Unfunctional Manufacturing batches ???

jouni · ‎2005-11-02

Posted on November 02, 2005 at 13:03

jouni · ‎2011-05-17

Posted on May 17, 2011 at 12:08

!!PROBLEMS with UPSD3254-U40U6!!

1) Devices with MFG-code 5AXCX 0401 work OK

2) Devices with MFG-codes AANFT V5/TWN AA 414 or AANFB V5/TWN AA 413 has a lot of different PROBLEMS with same Hardware:

-Clock (40Mhz/3rd ot) does not allways run or runs with 1/3 frequency (=13,33MHz)

-Devices does not allways program (with Flink/JTAG)

-when programmed with success, and clock runs, does not allways 'do' the program

-when 'does' the program, some hardware still (like peripheral I/O-mode) does not function

-and so on ...

If you have any explanation or same kind of experiences, please inform. Like to get ST's opinion as well.

This feels strongly like some not qualified mfg-batch(=shit-batch) to me... (is this actually possible ?!?)

jdaniel · ‎2011-05-17

Posted on May 17, 2011 at 12:08

While it's entirely possible that ST made a batch of bad chips, there's something very suspicious in your scenario that suggests that's NOT the case. The oscillator in the uPSD is the same as what's in the 8231, namely a ''Pierce'' oscillator. This doesn't do ANYTHING to control the resonance point of the crystal. If the crystal ends up in 1st-order mode when you intended 3rd overtone, then either you didn't design the 3rd overtone filter correctly, or the actual implementation of your board has shifted the resonance point.

What values and configuration are you using for your 3rd OT filter? Have you actually measured the values on a production unit with an LCR meter?

IMHO, before worrying about a possible problem with a batch of chips, you need to set your sights on the DEFINITE problem with your harware.

joseph2399 · ‎2011-05-17

Posted on May 17, 2011 at 12:08

If after taking another look at your hardware, and you feel that there is still a ''possible'' issue with the chips you have, please contact your local ST Salesperson, or Distributor, and have them send some parts in for Faiilure Analysis. this way you'll know if the issue is with the uPSD or with your hardware.

jouni · ‎2011-05-17

Posted on May 17, 2011 at 12:08

OK, I just got a new pack of UPSD3254A-40U6's (119pcs) in original sealed bag (dated 16 September) with trace codes AA532ZHR V5 7B & AA529XWT V5 7B from Arrow (USA).

Changed new processors to two cards (without any HW changes)==>>

EVERYTHING WORKS SMOOTH AND NICE !!!

The other 20 pcs of cards are now under work, I'll be back when they all are tested ...

jdaniel · ‎2011-05-17

Posted on May 17, 2011 at 12:08

jones,

I can believe that, but that's still not going to tell you if there's something wrong with your oscillator circuit. Certainly there's a range of input capacitances into which a batch of chips could fall and still be within spec. If the total capacitance in your filter/clock circuit gets shifted by this amount and ceases to function, then you might not have bad chips, but still a hardware issue.

Take a look is all I'm saying. I've seen problems like this before and I HAVE seen hardware issues be the problem despite the trial-and-error testing pointing to the contrary. And while you might change over this whole run of boards to the new batch and have them all work, this problem could certainly still rear its head again with the next batch if the chip isn't really the problem. Here's another possibility to test: Get a 1st-order crystal for your clock frequency if one's available (citizen seem to be making 1st-orders in higher frequencies than some other companies right now). If that works after removing your filter and hooking it up, it suggests a filter problem. Failing that, try the chips with an external clock oscillator and see what kind of results you get.