Minor correction from last week. IEC 60950-1 was withdrawn in December of 2020, and has been replaced by IEC 62368-1 (Product Safety testing standard for “Information Technology Equipment” and Audio Video Equipment)
- How was it?
- Favorite thing
TI soic oddities?
- Large indicator mark
- Laser marking has the TI logo but the line goes through the dimple
- Very offset marking
- Sharp cut body?
- Based on the original core of the SSM2040 (Solid State Micro Technology for Music )
- Voltage Controlled Multi-Mode Filter
- Four variable transconductance cells with i-v buffers
- Temperature compensated
Interesting reverse engineering tactics?
- In regards to customer projects?
- How often do you have to reverse engineer products or designs?
Special thanks to whixr over at Tymkrs for the intro and outro!
About The Hosts
Parker Dillmann is MacroFab's Co-Founder, and Lead ECE with backgrounds in Embedded System Design, and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. He also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components. Parker graduated with a BS in Electrical and Computer Engineering from the University of Texas.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Host 2 00:11
Hello, and welcome to the Mac fab engineering podcast. We're your hosts, Parker, Dolman
Host 1 00:15
And Steven Craig.
Host 2 00:17
This is episode 289.
Host 1 00:21
So a quick correction from last week, we had talked about IEC 60950 dash one. Yeah, it's a safety regulation document for UL, basically.
Host 2 00:34
And in regards to creepage and clearances,
Host 1 00:38
Well, among a lot of other things, and I had suggested that if you were interested in biofuels certified go to this document. Well, I was incorrect in the fact that 60950 dash one was phased out at the end of last year and has been replaced by 62368 Dash 150 testing standards for information technology equipment and audio video equipment. So just a quick little correction on there. Check out IEC 62368 dash one.
Host 2 01:15
So is there any major differences between that document and the older document?
Host 1 01:20
I'll be honest, I have no clue. I don't know. It's just I was I was corrected. So I thought it was it was worth mentioning the so I'm sure there are plenty. Just if you're if you are searching for UL standards, I'm sure you can find the old 60950 and reference it out there. But that's not what newer products are going to be referenced by.
Host 2 01:46
I'm trying to do a quick Google see what the big differences.
Host 1 01:50
I would not be surprised if there's like a big list of every single difference between them somewhere out there. Or, or I'm sure there's a list that you can pay for that tells you all the differences between them.
Host 2 02:01
Apparently one of the things does aims to create a more future proof document by requiring that manufacturers show known hazards have been considered. And their products have been built to be safe for use in their intended environments. Interesting, basically gives designers more flexibility in their design and valuations. You have to read up on that. That's a DigiKey. Article actually.
Host 1 02:29
So you just got back from DEF CON, right.
Host 2 02:32
Hold on, Casey, eight AP F. posted something from mighty ohm on Twitter. Oh, unfortunately, it's not a link to the differences. Just two milestones. Twitter accounts. got excited there for a second. Oh, yeah. So yeah. So it might sound a little bit tired.
Host 1 02:57
And when we say just got back, like, it's just, yeah, like,
Host 2 03:01
I got off the airplane like an hour and a half ago. You're in Houston. Actually, I'm not that tired. I'm just more I'm more just worn out just from five days of DEF CON.
Host 1 03:14
It's a lot, right.
Host 2 03:15
It's a lot. It's a lot of fun. Had a blast. This is my second in person DEF CON. So my third DEF CON because 28 was all virtual DEF CON 29 which was this year's was like half virtual half in person. It was interesting because my our first DEF CON was DC 27 And it's like, it was like bananas like you cannot walk around how many people are there?
Host 1 03:42
It was like 50,000 people right? Yeah, something like
Host 2 03:45
That. I actually kind of liked DC 29 A lot more. Because there were just less people so like when you wanted to go see something or talk to someone like you could actually do it. It sounds like with the during the closing ceremonies basically, DC 30 which is next year's is going to be all in person unless something like COVID 21 shows up or something like that, you know? Because who knows at this point
Host 1 04:14
Host 2 04:16
Yeah. So okay, so if anything, basically I think their plan is to have a DC 30 all in person. So we'll have to see. But they already announced the dates and stuff for it. Which means I need to go get my my hotel for that already. My favorite thing for Def Con this year. Oh man. I went to more talks or well and watch more talks. So that was a lot of fun. I didn't get to do a lot of the video village stuff like I did in DC 27 I think that's one thing I'm gonna do a lot to do more of and 30 is like the car hacking village. Like I think I'm just gonna like post up in car hacking village.
Host 1 05:05
So I guess. Okay, so for those who don't really have an understanding of DEF CON, give it a give a description of it.
