MS-6590 KT4 Ultra - short on vcore

[carlitosbala]

Hi,
I continue my journey trying to learn how to repair a motherboard, I still have a few after the Netvista ended up working

I am trying to fix an MSI MS-6590 KT4 Ultra, that I discovered has a short on Vcore.

The board turns on without issue if the cpu is not installed, but it turns off after a few seconds with the CPU on.

After analyzing the schematics (attached) I found the vcore output in page 39 (Voltage Regulator Module) is shorted to ground. I took every component referred to in that page out of the board for testing and they are all working, the short is there even without them, on the PHASE1 and PHASE2 pins of the HIP6022 (i also took the 6022 out for testing and it seems to be fine)

So, what should be the next step? Should I take each and every capacitor between vcore and ground out for testing? Should I suspect the SIO? I didn't want to take that one out because I'm not sure I'll be able to put it back, but I guess it's a suspect.

Or does this look like the board is toast and I should simply bin it?

Thanks in advance

[piernov]

Re: MS-6590 KT4 Ultra - short on vcore

Exact resistance to ground with and without CPU?

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

Without: 27.3 Ohm, with: 13.2

Measured it from the chokes in the VRM to a ground point from one of the screws

[piernov]

Re: MS-6590 KT4 Ultra - short on vcore

13.2 ohms with CPU sounds fine. Without CPU it should slowly increase above a few hundred ohms.
No voltage on CPU VCore though?

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

The resistance without CPU doesn't increase, it stays at that value.

I forgot to mention something: if I measure the voltage on pin 2 of the 6022 (PWM2) and turn on the board, I get to read a voltage of 1.8V before the board shuts down.

And just to clarify: I'm not seeing the usual "fan moves for a split second and you miss it if you blink", what I see is the board turning on for a full second or so and then shutting down.

[momaka]

Re: MS-6590 KT4 Ultra - short on vcore

Could we see some pictures of your board?

It could be that the KT400 Northbridge is powered from the CPU V_core, hence the 27.x Ohms resistance without CPU. I know Intel used to do that on a lot of their i845 boards.

If not, then that 27 Ohm resistance could be a partially-shorted ceramic capacitor.
If you remove the two output torroids on the CPU V_core, that should allow you to isolate where the "short" is - i.e. on one of the MOSFETs or on the CPU V_core output itself.

[piernov]

Re: MS-6590 KT4 Ultra - short on vcore

You're right, northbridge is powered from CPU VCore, I didn't think of that. So it's most likely fine.

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

Could we see some pictures of your board?

Attached a photo. Any part you want me to show in more detail?

If you remove the two output torroids on the CPU V_core, that should allow you to isolate where the "short" is - i.e. on one of the MOSFETs or on the CPU V_core output itself.

Could you please explain this a bit more? I mean, i know you mean removing the chokes and testing, but I'd like to learn the background behind the suggestion and what test do you want me to do

Regards

[momaka]

Re: MS-6590 KT4 Ultra - short on vcore

Attached a photo. Any part you want me to show in more detail?

Thanks!

This will do for now, I think.

A couple of things I spotted here...

First, I see possibly problematic capacitors - United Chemicon KZG (6.3V 3300 uF brown caps) and Teapo SC or SM (6.3V 1000 uF ?? green caps). Both of these are know to fail and in fact might have failed without showing visible signs - especially the Teapo. But the KZG usually do bulge. I'm not convinced this is the problem with your board at this point, but it is something we should keep in mind.

The second thing I see - no thermal compound on the CPU die.
When you were testing the board with the CPU, did you fit a heatsink with thermal paste on top of the CPU? If not, then that CPU is probably dead now. AMD CPUs from that era and probably all the way up to the Athlon II era could easily get damaged if ran without a heatsink. So if you did have a heatsink attached, then nothing to worry about here. But if you didn't, that CPU should now be considered suspect and not used for any further tests.

Could you please explain this a bit more? I mean, i know you mean removing the chokes and testing, but I'd like to learn the background behind the suggestion and what test do you want me to do

Well, if you desolder the CPU V_core output toroids, this will separate the CPU V_core power switching MOSFETs from the CPU. Thus, now you should be able to test all the MOSFETs between Source, Gate, and Drain pins without finding any low resistances. If you do still find a low resistance, you may have a shorted MOSFET... though that's usually rare and usually the PSU will shut down as soon as you try to power up the PC from the PWR switch.

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

First, I see possibly problematic capacitors.

Well, I've bought several boards to play with, all in the 486 - Pentium4 range, and my idea is to restore at least one of each generation, including a complete recap for the ones that I can get to work. So replacing those is part of the grand plan

The second thing I see - no thermal compound on the CPU die. When you were testing the board with the CPU, did you fit a heatsink with thermal paste on top of the CPU? If not, then that CPU is probably dead now.

