Hi Welcome to the Eev blog an Electronics Engineering Video blog of interest to anyone involved in electronics design. I'm your host Dave Jones Hi. It's industry story time now. let me set the scene for you.

Okay, we've got a couple of thousand boards going into production. They've been produced for you know, quite a long time now and this is just yet another batch of boards going through and it's overseas. it's in China and we get the call back saying hey, there's you know something wrong. These boards don't program properly.

What's going on? Can you investigate? Okay, let's check it out. This is what happened. Okay, basically it's got a um each board okay, we've got has what has one of these uh Dallas one wire, uh, programmable devices on it. It's a DS uh 2502.

We'll take a look at the data sheet in a sec, but it's basically a 1 kilobyte uh e Pro from identification chip. and we put this little chip on pretty much every board that we manufacture as an ID chip because out of the factory these are really cool out of the factory. it comes with a uh, a laser um exed ID in it which is unique. No two chips have the same ID so they're great for individually identifying uh boards you know when they were manufactured and etc etc.

and um. also you can store store up to 1 kilobyte um of information in there as well which is Handy for like a product ID So we'll have our little product ID tag in there. You know it'll have the product number and it might have some other info as well and that's written in at the time of manufacturer and we've getting the report back that Well, this information wasn't being written properly. it would just be it had write something but then it had read back.

Just all garbage. Can we take a look at it? Okay, let's see what happened now because this was in in China and we didn't have anyone on the ground over there to investigate directly. You start with the usual things. oh, you know, tell us what part um, you've loaded on there.

Give us, show us a photo. Is it mounted properly? That's what you might think. Okay, they' You know they've programmed the machine properly, they put it in backwards or something dumb like that, or they've accidentally used another part. It happens something like that.

But because we knew that the um uh from information that we got that the Uh ID was being out of the chip correctly by our processor, then we knew. Okay, at least there's communication going on there and there's something wrong. The programming process. The info that we programmed into this chip was just.

it was just readback as garbage. It was incorrect. So we got a photo of the part. it's got DS 2502 on it.

No problems at all. It's mounted the right way. Everything looks good. No problem at all.

And because we could read the chip device ID we knew that it was communicating with our processor. Not a problem, it was just reading out garbage And that was a problem. So we thought aha must be something wrong with the programming. Now this, uh Dallas semiconductor Maxum I still like to call them Dallas Semiconductor cuz I like Dallas Parts I don't like Maxim Parts but the same company, eh, whatever.

Anyway, the Dallas DS 2502 This is actually a 12vt programmable device, so normally it operates at 3.3 or 5 volts. But to program the information in there to write the information in once you got to provide a 12volt programming pulse you know, bang for a couple of micros seconds before you write each uh datab bit and we thought, okay, there's some 12vt circuitry on there. Is that the problem? Ah, so you check out all these things and know you get down to the point where you start ruling out these little uh problems happening everywhere and it looks like well, there's something wrong with the chip, what's going on So when you start eliminating some of the real basic dumb stuff, you start to get down to the circuitry involved and this is what. Basically what we had uh here because we thought, okay, there's something weird going on here and we've seen issues before, similar sorts of uh problems and stuff like that.

so we thought there might be something going on here. Now basically this is our DS 2502 ID uh chip that we're talking about. It's got a series protection resistor in here. It's got a pull-up resistor to VCC which it needs cuz it's a one wire.

uh bus. A one wire needs to pull up because this all this chip can do is pull down and it can be parasitic powered and we'll look at that later and then it's got a Zena diode for protection because um, well and we've got it going off to an external programming um programming uh header on the board that you can program the device through. and also there's a 12vt generator as well. but I won't draw that on here because we've already checked that and it does generate.

um, its 12vt programming pulse. Now the reason this uh circuitry here is because but is because it goes off to an arm processor over here or an Fpga depends what the product is, it goes off to it and um, it's got to talk to the device as well as uh, being able to externally program it. Now when you program this thing, you put your board in your test jig or you use some other circuitry to program it some onboard circuitry in this case. then um, it generates a 12vt programming pulse.