Host 2 05:12
So DEF CON is a security conference, I guess is a good way to put it. It's mainly aimed at at software security, hackers and that kind of stuff. But there is a small and growing like hardware security group. And what I do is, if you listen to previous podcasts, the AMA XR group, I'm part of that we build badges, electronic badges for hacking and the lols, basically, for DEF CON, well, not for DEF CON, but their unofficial DEF CON badges.
Host 1 05:54
Well, and it's not just you, there's like tons. There's hundreds of groups, right that do, maybe not hundreds, but there are a lot of groups that make badges for it.
Host 2 06:05
Yes. So, yeah, time to start planning the batch for DC 30. So interesting enough. Art DC 29 badge, we actually had to shift gears back in like, April, because of the chip shortage. Basically, the manufacturer for our microcontroller, like straight up ghosted us in April, we were like, Okay, we need to buy like 1000. Right. Gone. And so we had to switch gears, which is why we went to that this discrete design. badge, which was a, a 555 timer, that blinks LEDs. So actually, we've been technically working on the DC 30 badge for basically a year half and a half already. So because we're going to roll but we didn't, what we were planning on doing for 29 We're going to roll that into 30. And basically just have a huge banner. Year have a badge for, you know, 30th anniversary at DEF CON. Yeah, so
Host 1 07:13
When I guess you guys, I mean, that's kind of nice, because you guys get a little bit, a little bit more breathing room, right? Because you're, you have a lot more done towards the 30 badge already.
Host 2 07:23
That and we are our group, we tend to, like get stuff done early and make sure stuff is working and, like order everything it's like, and basically we can make sure we're done like a month ahead of schedule, like a month before DEF CON. So we have like a month of like, just make like smoothing everything out and then get to take a little break and then get to go to DEF CON.
Host 1 07:51
So then, okay, so you were mentioning villages. So let's go back to DEF CON real quick. What's the village at DEF CON?
Host 2 07:57
So villages is kind of like a subgroup, I guess at DEF CON.
Host 1 08:02
Lies almost like a just a room.
Host 2 08:04
Yeah, yeah. It's a room or a section of a room. That's like just a group of people that have this set of similar interests. And they have their own talks, and their own events and their own challenges. And that kind of stuff. I'm just trying to absorb does DEF CON because it's just, it's just crazy.
Host 1 08:23
And it's huge. It's huge.
Host 2 08:25
So I'm just trying to lots to in person def cons, I was just trying to like, figure out how the con worked. And I think I kind of have a grasp on it now. So I'm hoping DC 30 I like post up a couple villages and just like do that kind of stuff.
Host 1 08:46
But he'd be a little bit more active as opposed to just like passively suck soaking everything in.
Host 2 08:50
Yes. Yeah. Exactly. Though this year I think on I think it was Saturday. I didn't really want it because Saturday was there was a bunch of smoke from we're what is on fire right now. Like everything.
Host 1 09:08
It's somewhere in the Pacific Northwest, I believe.
Host 2 09:11
Okay. Anyways, all that smoke actually made down to Vegas. So I really didn't want to get out. It's like you open up your window and you couldn't see like the hotel across the street. And so I did is I just watched all the talks on the TV. And unfortunately you can't see oh, like the village talks. Because those aren't recorded
Host 1 09:36
The because there's village talks and then there's like mains? Like yeah, it's kind of like
Host 2 09:40
Track. Yeah, the call like track one, track two track three talks. And the great thing is though, if you're at the hotel, actually this year in last year, too, they were all on Twitch like they have like six Twitch channels that you can actually just watch them too. But I was watching them on on the hotels too. v. And I just spent basically a day and a half working on the badge, not our badge, but the DEF CON badge and solving those puzzles. And this was the first year I actually finished it. Oh, cool. Well, actually, me completed it because the last two dc 27 and 2828. I started. And I just couldn't wrap my head around the because it was like an audio puzzle kind of thing. Oh, yeah, it was like, it was on a cassette, right? It was on a cassette tape. I didn't really get far with that one. And 27 I was just so I was doing so much Anna XOR stuff I just never got around to basically working on the badge then. But 29 I'm like, Okay, I'm gonna spend some time just listen to the talks. And I'm gonna do the badge, and I completed the challenge. So that was a lot of fun. Cool. Some of the some of them were pretty rough. And then only one challenge because like, there's like, there's several parts. Basically, it was like a series of puzzles that just built on each other. And then one, there's a ginormous cipher that is based off of it's the same cipher they use in Nicholas Cage's national treasure with like, the, like the numbers, it's, I can't remember especially Oh, but it's like a book cipher, basically, like, it just has, like, you have a big book, and then it's referencing page blog line, line nine, character 26. And so you go and go, Okay, that's E, and then you write e down, and then you go down the list. So I've once I figured out that was that cipher, and I like looked at, like, it had to do that for like, like 80 characters or something like that. And it was like, through all these different, basically was going through the, the programs, like the PDFs of each DEFCON. And I almost got to the point where I was like, Oh, I could probably just write a Python script that just scrapes the text out of all these PDFs, and then just feed in that and I'm like, you know, already. In my mind, I like I've solved it. I'm just gonna, like, look the answer up for the, you know, the key. Yeah, I knew how to do it. I'm just like, man, that's gonna take me like, two hour, two hours of like, manually doing it or maybe an hour of code units. like, Ah, screw it. I'm just gonna. I'll cheat on that one. So
Host 1 12:34
That's, that's like a, I don't know, like a 10% cheat.
