FPGA Hardware/Cell Diagnostics

FPGA Hardware/Cell Diagnostics

Post by Nevi » Sat, 19 Feb 2005 04:17:30

We are experiencing problems with our VirtexII FPGA. Preliminary
debugging indicates that it may be bad hardware. We want to verify that
the cells in the FPGA are good. Is there any kind of diagnostic tool
available to scan FPGA and verify hardware integrity? Thanks in

FPGA Hardware/Cell Diagnostics

Post by Ray Andrak » Sat, 19 Feb 2005 11:05:19

Talk to xilinx. They test the cells before shipping, although it is
possible a fault might have made it out undetected, in which case they
will be VERY interested in knowing what the fault is as well as how to
detect it. It is extremely rare for a cell to go bad later, but again
there is a possibility. Generally speaking, if the FPGA does have a
hard failure, it is usually going to occur on an I/O that got
overstressed (often by ESD). What are the symptoms that lead you to
believe it is a hardware fault?

--Ray Andraka, P.E.
President, the Andraka Consulting Group, Inc.
401/884-7930 Fax 401/884-7950

"They that give up essential liberty to obtain a little
temporary safety deserve neither liberty nor safety."
-Benjamin Franklin, 1759


FPGA Hardware/Cell Diagnostics

Post by Bob Perlma » Sat, 19 Feb 2005 12:36:51

Hi -

On 17 Feb 2005 11:17:30 -0800, "Nevin" < XXXX@XXXXX.COM >

If you have strong evidence of a hardware error, you may want to talk
to Josh Rosen at Polybus Systems, who's done quite a bit of work on
FPGA test patterns. For more, go to http://www.yqcomputer.com/

But I agree with Ray. While I've seen I/O cells get fried (and maybe
someone whose e-mail moniker is electrostaticman has seen this, too),
I've rarely found internal hardware faults in FPGAs, and always make a
point to look elsewhere first.

Good luck,
Bob Perlman
Cambrian Design Works

FPGA Hardware/Cell Diagnostics

Post by larth » Sat, 19 Feb 2005 17:31:14

If you are VERY certain that your Xilinx FPGA is defective, you can
ask your distributer (if you bought the FPGA from an authorized
source) to start a process called RMA, where you are requested to
answer a number of questions and are eventually (if you qualify) given
an account number to send in the defective device to. Xilinx will then
inspect and E-test the device to find out what went wrong. As said,
normally you would expect defective I/O from ESD damage and I don't
know if that would qualify, but if you encounter an internal error
that can be attributed to a logic cell, it might.

Good luck!