7
\$\begingroup\$

I'm debugging a PCB which has a AT Mega 128L and two CPLDs mounted on it. The microcontroller actually works OK but I'm unable to program it via JTAG (and I don't have ISP or any other headers on the board). I was obviously able to program it earlier.

When the problem began, I checked for continutity between all JTAG pins. The Reset pin was problematic and I soldered a wire onto the Reset pin and connected the other end onto the JTAG header. I assumed this would have fixed the problem, but nope. I still get the same error: Failed To Identify Target.

But while doing the continuity checks, I came across something more alarming. The meter told me I had continuity between my 3.3V rail and Ground! Not only that, it also beeped when I checked for continuity between 3.3V and 1.8V. I am quite sure that a large current is not flowing because the power supply isn't hot - I know I should have measured it but I had to leave work for other purposes. But none of the chips on the board are hot either.

My colleague whose working with me on this board told me this is actually normal and its actually because the some pins are internally shorted in some of the chips. He doesn't seem to think this could be the cause for our JTAG problems. I'm quite skeptical of this claim because when is it good to have continuity between Vcc and Ground?! But I am quite inexperienced with such things and hence I'm asking this site. Is it normal to have such continuity?

Note: I know that capacitors will intermittently give a beep but once charged, its essentially an open circuit. In this case, I've held the DMM leads onto the pads for for quite some time and it gives a continuos beep.

\$\endgroup\$
2
  • \$\begingroup\$ How much current is drawn from the supply when switched on? What voltages are on the supply rails when switched on? \$\endgroup\$
    – markrages
    Oct 22, 2011 at 18:48
  • \$\begingroup\$ Is there a continuous beep or just a short beep? If it is a short beep it may be just the decoupling capacitors, which is normal. \$\endgroup\$
    – starblue
    Oct 22, 2011 at 21:14

3 Answers 3

3
\$\begingroup\$

Swap your positive and negative meter leads. Do you still read a "short"?

Check the power consumption of your board when powered, or better yet, put it on an adjustable current supply with the current limit all the way down and see how far you have to turn it up to get out of current regulation mode and into voltage regulation mode.

Most likely your jtag problem is unrelated to the meter reading. If you can't pull some of the parts out of sockets, try slowing the jtag clock.

\$\endgroup\$
1
\$\begingroup\$

Shorts between power and ground are always the most frustrating. Because so many things connect between power and ground, it means that the problem could be almost anywhere.

Also, if enough things use power, the resistance measured between the two can be low enough to trigger the beep. What is the actual resistance between power and ground? Often 50 ohms or lower gives a beep.

When I get this problem, I usually start removing components (or, if possible, lifting the power legs of chips) until it starts working.

\$\endgroup\$
2
  • \$\begingroup\$ Thanks for the response. The resistance was 3 Ohms. I really don't think it should be that low - I only have two CPLDs and a MCU connected. How can they draw 1A of current?! \$\endgroup\$
    – Saad
    Oct 22, 2011 at 14:58
  • 1
    \$\begingroup\$ Are you actually measuring one amp of current, or are you inferring that from the "ohm meter" reading? It is possible for CMOS chips to get into a state called "latch up" where they draw extreme amounts of power and consequently get painfully hot - sometimes whey will recover and sort of work (though you can't trust that board again) other times they are gone for good. But if you haven't actually measured an amp of current or burned your finger, but are merely inferring it from the ohm meter, think about what a diode normally does... \$\endgroup\$ Oct 23, 2011 at 0:25
0
\$\begingroup\$

If the board is powered off, then your 3.3V, 1.8V, and ground rails will all be connected by the protection diodes in your ICs. That doesn't mean there's a short! On many multimeters, the continuity checker doubles as a diode tester. Look at what voltage you're getting from this "short". If it's in the 0.3-0.7V range, then you're seeing a protection diode, which is nothing to worry about. Measuring the voltage on the 3.3V and 1.8V rails while the board is powered up should tell you whether there's a problem. Measuring the current from the power supply is of course also a good idea.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.