Host 2 12:38
Yeah, I already figured out the solution. I just, you know what, I didn't do the math. You didn't
Host 1 12:43
Do the legwork. Let's just do the legwork. And what's funny is like, you didn't do the math, nor did you write the code.
Host 2 12:51
I just I'm like, Okay, I know how to make this. Solve this. So yeah. Oh, and there was like a discord. For DEF CON. Apparently, there was one last year too. I just didn't know about it. Which might have helped with the audio tape cassette challenge. But um, yeah, that was it was a lot of fun with that the score two.
Host 1 13:18
Last week, we talked a bunch about the badge that you guys had designed to take out there. So how did everything go at DEF CON? How was it received? And what were the results of it?
Host 2 13:29
Um, oh, you can I mean, the annex your badge? Yeah. Yeah. My people loved it. It was it was interesting, because there are three different groups or thought groups of like, figuring it out. Okay. Some people were, some people were like you, Steven, like, foreseeing a schematic, and like showing us like schematics. And like, is this how it supposed to be? And we're like I don't know. I don't know.
Host 1 14:01
Too much vodka. I have no idea. Yeah.
Host 2 14:05
Some people were going through the cipher. And then the cipher was, that's definitely the easiest way to do it. And then the other way was, people were brute forcing the cipher because you can also attack it that way, basically, mathematically analyzing the symbols and figuring out the cipher that way. So it was great. People really liked it. We actually had so this badge this year was like Matt Damon themed, like heavily like the packaging was damaged nidos so we had big cardboard cutouts of Matt Damon, and that was definitely a lot of fun. We actually like our group got a like a call out at the Uh, closing ceremonies for Matt Damon.
Host 1 15:03
So Oh, that's cool.
Host 2 15:04
There's a lot of fun, basically, for shenanigans and that kind of stuff. So that was a lot of fun.
Host 1 15:12
Did you? Did you see a lot of people with completed soldered badges?
Host 2 15:16
Um, not as many as I would hope. But I did see some. So I think most people were, they figured it out, but they just didn't solder it up yet. And so they were just wearing the bear badge.
Host 1 15:32
It kind of exactly what we're talking about it like, it's sort of like they got the point and maybe didn't do the legwork.
Host 2 15:38
Yeah. Because it is a lot of soldering. Does Yeah, I It would probably take about an hour for someone to solder only if you actually solved the cipher. And then all the components, it probably takes about an hour to solder together. And the hardware hacking village probably just didn't have the bandwidth of soldering irons to handle that. Because we thought we probably gave away like 400 500 badges or something like that. Oh, well. Yeah. There's a lot of fun. That's walking around with a matt damon cutouts and getting people to take pictures and put stickers on Matt Damon, was a lot of fun.
Host 1 16:18
So How about how about some of the talks, anything noteworthy anything? That was really cool.
Host 2 16:25
One that, unfortunately, I missed, there was a talk about Apple's new chip that's in there, like the cryptography that's in there. And one of their ICS like someone was trying was working on reverse engineering that fortunately, unfortunately, I missed that talk. But the one that I liked the most I did catch. I'm trying to recall what the company's name was. It was a company that was moving basically analyzing, like, if you're making a I think it was a DNS company or something like that. Like, if you're What country is the best for, like privacy? And like going through, like laws and, and? And all that kind of stuff. I thought that was very interesting talk. Like, what what constitutes a telecommunication company in certain countries and what's not? Yeah, all the talks are on on def cons website. So you can go check that out. So I need to go find that that apple one.