So normally it talks 0 to 3.3 Vol digital down here and then you'll get a big 12v programming pulse. And if you don't have any protection circuitry that can blow up your processor or your Fpga or your other device which is trying to talk to it now, um, these values actually here can actually load down the line from the Uh external programming circuitry. So we thought, okay, maybe they've loaded a wrong part on here, but we checked all that and it wasn't. There was nothing wrong with it.

It all looked quite good. No problems at all, so we had to investigate further. Now further testing had to wait until we got uh the assembler to send back a sample from The Reel cuz they had some leftover. Parts it was Man on one of these reels and here's a it's on like tape like this if you can see that the uh, there it is there.

The Uh chips are actually uh, embedded in the little tape on the reels like that. so they come straight from the manufacturer uh on a reel. So we had to wait until we got some samples back and then uh, what? I did is I actually uh mounted one cuz I didn't want to mount it on the board under question I just wanted to directly talk to it with the original uh Maxum um USB programmer the one wire programmer device that we've got so solded it onto a blank board like that with no circuitry whatsoever. It was just the chip itself, so there could be absolutely no question that any other circuitry on the board was interfering with it.

So I sold it on the board Use the programmer and here's what I've got. Check it out. I Actually loaded up right? A good chip first to get a reference and this is what a good chip looks like on the Reel from the factory, it's supposed to be blank. It's supposed to have Fffff in all the data areas and as you can see it, there's a device.

um ID up there in the Uh top left hand corner. It tells you it's a Ds25. A2 It identifies the chip. No problems at all.

So our system's working. We got a good bench line to work with. Now here is the Uh. Here is one of the chips that they sent back.

Brand new chip from: The Reel hadn't been sold it on the board yet. And check out the data dump. Look at it. it's not blank at all.

It's got data in there. It looks like it's corrupted. Bingo We've found our culprit. Okay, this chip is supposed to be completely blank from the factory unless you specifically ask for the manufacturer to actually uh program it for you.

That's a service that a lot of Manufacturers will do. You can actually have it programmed. It comes on your real already program with the information that you need for your particular product. So Bingo we found our culprit.

The chips weren't blank for some unknown reason. Were they just corrupted with ram? random data? Well, we didn't know yet, but we knew that they were supposed to be blank. Now, the problem with these devices is that they're what's called a right once uh E prom memory. so you only supposed to write to them once.

That's why they all come programmed with all Fffff. And once you write a bit of data to a zero, an individual bit to a zero, you can't change it back to a one again. So they're effectively a right. once.

E S problem, you only get one shot at it. If you accidentally write the wrong data on there, you got to suck the chip off and put a new one on. So those chips were absolutely useless to them. It's not useless to us.

It's not like we could raised them and then reprogram them. That would have solved all our problems Now of course, when you look at this data, just the hex stump like that, it looks like it's just random data, right? Something's gone wrong with the programming process. Ah, you don't know? Um, so I thought. Rather than delve into the software and trying to debug it and it's in a different country and ah, it's all too hard I Thought I Just translate the uh, the hex data dump into Asy just to see what was going on.

So here it is. Take a look Bingo It is not random data. There it is in plain asy. Uh, Copyright 1995 Motorola Corp You got to be kidding me.

Where did this come from? So Bingo There you have it. We found the culprit. This isn't just random data Rand you know, produced some fault in Out programming system. No, these chips came pre-programmed on the damn reel.

okay with some sort of Motorola firmware. I've got no idea what Motorola product it's for or you know where it came from or how. We still need to do further investigation on that, but clearly we've been given dodgy chips. so how did it happen? Well, as I said I got no idea.

we have to do further investigation. So I'm not going to blame Maxim just yet. But I've talked to the purchasing guys he purchased uh, these things directly on the reels and they're saying that they always buy from either directly from Maxim or from uh, a an authorized component distributor like Digi key Mouser or somebody like that because we're very careful about this. We've been, you know, bitten by this before in the past you buy parts from the gray market and they're fake or something like that.

