Dell Precision T3500 - 09KPNV motherboard repair

[wasserprojekt]

Hello!
I've been given a batch of 3 Dell Precision T3500 in pretty bad shape. After cleaning all the hardware, most of components were working but one graphics card and a motherboard. I used one of the working motherboards to test all the memory modules, all PSUs, the CPUs and expansion cards.

I am trying to ressurrect the broken motherboard. No beeps but fans work. Computer turns on for three seconds then shuts down. Board's amber LED behaves as expected. Board's green LED near motherboard PSU connection turns on when board's powered, as expected, then turns off when it shuts down. The CMOS battery is working, I tried clearing CMOS by the regular means, tried some jumper configs seen on the internet. Replaced processor, one memory stick only. No success.

On further inspection, noticed the CPU does not warm up. I tested some random components on the motherboard with multimeter. Found 3.3V, 5.1V, 12V in places they should be (on components near PCI-E, USB, fans, LEDs, etc...). However, not surprisingly, no power on VRMs, though there is power on capacitors near them. Then I tested the capacitors (I think they are capacitors because of the component name, C1907) near the 10-pin processor connector and I noticed one burnt up small cap. The other ones of this group all have the expected voltages (12V).

I suspect the problem arises from this cap, although I don't know if there are other broken components. However, I cannot identify this cap, as there is no identifier.

I would be grateful if someone could give me some advice on this and helping me identify this cap. I also searched the whole google for the schematics and board view, but I have been unsuccessful, making the whole thing harder to solve.

I expect people tell me to just replace the whole motherboard. However, I would really like to repair it, as this is my first attempt at doing so. It's the perfect situation to learn something more without catastrophic consequences.

I have attached some photos of the culprit.

Thanks a lot for your patience!

[Schorsch87]

Re: Dell Precision T3500 - 09KPNV motherboard repair

I would say, that on this places inductors have to be, because of the naming with L114 and L101. As for testing I would solder in some fuses to check whether the mainboards starts up.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Thank you very much for your reply. Going to try that. Any recommendation on which specifications should the testing components have? Sorry, I am in the noob side of the moon eheh

[dicky96]

Re: Dell Precision T3500 - 09KPNV motherboard repair

It's hard to see from the Pic but I am not convinced this small capacitor is causing your problem. If it is a capacitor that is.

Do you get any Vcore for the few seconds the board powers on?

Do you have a PCI POST test card - they are very useful to help to see what is happening and also very cheap.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Thank you for your answer!

I will check the Vcore voltage and I'll report back later. But although I get voltage on the capacitors just below the VRMs, I don't in the MOSFET themselves (can be that I don't know how to measure them as I should haha).

As for that cap/inductor/unknown-motherboard-object, couldn't it be a major problem? The other systems have it, and it was the only component at least visually damaged. Isn't it possible that is an inductor like Schorsch87 said? Without the schematics I'm too ignorant to identify it

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Well, here it is. I didn't find the usual places to measure VCore, so I kinda improvised, I don't know if I was doing it right.

Hope this image is useful.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

I would say, that on this places inductors have to be, because of the naming with L114 and L101. As for testing I would solder in some fuses to check whether the mainboards starts up.

Could You please give me some insight over the choice of components to try the replacement? I am having difficulty on identifying the inductor specifications, as I cannot find anything on the component which could help me identify it. I believe it's a SMD ferrite inductor, but no more than that.

[momaka]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Since the motherboard seems to only run for a few seconds, try this:
Plug the 24-pin ATX connector to the motherboard but do NOT plug in the 4-pin/8-pin CPU power connector. Then try powering on the board like that. Do the fans and PSU stay turned on? If yes, you might have a short-circuit on the CPU VRM. In such case, checking the MOSFETs would be the first step. Usually bad MOSFETs develop a low-resistance between Gate and Source or Gate and Drain... though not always, of course. When checking for short-circuit on MOSFETs on the Gate connection, it is actually a good idea to keep the CPU in, as any short-circuit to Drain or Source can show the resistance of the CPU in some cases (and that would indicate a bad MOSFET.) However, when checking Drain-Source for short-circuit, have the CPU removed so that your lower MOSFETs don't appear as a "fake" short-circuit from the CPU's low resistance.