Host 1 17:38
I'm curious. It was a what was the result of that one talk? What was the most? I don't know, secure? Country? Oh,
Host 2 17:44
They ended up moving to Switzerland? Yeah. All right. Because like also the things that we're talking about, like treaties, and that kind of stuff, because that also impacts local laws and that kind of stuff. Then there was the talk about the mag, so So DC 27, it used near field magnetic coupling to communicate between the badges. So you had to be really close. And it's not like near field. And like, it's not like NFC. But it's like, it's a magnetic couple. And so if someone was trying to reverse engineer how that worked, because basically, in the, what he discovered in the DC 27 firmware was, there was a export, he can use. Basically, if he was able to push more data over the push more data over over the communication, he can actually like remote execute code, through a wireless attack, basically onto the badge. And so he proved it out by basically, like, just manually, like, hooking up UART to it and dumping into it into the code. And so then it worked. So he's like, Okay, now how do I do it from like, a wireless perspective. And the problem with that, that chip, the integrated circuit that runs that, there's like, no data sheets that exists, and the company wondering, like, talk to you to like, make, like, how does it work and that kind of stuff. And so he reverse engineer that entire, like, protocol and stuff. Oh, wow. Yeah, it was that was pretty intense, like, trying to figure out like, because it's not like just regular like, oh, it's zeros and ones like there's like a rotating mask. And the symbols are rotating as well. The for to obscure the signal. Because he's like, when you first look at it, it just looks like straight noise. And it's like, okay, how does that actually talk? That's cool. Yeah.
Host 1 19:59
Yeah, that was a real The neat badge we had, we had Joe Grand on the talk about the design of everything in there, right.
Host 2 20:06
But yeah, how? Basically like, I think I think what Joe Grand said about that was the company that made that IC basically gave them like a chunk of C code. And I was like, this is how, like, use this right feed it the data and it will work with the chip. I think that's how it was I can be misremembering that that was like two years ago.
Host 1 20:29
So So So you've already kind of mentioned this a little bit, but DEF CON 30. What do you want to see? And what do you want to do?
Host 2 20:37
The I want to do more village stuff for sure. I'm hoping to see more hardware stuff this year, the hardware stuff, kind of, but was tampered a bit. A lot of the villages like either canceled or, or went virtual this year as well. I'm hoping basically car hacker villages like they're in full strength. And I will hang out there and wrench on things. But yeah, that's what I'm looking forward to. I want to do more village stuff. We have a big project. And an extra has a big project for DEF CON 30. That's about all I can say about it. It's uh oh, it's also the name for it. That's public, called Project snacky.
Host 1 21:22
Snacky. Right. I think you mentioned that last soon.
Host 2 21:25
Yeah, that's all I can say about it yet. Okay, uh, probably started leaking out information about that as we work on it. Because that's gonna be kind of like a side project. So I get to talk about
Host 1 21:39
It. Okay, you know, actually, because of the badge this year, and what we talked about on the last episode, a lot of a lot of our listeners reached out and were saying like, Oh, I wish I would have gotten this badge or whatnot. So when it comes down to NX, or and these kinds of badges, or even DEF CON, and all the things around it, I guess let's start with with an odd X or what's the best place for our listeners to go at about these kinds of things so that they could participate next year?
Host 2 22:10
Follow and XOR on Twitter. It's 100%. What it is, everything that we do is like on that Twitter.
Host 1 22:19
And that's an exclamation point.
Host 2 22:22
Yes. Okay. I think it's actually because you can't have explanation points. I think it's a n d n. X or
Host 1 22:29
Oh, okay. Yeah, I wasn't aware. Okay. Great. And then And then yeah, I guess Parker throughout the year might if we all cross Leka. Some information here on the on the podcast?
Host 2 22:41
Yeah. I try to be pretty tight lipped. But the project snacky is going to be a pretty big project. And I'm really looking forward to
Host 1 22:50
It. So hey, with how many people listen to this kind of stuff and already go to DEF CON, I think you should talk to the anon XR crew and ask if you can leak certain information. Use this. I use the podcast as a vector.
Host 2 23:06
We kind of did that with last week's episode. So
Host 1 23:10
I look but you know, I think it'd be cool to plan a little bit.
Host 2 23:13
Yeah, yeah, for sure. I just All right. Well, I can't wait to sleep tonight. Honestly, I did get to sleep more than I have in like the past like 10 weeks, because you probably also
Host 1 23:27
Drank more than you had in the past 10 weeks. Because that's that's one other part of DEF CON that. I mean, you're in Vegas, right? You're just walking around all the time. Like it happens, right?
Host 2 23:41
Work a cipher into the podcast. Ooh, I like that idea. DJ, work a cipher into the podcast. That might be way too much planning.