So there's only a couple of things that this can be one. is that, uh, they're actually fake gray Market chips and they've been. uh, they've already been pre-program for some other customer or they ripped them off boards and then they've Reed them because it's not like the these things, they just shove them back in tubes or something like that. you've got to actually Ral them.

Um, onto original things, put the uh, put the tape on them and all sorts of stuff and well, you know it's not hard to do that, but you've got to have the equipment to actually do it. And a lot of the gray Market companies they specialize in doing that. they Rebrand the chips and they actually, um and they repackage them and make them look like brand new and you get sucked in. They either don't work at all or they're close or they're pre-programmed with someone else's firm.

So that's a possibility. or the more likely scenario because we bought them from authorized s sources I don't reckon that they're gray Market uh chips I reckon they've been supplied. Here's my guess from Maxim themselves directly from the maximum Maxim inventory pre-programed with somebody else's firmware I reckon they've programmed them, screwed it up, put them back on the shelf and sold them to to us in this case. So I reckon that's the a much more likely scenario than getting uh, gray Market chips cuz we're very careful where we buy them from so we still have.

We haven't talked to Maxim about it yet so you know it's it's all still out there. we haven't thoroughly investigated, we haven't even got the original reel back yet. but I just thought I'd share this with you as an experience of it's not I'm not going to say it's common, but it has happened before. I've had issues like this before so it's definitely it.

Definitely happens out there and it's something you got to consider. Now one of the uh things that you also do is take a look at the chip itself to actually um, see now I've got my trusty little Xtec MC 108 um portable microscope here I Don't reckon it's that great for uh, soldering to soldering inspection, but it's brilliant for looking that uh, the numbers on top of chips. so let's take a look at it. Now let's actually take a look at a good chip.

Here is a good chip and as you can see, DS 2502 and it looks like it's been laser marked I'd say you can tell by the really sharp um edges on there and if you actually look at Angles under certain lights you can, you can tell it's laser etched as opposed to it's some sort of silk screen uh process. Now the plus mark up there indicates that it's Ros compliant. R HS and um, now the numbers down the bottom I Couldn't find anything on the data sheet uh to indicate what those numbers are, but my guess is that 9 is the year, 09 27 is the week of manufacturer and D3 I Don't know it might be some sort of batch number or um, some silicon revision or something like that, so we we'd have to ask Maxim direct to actually do that. Now let's let's turn it up a bit and as you can see, the uh, the figures are nice and sharp.

No problems at all. And here you go. here is the faulty chip. Now as you can see, the um, the silk screen isn't nearly as big or the identifying marks aren't nearly as big as they were on the other device and I think they're still laser, uh, etched I think but they look substantially different.

Um, now as you can see, it's um 08 as well. So considering these are supposed to be fairly new parts, if that is the year code uh down here, then we've certainly got old chips. If it is 26, we, um 08 and B2 is different to the other device. Now you know that's not telling evidence that it's a fake chip and I'm not going to claim that it is fake, but um I guess you know Maxim could tell us uh, something.

you know if there's actually an issue there and if the chips are real. but they certainly have been pre-programmed with someone else's firmware. So there you have it. There's the problem with the Maxim chip that really, it came down to the last possible thing I thought it would have been which was a pre-programmed chip because you just assume that the chips you're going to get from the manufacturer are good and that happens.

You know all the time. you basically got to rely on that. and there's all sorts of other things that would cross your mind as being um, the failure mechanism for something like this before a pre-program chip. I Couldn't believe it when I finally saw that copyright motor roller in there.

Unbelievable. But hey, that's a real story. There you go. watch out next time.