Noted. I do have a heatsink I put on top of the cpus I'm testing when I turn on the power, so I'm not letting it overheat (I hope). I've only had this board+cpu for a couple of months and to be honest I can't be sure the CPU is OK, my only kind-of-assurance is that I have tried several other CPUs on this board with identical results.

Well, if you desolder the CPU V_core output toroids, this will separate the CPU V_core power switching MOSFETs from the CPU. Thus, now you should be able to test all the MOSFETs between Source, Gate, and Drain pins without finding any low resistances. If you do still find a low resistance, you may have a shorted MOSFET... though that's usually rare and usually the PSU will shut down as soon as you try to power up the PC from the PWR switch.

Right, it's obvious after you explain it I was not paying attention hehe

I took the two inductors out, and now with the CPU installed I get "open circuit" to ground in the solder point that's closest to the mosfets, and 13.8 Ohm on the other. The board no longer shuts down with the CPU installed, and the mosfets seem to be fine: they pass a quick test with the multimeter and with the PSU on I can measure the expected voltages on each pin, except of course those that are actually connected to ground

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

I'm about to give up on this one, I don't know what to check, and seriously considered removing each and every capacitor between vcore and ground, but there are just too many of them, and I have never soldered an smd component, so I'll definitely screw up if I go that route. Not that I have anything to lose anyway.

I was looking at the board and found these FM points, listed in page 46 of the schematic, what are they? Can I use them to diagnose parts of the board?

[dicky96]

Re: MS-6590 KT4 Ultra - short on vcore

@carlitosbala
No need to give up.
As Momaka and Piernov have already told you - if you have 13R or 27R on Vcore you do NOT have a short there!

Just look at what ohms law says - I = V/R so if Vcore is 1.8V and R is 13R then I is 0.13A. So just 130mA. That is not a short.

Without the CPU then I is even less 1.8V/27ohms= 67mA

To convince yourself of the truth of this, seeing as you removed the two inductors try the following simple test:

1. remove the CPU then get your bench PSU (you do have one yes?) and set the bench PSU output to 1.8V

2. Connect PSU -ve to Ground and connect PSU +ve to the side of the (removed) inductor that goes to Vcore (the side that goes to all the capacitors and the CPU)

3. How much current does your PSU says is flowing now?

Once you accept you do not have a short come back and we can try some other things. In the very unlikely event that you really do have a short (several amps are flowing at 1.8V) then use your finger and try to find a capacitor is getting hot.

rich

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

@dicky96: point taken. Forget about the short

I don't have a bench PSU yet, I will get one as soon as my budget allows it

As I mentioned before I'm a beginner, that means I may fail to see a problem more experienced technicians would identify immediately. I have no issue admitting it, but I tend to clarify it so the people I ask for advice doesn't assume that I may have done / checked anything in particular

I believe I'm quite good at reasoning, and I can read a schematic, so I hope that if I get a possible lead to investigate I'll get an answer. So going back to the very basics, the question I need to answer is "why is the board shutting down".

If power doesn't seem to be a problem, then there must be a signal somewhere telling the board to shut down. But which kind of signal (present or missing) would do that?

BIOS issues I've seen cause the board to just not start, but anyway I can't reflash the chip at the moment because the eeprom programmer I bought like 3 months ago is yet to arrive. I may just need to eat the cost and buy it locally, which I didn't do in the first place because it's like twice as much as overseas

North/South bridge problems I've seen others describe (haven't diagnosed one myself yet) manifest themselves as the chips overheating as soon as power is applied, which is not the case here.

Another thing is that I've mostly followed (read / watched) people fixing issues in more modern boards, like 775 and newer. There are some components in these older boards that I can't map to what I'm seeing others do. For example, I also have a Socket 370 board that shuts down as soon as it turns on that instead of two mosfets has a mosfet and a schottky diode, and I think what I really need is some good learning resources on how these work...

Anyhow, I'm digressing. What would you suggest me to check?

[dicky96]

Re: MS-6590 KT4 Ultra - short on vcore

That's good analysis and reasoning - yes you are right, you have some signal shutting Vcore down

Without going further directly into your problem, here are two threads I suggest you first:


https://www.badcaps.net/forum/showth...hlight=ms-6178
https://www.badcaps.net/forum/showthread.php?t=93613


This is a very simlar board to yours - no I didn't fix it yet ,but it should give you a very good idea how to move forwards from where you are now and you may have a bit more luck than I did so far - if something is not clear from what I was trying to do, just post back here.