Host 1 23:53
That sounds like work. Okay, so speaking of work, he see that segue there. I ran into something interesting today that I haven't run into before. And it's funny because like I was, I was thinking of things for the topics to talk about tonight. And this happened today, I ran into something that like screams wrong part or counterfeit part today. But I still am not fully aware of of the whole situation. So so I'm at work where for most things that we run on a on our pick and place machine. The the operator before we do a large run, they'll they'll bring the first 123 boards to me and I'll do a solder inspection. And I also check component rotations and things like that. So we have basically our pick and place operator. Loads of the machine checks all rotations of things, and then myself or someone else comes over there. We check out Free feeder, we check it, we check rotations. And then once it comes off the machine, we do another check. It's it's worth spending the labor time upfront than the labor time in the back because it's way worse labor siccing stuff. So So I, a, and I'm scanning over the board, everything's looking great solder joints are looking awesome. Like I've got no problems with it. And then I stopped on an soI C eight component which is just a it's a it's a jelly bean op amp. It's a to 07 to just classic op amp been around forever. things. But the package looks super weird. I've never seen an SOC eight looks like look like this. Now it fit the the pads fine. Nothing, nothing wrong there. But okay, so it's branded as part. All the laser marking is way offset on this part, almost to the point where like it's hanging off the edge. In fact, I think it may have even been off the off the edge of the component. There's also a big laser across it, which I think ti does that some of the time. So like that's, that's fine. That's like there's no one side. Yeah, there's a pin pin one side. Yeah. There's also the big circular dimple on on pin one, but it's normal, like it looks quite a bit larger. And that line goes through the dimple, which you don't normally see. No,
Host 2 26:28
Usually the laser doesn't go through double.
Host 1 26:31
Right. Right, exactly. And it looks odd. And then there's there's one other characteristic about this that that she had never seen before. And maybe it's a new form of manufacturing. Or maybe it's something weird, you know how? Okay, so that's why I see eight, if you look at the sides of the package there chamfered. Right. And I'm not just talking about like a pin one. Besides have a chamfer on them. Yeah, for manufacturing of the actual body of the IC. And then where the two pieces come together. There's usually like a flange, or like some
Host 2 27:06
Reduction. The injection mark,
Host 1 27:09
Injection mark, right? Yep. Okay. So on the two sides of this component where the legs leave, those chamfers exist. And there's that there's that mark, but on the the I guess the front end and the back end, the sides that are on pins one, four and five. It's straight flat, like it was cut with a? It's like it like think of if you could extrude Yeah, so I see eight and then just chop them like chop Snickers bars, right? Yeah. I've never seen that before. And it was very clearly cut, it has cut marks on it, like a diamond, diamond blade came across it. And on top of that, if you look at the side of associates, sometimes you can see some of the pin frame that Yep, the legs with this, it's very clear that pin frame you see like a very exposed side to it. So a lot of these characteristics of this does not screen Texas Instruments to me. It screams like some something else. Branded something else or counterfeit or whatnot. So of course, like I hold the production line isn't like what is this? I've never seen this. And lo and behold, we did buy two new reels of these TLO seven twos with them and soldered them on to another product that I know uses the n o 72. And it's functioning. I mean, it does exactly what I expect it to nothing, nothing unique there. So, you know, I've only I've only tested one I want to I want to do a little that. Because it is like wait, whoa, like this is very odd. It doesn't match the datasheet. The datasheet has no call outs about those kinds of things. Although most of the time the datasheet doesn't call out like the body should look exactly like this. Unless Unless, you know sometimes, you know, laser marking, it'll be marked here. It'll be marked this way. This particular data sheet doesn't have specifics on that. But but the fact that you know it has it has the pin one indicating line going up pin one dimple that's not very common characteristics. The fact that all the lasering is offset and I checked like multiple of these things, and they're, they're consistently off. It's not like I just got a bum one and then iced package just screams like, perhaps because of the chip shortage, Texas Instruments had to do something different or do something a little bit quicker or just implement something that maybe they do have a new like, Chip. Just you know. But you know, it's funny because like you just don't know it bucks the standard I've never seen This before so it's like, well hold production. I just want to make 100% Certain on this. Yeah, I
Host 2 30:03
Would, I would contact ti with the batch number or that real.
Host 1 30:10
Yeah, yeah. Like I said, I'm happy because it is functioning, you know?
Host 2 30:14
Yeah. So I would be like, is this post is this new supposed to her this? Yeah.
Host 1 30:19
Is this as seen that hit me up? If this is like a standard in SLIC production? I'd like to know that I've just never seen it before. You need posting pictures? Yeah, yeah, I you know, it's funny. I got home from work today was like, Why did I do that for the podcast? I'll ask one of the guys at work to take a picture tomorrow and send it to me.
Host 2 30:43
Awesome. Yeah. I'm interested. I want to see what ti says. I bet you so would you buy him from?
Host 1 30:52
Yeah, I didn't purchase it myself. So I don't know yet. I just, it was at the end of the day was the last thing I was doing. And it was like, Oh, shit, we gotta we gotta figure this out real fast. So probably like I said, like I did. I did the brute force off another one from a known working board. soldered this on. Okay, I can at least sleep tonight because I know that that one is working.
Host 2 31:17
Yeah. Yeah. Hopefully not eBay.
Host 1 31:21
Oh, no, we don't we don't do that level.
