Thu Oct 25, 2012 10:07 am by wrf
Got back to diagnosing the TGP V last night. No break thrus, but there is a little new information.
Symptoms are now stable. When I plug the DSP board into the mother board, power consumption spikes up to 67 watts and U57 on the mother board quickly gets REAL hot. U57 is an LM317T regulator mounted on a heat sink. Silk screened voltages indicate that it provides 3.3v. Chasing connections with my meter shows that its 3.3v go to the DSP board connector on pins 24 and 25. Unplug the DSP board and power drops to 37 watts and U57 remains at a stable temp.
With the DSP board unplugged, voltages on U57 are 2/3.3/11.4 on pins 1,2 and 3. With the DSP board plugged in they are .5/.87/10. There is a regulator on the DSP board - U2. Its voltages are .5/1.8/10. It too is an LM317T. I looked up the data sheet on the DSP processor and it uses both 3.3 and 1.8v power connections. With the DSP board removed, I measured resistance between pins 25 and 1 on the connector. I believe these two pins to be +3.3v and ground. Resistance is .5 ohms.
My conclusions so far:
- U57 on the mother board provides +3.3v to the DSP board via the connector on pins 25 and 1.
- U2 on the DSP board provides +1.8v to the DSP processor (and maybe the memory).
- Something is shorting the 3.3 power on the DSP board.
What have I done so far?
- Blew the DSP board out multiple times. No changes.
- Took the board into work and handed it to the QA guys. They went after it with their microscopes and couldn't find anything.
- There are six electrolytic caps on the DSP board. I took them all off and I still get .5 ohms on the 3.3v power (pins 25 and 1). All the other components on the DSP board are surface mount.
Open Questions:
- What is shorting the 3.3v power on the DSP board?
- Why the earlier measurements where I got 3v from regulator U2 on the DSP board and power consumption of about 43 watts?
I'm still hoping that its something stupid - some debris in the wrong place or something. My fear is that a device on the DSP board has failed in a mode where it's shorting the 3.3v power.
I'm pretty much stumped at this point. Any suggestions short of turning the TGP V into a planter would be welcome.