PC won't boot (start/stops) when internal HDD is connected to PSU

[b_profane]

I have recently replaced several of the main components of my system due to a power failure after returning form holiday. Long story, which can be found here.

After sorting out this PSU situation, I ended up with a quasi-new system, which is detailed as:

- PSU (EVGA 850w supernova, NEW)
- motherboard (MSA 970 Gaming, NEW)
- CPU (AMD 8-series, NEW)
- GPU (Geforce Titan X, about 1 year old)
- HDD 1 (250 gb, used for apps, about 4 years old)
- HDD 2 (Seagate 1.5 TB, used for data, about 4 years old)

Attaching all components together to the new PSU causes the computer to start / stop (i.e. all fans & lights come on, but only for about 1 second, and then it all stops). Attaching them one at a time, and I have narrowed it down to HDD 2 that is causing the problem. If I disconnect HDD 2, the computer boots fine. I have tried different variations of connection including attaching to different positions along the SATA power cable, using different SATA power cables, and connecting only HDD 2 and not HDD 1. I have also cleaned off all dust thoroughly. None of these have worked.

So I have now detached the PCB on the drive to have a look at the diodes. I am following advice in this thread. And particularly I am following the well laid out instruction by fzabkar which is here.

There is no physical damage noted on the circuit board. Seagate has made this very difficult to read, but I believe I have identified the two diodes next to the power connector as the +12V diode (labeled BUX) and the +5V did (right next to it, labeled 5J). I have tested the resistance of both of these with the multimeter.

The +12V diode tests as "1" if the setting is on 200 ohms. If I move to 2000 ohms for one side, it tests around ~650. If I move to 2000k ohms for the other direction it tests around ~600. Does this sound about right?

On the other hand, the +5V diode tests 0.6 in both directions (using 200 ohm setting). Does this mean it is dead? Should I be snipping this? As this is my primary data drive, I don't want to cause any permanent damage. I do feel I can rely on the new PSU however, so not worried about the temporary risk to over-voltage. More interested in getting this drive up once to transfer all the data to my external one.

Also I have done some voltage tests on the PSU with and without this drive connected. With the drive connected it seems the red wire (+5V) is never activated (stays at 0.00 all throughout boot). The yellow wire (+12V) jumps to ~10-11, and then quickly drops back to 0.00. Thus the start-stop described in the title of the post.

I will try to upload a couple of pictures of the PCB if you bear with me. Ok, have uploaded a couple of images of the PCB.

LINK TO PHOTOS

Cheers

 

[b_profane]

A shameless self-bump. Photos of HDD 2 uploaded.

 

[b_profane]

So an update on my on-going detective work, and also so more accurate data on the TVS diodes / resistors as I have been reading more.

First of all the old PSU was diagnosed to have a defective (read blown) red-orange regulator. Which would explain why HDD 2 is start/stopping the boot, as it suffered over-voltage on the +5V current (see below). If I disconnect HDD 2 from the system it boots fine. But if I try to install Windows (this is a fresh mobo), it won't detect HDD 1. So readings on the PCBs of these two HDDs are as follows (all diode tests done using diode test setting on multimeter; all zero-ohm resistor tests done using 200 ohms setting):

HDD 1 (powers up but not detected)
- +5V TSV diode: appears to be closed (reads 0 ohms in both directions)
- +12V TSV diode: appears to be open (reads ~650 in one direction and OL in other)
- zero-ohm resistor next to +5V diode: reads 5.9
- zero-ohm resistor next to +12V diode: reads 0.5 (which is same reading as touching two multimeter nodes together)
- SATA connector: Testing +5V-GND resistance is 5.9

HDD 2 (starts/stops)
- +5V TSV diode: appears to be closed (reads 0 ohms in both directions)
- +12V TSV diode: appears to be open (reads ~650 in one direction and OL in other)
- zero-ohm resistor 1*: reads 2.5
- zero-ohm resistor 2*: reads 2.5
- SATA connector: Testing +5V-GND resistance is 0.5

* the resistors are not situated on the PCB in a way that identifies them (see photo)

Snip, solder, both or neither?? Any help would be much appreciated. Cheers

 

[b_profane]

So I never heard from anyone here, but after thorough research on similar problems around the web, I've managed to solve my issues. So I'm posting here the conclusions in case it might help someone in the future without having to do so much research. All multimeter tests below were done using a 2000 ohms setting.

HDD 1
- 5V diode was closed; 12V diode was fine.
- The zero-ohm resistor nearest to the 5V diode was reading 3 ohms. The other one was reading 0 ohms (fine).
- I snipped out the 5V diode with flush cutters by nibbling away on one end of its connection until it came free. Then I bent it up and snipped off the other side. I left the 12V diode alone.
- I then took a soldering iron and loosened each side of the defective zero-ohm resistor and then snipped it out with the flush cutters. I put a fresh blob of solder over the two pads that were left (this was a bit tricky as it was effectively my first soldering ever).
- RESULT: Drive works! It boots up just fine and I grabbed me data.

HDD 2
- 5V diode was closed; 12V diode was fine.
- The ohm resistors on this board are two-ohm resistors (designated by the 2R0 on the top of them). Both were reading 2 ohms, so were fine.
- I snipped out the 5V diode with flush cutters, same as above. I left the 12V diode alone.
- Nothing to do on the ohm resistors.
- RESULT: Drive works! It boots up just fine and I grabbed me data.

Anyway, happy tinkering, people. I guess I can't click solved on my own post...

 

[westom]

Anyway, happy tinkering, people. I guess I can't click solved on my own post...

Some background for future reference. Better meters use less than 0.7 volts to test conductivity. Then other components (that are not shorted) will not conduct - will not act shorted.

Use a second meter to measure voltage from probes on the first meter in 'conductivity test mode' (or in ohms). A higher voltage makes it difficult to find shorted component such as that diode.

Also available are conductivity testers that use less than 0.7 volts to detect a short circuit (ie that failed diode). These also do not conduct power 'backwards' through other semiconductors.

Kudos for truly sticking with and solving the problem.