Host 2 31:27
Oh, man, I had to do that for I actually, it's funny is for peanuts, or when I was getting this current. That whole, like 250 unit run. I tried to buy parts on Amazon for current sensors, and they ended up canceling it. Oh, shame. Yeah. I think this basically like they thought they had some in stock, they just didn't. So though we're going to completely different current sensors just for an brand. Just to get away from Allegro, basically, Allegro just can't seem to keep stuff in stock right now.
Host 1 32:01
So you're doing a revision on the board? Yeah, we're just rolling
Host 2 32:05
To it for C. And also, like, that's those are also the only non deleted packages on the entire board. And so you have to like X ray them. We're going to move to current sensors that are needed. And so we can just get away with just doing aeoi and be fine with it. So yeah, that's a good move. Yeah. Yeah, basically,
Host 1 32:32
Maybe not necessarily cheaper on your end, but for the manual, the whoever's doing PCBA I'm assuming it's macro fam. They their life is a little easier. Now it's
Host 2 32:43
Cheaper for me to Oh, is it? Yeah. Do you know?
Host 1 32:48
Does does macro fab inform you? If if you need to do X ray, is there like an x ray service? Or is that just kind of included? If needed? Be?
Host 2 32:57
Um, little both I guess.
Host 1 33:02
I It's per case basis.
Host 2 33:05
Yeah. per case basis. Yeah.
Host 1 33:07
I think West X ray from Mexico. Yes, you
Host 2 33:10
Can. Cool. Yeah. I think for this we probably x rayed. I think we actually like basically only like 10% of them. Because it's not that crazy of a package. It's not like a BGA. Right. So basically, we decided 10% was probably fine. The X ray, we didn't have the X ray, all of them.
Host 1 33:32
That x ray must a long
Host 2 33:36
Runs about 60% of the day. Okay? It's not, it's not an automated inspect. X ray. It's a manual. So it's mainly for spot checking. Got it? I would love to get it in line x ray. That'd be awesome.
Host 1 33:54
Oh, yeah, that'd be crazy. Just just buy an old TSA X ray.
Host 2 34:02
I mean, oh, man. When we were first starting Mac five, we were thinking about getting those like really sketchy dental X ray machines that had like a lead box.
Host 1 34:11
Now those are those are sketchy off at eBay.
Host 2 34:14
Where you can still buy him.
Host 1 34:16
So I had a my first company we did buy one of those and they weren't. But they're sketchy. Yeah,
Host 2 34:23
Because I was like, it's like this like it's basically like, you know, dentists in the heat in the in the dentist, your dentist moves the X ray machine towards your face. And it's got like that like tube that just shoots out all the beams. It's like just one of those except that it's just like free floating like just has a plug in and oh, it's just the fact that it has no Interlock System. That's what's sketchy.
Host 1 34:54
What we had well first of all, okay, so like it was it was in we didn't have I have it in the engineering lab, it was on the production floor because they used it there. But it was used by, like trained individuals. And it was in the corner away from people. So it wasn't right next to the secretary machine. That was blasting them or anything like that. But but it was very similar what Parker saying like it's just a box with a tube on it on, like, the other end of the table was a lead plate, basically, at a school somewhere in between. You can shower sketchy then, right? Yeah. And the thing was like, you could put it on continuous operation, like it wouldn't like, like, it wouldn't do like pulsed operation, you just flip it on and turn the juice up. Yeah, one of the guys one of the production guys, whenever the abs are not mechanical engineer, the manufacturing engineers. They had purchased it and they they got it all set up when it was brand new and fresh, and he's like, hey, it's working. Any any random stacks, and he put a picture on on my desk, and it was freaking hand like, oh, no, yeah, no straight up, and he can see all the bones and everything. I'm like, Dude, no, no, no, no. No. Yeah, well, I mean, so it was working, I guess. Oh, that's scary. Yeah. Yeah. You know, what was really interesting about that. Okay, so we had an entire safety team at that building, there was the building. He's in it, and there was almost 500 people in that company. So there was there was an entire safety team. And you had to pass everything from these guys. If you purchased latex gloves, you had to give these guys a receipt. And they had to, you had to have the MSDS like these guys were were down to everything. I never once heard them peep about the fact that there was an x ray machine that was like an open frame x ray machine and like one of these that never understood it. Maybe they didn't. Maybe they didn't know. Like I
Host 2 37:02
For those that are listening to the podcast, I am shaking my head really hard.
Host 1 37:07
When there's just a thing like that we were really safe in that building. I mean, it was legitimately like, everything was really well done. Except for that. I don't know. Yeah. Maybe they didn't know. I'm sure they do now.
Host 2 37:22
I hope so. The fact that the person tested it by just sticking his hand in the beam. Hopefully he's okay.
Host 1 37:29
Oh, he's fine. I mean, I know he did it quickly. He grew three thumbs. Let's just put it this way. Don't do that. Boys and girls. Never.
