Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
14 years ago

TMS stuck low on CYCLONE III

Yesterday we had been happily programming our cyclone III based prototype via JTAG, Quartus and a USB blaster fine until all of a sudden we could no longer detect the JTAG chain!

I traced the problem to being TMS stuck low.

Removing power from the board I could measure a short circuit (1.5 ohm)between TMS and gnd.

Removing all connections to TMS except the FPGA the problem persisted. I applied 0.8v to TMS briefly and the resistance to gnd increased to 14 ohms! I repowered the board and applied 4 volts to TMS, powered down and rechecked the resistance to gnd. It had increased to 200 ohm!

I can now re-program the FPGA fine using the USB blaster but TMS only rises to a couple of volts as it is still being loaded bythe 200 ohm to gnd!

I would put this down to a single damaged FPGA BUT! I have another board returned from the customer with a reported JTAG fault. I have checked the TMS pin and lo and behold its resistance to gnd is 70 ohms!

VCCIO on all banks of the FPGA is 3V3

Can anyone please explain the possible causes for this strange behaviour?

Cheers

Dave

3 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Are you sure you don't have an external micro short circuit between TMS and ground, such as a small metal piece or some solder?

    It could also be a clamping diode to GND that is failing. This could be caused by an ESD pulse or a negative voltage applied by accident to the pin. Do you have any kind of protection between the FPGA JTAG pins and the connector?
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Thanks Daixiwen

    Pretty sure it isn't a micro short as I have the same problem on two boards. Also I have checked very carefully, cleaned and re-cleaned .

    Your second suggestion is possible though. There is no protection between the JTAG header and the FPGA. A negative voltage could have been applied when the USB Blaster was plugged in to the JTAG header on the board. This could have damaged the clamping diode. The engineer tells me he may have had the JTAG disconnected while he ran some tests then plugged it in to re-program the FPGA! Which then failed to detect the chain!

    Thanks

    Dave
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Hi Dave,

    I just had the same problem with the TDO pin on the JTAG. I have a cyclone IV chip that has been working just fine. All of a sudden the JTAG stopped working. The TDO was shorted to ground with 40 ohms. It was still working, but just a lower voltage(.5V swing).

    Following your suggestion I applied 3.8V to the pin and the resistance went up to 200 ohm. Now it works.

    Like you I would like to know what happened so I can keep it from happening again.

    The JTAG is configured to work on 2.5V correctly and everything worked just fine up until it didn't. I only have protection diodes on the ASDO etc pins to the EPCS device as Altera suggested. They don't recommend anything if your just going to their programmer.

    Dan