Could You please give me some insight over the choice of components to try the replacement? I am having difficulty on identifying the inductor specifications, as I cannot find anything on the component which could help me identify it. I believe it's a SMD ferrite inductor, but no more than that.

Correct, they are SMD ferrite beads.
For testing purposes, you can just solder a jumper across each, as these are not directly responsible for power conversion. But apart from that exception, DON'T solder jumpers in place of inductors - particularly VRM ones, as that will burn other components.

[Schorsch87]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Could You please give me some insight over the choice of components to try the replacement? I am having difficulty on identifying the inductor specifications, as I cannot find anything on the component which could help me identify it. I believe it's a SMD ferrite inductor, but no more than that.

Hi,

momaka said it. For testing you can use a thin wire or maybe a fuse for bridging the contacts.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Thank you both a lot! Going to give a try to all your suggestions.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Good afternoon!

I was able to get a fully working motherboard for comparison.

I tried your suggestions! Starting by trying to plug just the 24pin ATX connector without connecting the 8pin CPU connector, the motherboard behaved exactly as before. I mean, it started, no beeps, four seconds, shuts off. Because the CPU and Case fans are connected to the 12V power rail that feeds the CPU (and as expected they did not start), I connected one fan to the HDD fan connector (which is powered by the 24-pin cable rails). The fan also behaved as expected (starting when the board powered up, shutting down after that).
This might be the regular behaviour of the motherboard (shutting down after 4 seconds if no CPU detected, what would mean the rest of the board is OK), as the good board does the same.

I also soldered a really thin wire to the inductors' pads. As soon as I started the motherboard, it "exploded"/shortcircuited right away.
I searched online for how to identify inductors in order to replace this one, however I believe I haven't the tools required. Or is it possible to do it with a multimeter?

So, I still believe the problem is confined to this CPU section of the board. Although I now have another board to compare, voltage measures are going to be difficult without a schematic and with the board shutting down every 4 sec. Any suggestion on how should I address this?

Again, thanks a lot for your inputs!

[dicky96]

Re: Dell Precision T3500 - 09KPNV motherboard repair

wasserprojekt
When you say you used a 'really thin' piece of wire, just how thin was that?

If these are ferrite beads like Momaka says then you should use a piece of wire similar thickness to the others. Try something like the leg off a resistor. Actually no, take some off one of the other boards for testing purposes.

To clarify, how many good boards and bad boards do you have now?

I like the way you measured Vcore - I assume you did that with no CPU fitted? In which case all the caps that read 0V across them are probably Vcore as you would expect about 1V

You also could do this with the CPU fitted (see pic, it will make it clearer)

You have six phases on Vcore
Each phase has an inductor
Each phase has two mosfets
One mosfet in each pair (phase) will have source connecting to 0V
The other mosfet in each pair (phase) will have drain connected to 12V
The other Source and Drain connect together and also to the inductor
You can check this all out with meter on resistance mode, make sure I have the mosfets the right way round on the pic.
Once you find the junction of the mosfet pair and coil, this is where you can measure Vcore

Now what Vcore do you have on a good and bad board with CPU fitted?

Do you have a POST analyzer card? They can be useful sometimes...

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Thanks a lot!

Haha, I now understand it was really thin. Like a "hair" of wire. No surprise it has gone bananas.

I have one good and one bad board. I acquired the good one for comparison purposes.

All the measurements I did on the caps behind the board were done with CPU installed. I have to mention that the CPU *barely* heats , although I have no idea if that's consequence of low or none voltage (most probably) or if it does not have time to heat up.

As soon as I can have my hands on the board again, I will measure all the components on that area in both dead and good motherboard and I'll upload them here.

Unfortunately, I didn't have the opportunity of buying a PCI analyzer card, which I'll promptly do.