Host 2 37:39
Never know, ours is a big, professional machine. So
Host 1 37:44
Yours closes right.
Host 2 37:46
Yes. That's Yes. The machine is like the size. It's like five feet by five feet by five feet. Yeah, it's big. It's a big machine. Yeah. What's what's interesting about that machine, though, it has a little window. That's like, you know, like an oven. It's got like a window you can see like your, your your turkey cooking in there, whatever. It's got one of those. It's a lot smaller. And, you know, led lines and all that stuff. So X rays can't bounce out. But it's like, why is that even important? Like nothing happens when you put the board in there. Why do you need to view it? I just that never made sense to me.
Host 1 38:28
Yeah, I guess once you close the door like the magic happens, right?
Host 2 38:33
Yeah, but it's like you don't need to observe your board getting x rayed.
Host 1 38:38
Not gonna see anything, right? No, well, hopefully you don't see anything kind of so so I got a quick, quick little tasty chip that I've been playing around with this week. And I thought it'd be cool to showcase so I went to Google and I typed in tasty chips. And there is a company called tasty chips. And I didn't even know this they make Eurorack synthesizer stuff. Which is interesting. They're from the Netherlands. I don't
Host 2 39:10
Count but hey, there they got some nice looking equipment.
Host 1 39:13
Yeah, they totally do. Yeah, their stuff is really clean. So go check out tasty chips. Who knows maybe we inspired them? Or maybe it's just coincidence.
Host 2 39:23
I think it's just by coincidence. You know what they they are not using our naming scheme. So no, we did not inspire them. So the naming scheme for tasty chips was your our first chip would just be m the second chip will be m m, etc etc. And you would pronounce it as you go. So for like the 12 chip it'd be m whereas first one just be
Host 1 39:55
I don't know how many of you know we back of our chat tasty chat for probably We all like by now, right now. Yeah, something like that. So actually, well, I want to I want to call it this one chip, but I want to call out the kind of the whole business that makes a handful of family. It sounds semiconductor. And the chip, I want to call out as the SSI 2140, which is a voltage controlled multimode filter. And yeah, it's point is in audio, and very specifically, synthesizers. Plenty of uses, you could use this outside of that. So it's a, it's based on the original core of the SSM 2014, which is, which is solid state micro technology for music. Basically, it's a older audio specific or synthesizer specific ICs. SSI is kind of the evolution of that. And they've basically taken those old designs and add a whole bunch of bells and whistles and brought them into modern manufacturing. And so this two mode filter has four transconductance blocks inside of it, each of which has its own buffer. And then it also has a VCA that allows you to reconfigure things and give you a whole bunch of different features. So what you can buy by how you configure these four constraints, conduction blocks, you can get to pole or four pole High Pass or low pass, you can get all pass filters, you can get state variable filters, you can get a whole variety, based on how you basically tie passives to this and reconfigure it, and all of it is voltage control. So if you want to swing the cutoff frequency of any of these filters, you just send it a voltage boundaries. And depending on how you set up your Rs and your C's, you can swing your your window around on it. Now, it also has some other features. First of all, it can be voltage, sorry, it can be volt per octave. So calibrate it, you can get it such that a one volt change on the control input gives you an octave difference in the in the cutoff frequency. There is the ability also to tie to make it temperature compensated. And that's pretty legit. So it won't drift around with temperature. But they also make it pretty easy where if you disconnect some pins, it just it'll drift around. And some people might want that has a really cool like that VCA that added extra fifth block that's in here allows you to do what's called cue compensation. So you can you can arrange this filter such that it has external or not External links back back to the input. And that allows you to emphasize frequencies at the cutoff. So you can do resonant filters. And one of the problems with traditional resonant filters is as you increase the feedback, you start passband on there, so you you end up turning a low pass filter into a like a peak bandpass filter. Well, if you do que compensation with a voltage controlled amplifier, at the at the very end, you can keep your passband flat frequencies at your cutoff frequency. So there's there's a lot of cool little things in there. So to get a second, check out the SSI 2140 over at sound semiconductor.com. Also check out the rest of their chips because they've got a bunch of cool stuff that's like these really hyper specific chip that work in there.
Host 2 43:48
I'm been going over the datasheet for this. What are fat keys?
Host 1 43:55
Oh, I mean, that's some kind of
Host 2 43:58
Okay, yeah, that's what it says. I didn't know if that meant anything.
Host 1 44:01
No, it's some kind of yeah, whatever word they like to use for something. I don't know what it's for. Yeah, exactly.
