cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
jensrenner
Contributor
Contributor
1,266 Views
Registered: ‎02-20-2014

Zynq counterfeit question

Hello forums,

I just received a new batch of SDR boards from our PCB assembler who also purchases the components. The design features a Zynq XC7Z010-CLG400C and has already been manufactured, tested, and sold in the past.

In the current batch though, for about 25% of the boards I am seeing an overcurrent together with the Zynq IC getting very hot within seconds. As of now, I have no reason to believe that this is a general design flaw or an assembly issue. (Otoh, it cannot be ruled out for the moment.)

There is one observation on all of the Zynq ICs: The QR code that gets lasered onto the chip case seems to be altered. Some of the pixels have been blacked-out in an abrasive way. I have never seen such a modification in the past. I have asked the assembly company to tell us the distributor where they have bought the components.

Has anyone come across Zynq SoCs or other Xilinx components with a changed QR code like this? Is that a regular thing which is done by manufacturers or should I worry whether or not our Zynq chips are legit? Could they be counterfeit or scrapped parts? Any hint would be very much appreciated. 

IMG_20200424_144055.jpg
IMG_20200424_144214.jpg
0 Kudos
7 Replies
tenzinc
Moderator
Moderator
1,175 Views
Registered: ‎09-18-2014

You really need to verify how the part was sourced. It's definitely not normal to receive parts with modified 2D bar-code. I just can't think of a reason to do that intentionally? NONE. Unintentionally, there could be a few reasons from bad manufacturing/assembly process/techniques to other bad logistical issues. You may want to do your due diligence with your PCB maker and verify with them that they did or didn't receive the parts like that before they assembled it on the board knowingly. Identify the distributor and make sure it's a XIlinx authorized and not just some flipper/wholesaler from Ebay... Make sure that they can state that the parts were delivered without such modifications. I am not sure if this is an hobby or a student project but if it were for a legitimate business it's an requirement to know where you are sourcing your components from. 

 

Reards,

T



Don’t forget to reply, kudo, and accept as solution.

Get started FAST with our UltraFAST design methodoly guides and don't forget to visit our Xilinx Design Hubs for additional resources and reference.

If starting with Versal take a look at our Versal Design Process Hub and our Versal Blogs

------------------------------------------------------------------------------------------------

jensrenner
Contributor
Contributor
1,162 Views
Registered: ‎02-20-2014

@tenzinc Thanks for your reply. This is commercial and not a hobby project. So far we've had no reason to believe that the parts purchased by our assemblers are coming from a questionable source or aren't genuine. Of course I have inquired about the distributor of the Zynq chips or if they can explain what has happend to the QR code. I am still waiting for their answer. I'll keep you posted ...

Jens.

0 Kudos
barriet
Xilinx Employee
Xilinx Employee
1,154 Views
Registered: ‎08-13-2007

I'm not associated with the quality team, but agree this looks "sketchy". I would ask to see the purchase order they used to place the order.

You can see the list of authorized distributors here:

https://www.xilinx.com/about/contact/locations.html#authorizedDistributors

Cheers,

bt

u4223374
Advisor
Advisor
1,121 Views
Registered: ‎04-26-2015

@jensrenner If you have the necessary equipment, the obvious place to start is getting a Zynq from a reputable supplier (eg. Digikey, Mouser, Avnet) and swapping out the existing part.

 

If it works perfectly then you have either a faulty chip or bad soldering. Since you've got a number of failed boards and it's likely that they all have similar problems, you could have another few sent away for X-ray inspection of the solder balls.

 

If it fails, then you have a different problem. Are your power supplies all good? Low noise, voltages exactly right, etc?

jensrenner
Contributor
Contributor
1,091 Views
Registered: ‎02-20-2014

@u4223374 We have business relations with a re-work company that is also doing x-ray inspections. At this point in time we would like to avoid the effort and cost for this. But ultimately this might be the way to go.

Power supply itself seems to be good. But it either detects an over-current on one or all of the Zynq voltage rails and shuts down, or continues to deliver power at very high currents (about 2 to 3 times compared to normal). I can even measure close to zero Ohms on some or all of the Zynq power rails. But if it was a shortcut on the PCB, I wouldn't expect the Zynq to get hot at all.

@barriet So far the assembler is not willing to tell us the exact source of the parts and has asked for the faulty boards to be returned. They just said they've bought the chips "on the free market" (whatever that may mean). But I insisted that they have to show us the order details. Should be no big deal if they have nothing to hide.

The more I think about it the more likely it seems that these parts are "salvaged" scrap parts that have failed factory tests. I just requested access to the Xilinx 2D Marking Application so I can check the 2D code on our parts.

0 Kudos
u4223374
Advisor
Advisor
1,077 Views
Registered: ‎04-26-2015

@jensrenner With regards to the chip getting hot - I've seen a BGA get hot because there was a short between solder balls. After all, there are extremely good metal-to-metal connections between the power/ground planes and the chip (via all the power/ground pins), so a short on the PCB is going to translate into a lot of heat transmitted into the chip.

 

That does provide another meaningful test - get the chip removed, clean up the bottom of it (so there are no visible shorts), and then re-do your resistance check. If it's a sensible value now, then chances are that the soldering was at fault. If it's still a short-circuit then you've got a faulty chip.

 

Edit: it's also possible that they're perfectly legitimate chips that have failed due to handling. In particular, almost all chips are moisture sensitive and will suffer if they're left in the open air for a period and then reflowed (as the moisture that has entered the chip rapidly turns to steam and does internal damage). This may be more likely now, if (for example) the company assembling the boards prepared all the chips for a production run (pulling them out of their packaging etc), shut down for a few weeks for COVID-19, and then went back to work and immediately put them through a reflow oven.

jensrenner
Contributor
Contributor
1,050 Views
Registered: ‎02-20-2014

@u4223374 Thanks for pointing that out. Of course I have thought about this possibility and it seemed somewhat unlikely, because I expect the assembler to make sure their solding process is ok (we've had issues with them in the past so I really hoped they have double checked their process parameters). As we have sent the boards in question back to them, they are probably going to x-ray the boards. Either way, they better find out what went wrong ...

0 Kudos