Note: All this time, I had my CMOS battery disconnected, because the piece of plasting holding it was broken. Although I also tried to power the PC with it held by a piece of tape, I did all my testing without it. Hope it has no relation with all of this.

Many thanks again!

[momaka]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Haha, I now understand it was really thin. Like a "hair" of wire. No surprise it has gone bananas.

Yes, the wire has to have substantial thickness to carry the current downstream. In the case of the CPU VRM, since the CPU can draw up to 100-150 Watts with some of these 1366 CPUs, you're looking at roughly 8 to 15 Amps of current flowing.

However, looking at the picture again, I see the blown ferrite bead is close to another inductor and pair of MOSFETs, suggesting this may be the CPU V_tt VRM. CPU V_tt doesn't need to provide as much power as CPU V_core, but you should still use thicker wire - 18 AWG or lower (i.e. 0.75 mm^2 or higher.) Also, check the MOSFETs in the CPU V_tt VRM without the CPU installed - that is look for short-circuit between Gate and Drain or Source pins, but also for short-circuit between Drain and Source pins. If it all checks out OK, check the voltage the same way dicky96 suggested above. I'm actually not sure if the Drain of the upper MOSFET will be connected to 12V rail or 3.3V rail (and hence the ferrite beads too), but you can check that with a multimeter. If it connects to the 3.3V rail, expect the current draw to be slightly higher, so use maybe 16 AWG / 1 mm^2 wire.

Note: All this time, I had my CMOS battery disconnected, because the piece of plasting holding it was broken. Although I also tried to power the PC with it held by a piece of tape, I did all my testing without it. Hope it has no relation with all of this.

It probably isn't the issue here... but IIRC, some Dell motherboards did have an issue with the CMOS battery not present. So if you can, try to solder at least a temporary battery holder connector on there, just to eliminate that as a possible source of trouble too.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Some measurements I did to the damaged board.
Couldn't find the 1.49V and 1.86V on the caps behind the processor anywhere.

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Yes, the wire has to have substantial thickness to carry the current downstream. In the case of the CPU VRM, since the CPU can draw up to 100-150 Watts with some of these 1366 CPUs, you're looking at roughly 8 to 15 Amps of current flowing.

However, looking at the picture again, I see the blown ferrite bead is close to another inductor and pair of MOSFETs, suggesting this may be the CPU V_tt VRM. CPU V_tt doesn't need to provide as much power as CPU V_core, but you should still use thicker wire - 18 AWG or lower (i.e. 0.75 mm^2 or higher.) Also, check the MOSFETs in the CPU V_tt VRM without the CPU installed - that is look for short-circuit between Gate and Drain or Source pins, but also for short-circuit between Drain and Source pins. If it all checks out OK, check the voltage the same way dicky96 suggested above. I'm actually not sure if the Drain of the upper MOSFET will be connected to 12V rail or 3.3V rail (and hence the ferrite beads too), but you can check that with a multimeter. If it connects to the 3.3V rail, expect the current draw to be slightly higher, so use maybe 16 AWG / 1 mm^2 wire.


It probably isn't the issue here... but IIRC, some Dell motherboards did have an issue with the CMOS battery not present. So if you can, try to solder at least a temporary battery holder connector on there, just to eliminate that as a possible source of trouble too.

Thanks a lot! I am going to try to boot the good motherboard without the battery to check if it's required and I'll try the thicker wire!

[Tambayan]

Re: Dell Precision T3500 - 09KPNV motherboard repair

Thanks a lot! I am going to try to boot the good motherboard without the battery to check if it's required and I'll try the thicker wire!

any update ??

[wasserprojekt]

Re: Dell Precision T3500 - 09KPNV motherboard repair

any update ??

Hi!
I had to abandon this project because I lacked the time to do it. But it seems it is a common problem this motherboard: there are lots of boards with the VRMs and other components around the CPU replaced. It seems it's beyond my abilities, as I could never find the schematics for this board.
Thanks a lot!