anything. But it functions. So that sort of
Host 2 43:46
Doesn't remember one thing, it remembers to fire everything at once.
Host 1 43:51
Well, what I think that is actually is the RAM is busted, so it's going to ram and it's like, Oh, give me everything on every step. Yeah. So every time it accesses RAM, it probably gets an everything command. So
Host 2 44:06
Everything's a one. That's what I'm
Host 1 44:09
Thinking. I'd really I have no clue on that. It's kind of hard to tell. I mean, I guess like I made the elf a little. I didn't actually probe the lines from the RAM. So I don't know what it's getting from there and
Host 2 44:21
Not LDA DLA you can do a the Rambis, I could
Host 1 44:25
Or I could just do what I did is look up what the RAM wasn't purchased those those, those chips, which this thing is old enough that it had ram with part number UWPD 444 C, which is obsolete, but apparently there's a drop in replacement. It's just an A P dip 18 package, and it's the 6514 s ram. So I purchased two of those, and I'm going to pull them off the board and shove them in and you know, fingers crossed, that's the only issue there. So everything else seems to work pretty cool. those are those are fun repairs, where it's just, it boils down. Like, to be honest, repairing the digital stuff is more fun than the analog stuff because what the analog stuff is, something's wrong. A lot of times it's not wrong. It's not binary, wrong, like binary wrong means it just doesn't work with analog stuff. Like if it's wrong, it's like drifting or it's going out or it's started to get noisy or something. And those are always way harder to fix than like, Oh, this one thing is broken. Replace it, and you're good.
Host 2 45:33
And well, and in some cases of that, you want it to do that. It just drift too far out of spec.
Host 1 45:42
Yeah, but that's like, you have to then know the spec. Yeah, exactly. Spec what digital is work or not?
Host 2 45:52
The what they call that back in the day. Chip was chip bending. Is that a circuit?
Host 1 46:02
Circuit? Circuit? Bending,
Host 2 46:03
That's the term? Yes, sir. Yeah, hey, I'm purposely make the analog section. Not good.
Host 1 46:11
Oh, well, yeah, with circuit bending. Like the original way It basically started is you get a wire and you just start touching things together and see what the result was. Actually, what's funny is, that's the first since I knew this entire thing ran on nine volts, I knew I wasn't going to hurt myself with anything. So, what I started with was just touching it everywhere and pushing on things and rubbing my fingers everywhere to see if I could affect the circuit in any way. Because a lot of times with through hole stuff, you hit a cold solder joint and just touching it fixes it right
Host 2 46:44
Adding enough capacitive like bend to it
Host 1 46:48
Right right well and that actually, I can. So this circuit has it has a gosh what is it has a Schmitt trigger oscillator in it that is what controls the tempo and and more than the tempo just it controls the oscillator for the processor, which is highly sensitive to two extra capacitance so you can actually overclock I'm doing quotes you can overclock this circuit by just sticking your thumb on the PCB spots but it does work. Hopefully those will come in I ordered those chips a few days ago and I still haven't even gotten a shipping notification but then again, it's Thanksgiving week and yeah, everything
Host 2 47:38
Yeah, well I ordered a replacement. Complete side tangent but same thing I guess ordered a cuz I've actually ordered it on on the eBay.
Host 1 47:49
I don't know I ordered this directly from IC website jamika Angelico or something like that. Okay,
Host 2 47:55
Nevermind, then. My story is not relevant anymore. Do you have another topic? Are we done for the day?
Host 1 48:06
Oh, I guess we've played through your topics anyway. Alright,
Host 2 48:09
Yeah. Remember we did? We brought up Ben at the very beginning of the pack.
Host 1 48:13
You know, I just saw I saw three bullet points. And you're totally right. We talked about three but no, I think
Host 2 48:19
I cheated. I talked about to mine in the row.
Host 1 48:22
Oh, I see how this is just trying to be that extra? Yeah. No, I think I think that's that's good for me now.
Host 2 48:32
Okay, so that was the Mac Feb engineering podcast where your host Parker Dolman and Steven Gregg. Let everyone
Host 1 48:39
Take it easy.
Transcribed by https://otter.ai