Host 2 44:09
See if it mentioned what that meant. And it's just a marketing term. It looks like
Host 1 44:14
One of the things is okay, so this is this is fully analog, but it's really easy to slap a DAC on this and control your cutoff frequency digitally. This would make a an excellent tracking filter. You so you could you could reasons outside of sound synthesis for using filters of this sort. Like I know in my first job when I was doing vibration sensing that we had a lot of need for tracking filters and easily set this up for this or even configurable filters. I would almost consider using this for it. I think the only reason I might shy away from this is mainly it might not be rated for indoor to a range. It's a overall temperature ranges. But you know, if that's not a problem for you, there's a lot of power within this also in with this one chip because the all the all the transconductance blocks are broken apart, you could set this up as a two pole high filter, and then have variable band passing. So you could kind of like pocket your, your frequency data wherever you want to. So there's a lot of cool stuff that you can do with these. So
Host 2 45:29
Alright, so to wrap up the podcast this week, the topic will be interested in reverse engineering tactics. With customer supplied stuff, I guess it's a good way to put it. This comes up in the big customer project I've been working on work. When they supplied us a board to do the communication, it's FSK, high voltage for doing downhole testing for oil and gas. And so I started looking basically the the comm board, they gave us his little flaky, it tends to like test, we'll be able to test like 1000 units, and it decides to like explode. And so I'm trying to I was trying to figure out why is it going to why is it exploding? Why are it's a lot of analog circuitry, why are some of the op amps getting hot? And I was like, oh, Steven will, like look at it one night, figure out why these op amps get hot. But I started digging more into the good thing is the customer gave us the source code for the microcontroller on it. So I actually just started looking through it. I'm like, okay, comparing it with the PDF schematic that we have, unfortunately, that's like the only way we have it. And the code, and I'm like, Okay, what's this op amp doing? Right? With the control signal? Well, like half the control signals that come off the microcontroller are just set the like one or zero at startup, and they don't do anything else. And so I'm like, Okay, if that's the case, what does the circuit do? Absolutely nothing. Why is it making so much heat? I have no idea. So I just started the soldering them off the board. And basically going through and like, yeah, basically, like, hard wiring stuff high or low and just removing op amps off the board. I basically removed like, about 60% of the components on the board. And it still functions correctly.
Host 1 47:29
Were they? So I haven't seen this? Yeah. Where the Do you know anything about like, we're guys trying to future proof this or add function, so it'd be added
Host 2 47:39
It was? Well, I know that they have a tester that's like what's called the surface box or something like that. And that is what they use to test the switches in the field. But it can't do everything it can't it actually designed to not be able to do the firing part. Because you don't want to accidentally explode something on the surface. That'd be bad. And so they took that design, and then modified it and then started adding the firmware. So it can do all the functions we need to test the unit with. It just never got finished. Really? Yeah. Hardcore munching.
Host 1 48:24
Host 2 48:25
Basically just I just like started going through and just like, well, that doesn't actually do it like it has, where it can like this able, like the communication, and like, but it's set to one and there's no command to make it go to zero. So you can disable communication. So I'm like, well, I'll just rip those parts off the board, because it doesn't matter. And so basically, I was I actually was able to reduce the power draw the board by like half,
Host 1 48:50
Just getting rid of op amps, I bet. Most of those op amps were paying the higher low, like they were probably at their rails and just doing nothing
Host 2 49:01
Doing nothing and just heating up. And what will happen is one would pop and actually just like, like hit and, and destroy basically the signal or take a take this board out. Basically I take the board out. And you know, a couple months ago, I had no idea how to fix the board or how the board worked. Now I have a little bit more understanding of how it works. But yeah, it's it's kind of at least at least I think they're more stable now. We'll have to see.
Host 1 49:34
Well, that's goofy.
Host 2 49:35
Yeah. So if anyone in chat or on Twitter, or on Twitch has any experiences like that, I'd like to hear it were basically like, you know, reverse engineering customer products to make stuff better.
Host 1 49:56
Sounds like in your case, just to make it work.
Host 2 49:59
Just No, it did work. It just didn't work for long.
Host 1 50:04
I would be really like a what you find that is like the actual cause of them causing damage because like, perhaps they were over driven or or something of that sort.
Host 2 50:17
Yeah. So I'd like to still talk to you. Unfortunately, I won't be we won't be able to stream that because it's customer data. But go over that schematic and be like, Okay, why? Why that chip is like, heating up and lasts for like, a couple of days before it explodes, right?
Host 1 50:38
Yeah. I mean, it's that just sounds like something's overcurrent for Yeah, just because that's long enough to just cook it.
Host 2 50:44
Yeah. And it just cooks it for long enough. It finally explodes and then takes the board out with it. But basically dumping voltage on a rail that's that's supposed to have that high voltage. Think well sign off the podcast.
Host 1 51:00
Yeah, I think I think that's a good one. So that was the macro fab engineering podcast. We were your host, Stephen, Greg,
Host 2 51:09
And Parker Dolman. Take it easy. Later one
Transcribed by https://otter.ai