[dicky96]

Re: MS-6590 KT4 Ultra - short on vcore

Actually no ignore those links - that is a PIII board not AMD like yours

1. I would start by having a look around the Vcore PWM controller chip U8

Here is the datasheet
https://pdf1.alldatasheet.com/datash...L/HIP6302.html

Also refer to page 39 of the schematic. See if you can figure out why Vcore is disabling


2. If that does not give you anything conclusive, have a look around the Thermal Protection Circuit on page 41 of the schematic - a false (or real) trigger there could shut down Vcore

3. See if you can find a datasheet for the SIO - it will probably tell you which signal starts up the VRM and what conditions need to be present for that to happen

4. Check the voltages on all the other supply rails (LDO regulators) - are any of the other supply rails switching on and off?

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

Thank you for the follow-up, I did notice the board was a different one (in fact I do have a working board of that model!), but I assumed you referred me to it as an example of the things I could try.

I started looking at the schematics to find signals to verify, so thanks for the pointers. I haven't had time yet to go back to this board, but will resume the troubleshooting soon.

Regards

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

I had a bit more time to check this board again.

I was re-testing what I had already tested, like making sure vcore is actually being generated before the board shuts down, at pin 10 of the U6 chip and in the positive terminal of the capacitors between vcore and ground.

As I don't want to overload my PSU, I shut it down after every test. After one of the tests I forgot to switch it off while I fixed some issues I had with my monitor, and after a few minutes the fan started spinning again, normally.

In fact, I discovered that from a cold start, it would run for a variable time (1-inf seconds), then "shut down" (now in quotes because it doesn't seem to be shutting down at all), but still respond to the power switch: if I short the pins on the switch for 6 seconds, then release them and short them again, the fan starts spinning.

And finally, after a few tests like that I touched the CPU and it was hot. Which means that 1) somehow the board is now sending power to the CPU, which until today had been completely cold, and 2) that cpu is probably dead now due to overheating during these tests

Anyhow, I seated the fan correctly now, and at least for now (who knows if tomorrow when I check it again it won't behave as it was until today) I can get it to turn on reliably. I noticed the northbridge also gets warm after the board is on for some time, but just warm, not hot. I can't get a beep from the speaker, so the board is still not working, of course.

I will try different CPUs, I have a couple that should work for this mobo, but I think I killed the one I was using to test the board.

[momaka]

Re: MS-6590 KT4 Ultra - short on vcore

And finally, after a few tests like that I touched the CPU and it was hot. Which means that 1) somehow the board is now sending power to the CPU, which until today had been completely cold, and 2) that cpu is probably dead now due to overheating during these tests

I know you mentioned in post #10 that you did have a heatsink on the CPU, but did you actually use a socket 462 heatsink? And did you clamp it down?

Reason I ask is because if you used a regular "flat" heatsink not meant for socket 462/370, the foam pads on socket 462 CPUs will raise the CPU heatsink above the CPU core so the heatsink would never touch it and the CPU will burn out. Also, the die of socket 462 CPUs is very small, so not clamping down the heatsink can overheat the CPU as well (in addition to the foam pads usually raising it above the CPU.)

Lastly, make sure you have a proper 3-pin fan plugged into the CPU fan connector. Many socket 462 motherboards have this protection that will either shut down the motherboard or just prevent it from POSTing if a CPU fan is not installed. This is because AMD did not implement proper CPU over-temperature protection probably until socket 939/AM2, if not later. Therefore, a lot of manufacturer's implemented this "feature" to prevent burned out CPUs.

So to recap the above statements:
1) always have a proper socket 462 heatsink installed (with thermal compound)
2) always clamp down the CPU heatsink
3) always have a CPU fan connected to the CPU fan header

I seen these mistakes more times than I care to remember, and even more so now with people trying to build retro PCs but not really being aware that a lot of the old hardware was a lot more "crude" in operation and could easily be damaged if proper installation instructions aren't followed.

I will try different CPUs, I have a couple that should work for this mobo, but I think I killed the one I was using to test the board.

At least you have spares.

I took the two inductors out, and now with the CPU installed I get "open circuit" to ground in the solder point that's closest to the mosfets, and 13.8 Ohm on the other. The board no longer shuts down with the CPU installed, and the mosfets seem to be fine: they pass a quick test with the multimeter and with the PSU on I can measure the expected voltages on each pin, except of course those that are actually connected to ground

Generally, I don't suggest to power up things with parts removed. In this instance, I suggested removing the inductors only if you needed to isolate a short-circuit on the VRM (which you didn't really have anyways.) But after this, always place back the removed parts or replace with proper equivalents.

Otherwise, it's rather pointless to power On the board and try to measure voltages with the inductors removed, because then the CPU VRM will not be outputting a voltage, which means the LPC/SIO will see the missing CPU voltage and possibly shut down some or all of the other power rails on the board. So in essence, you don't know if you will get reliable voltage readings.

Also, don't remove MOSFETs or capacitors and try to power On the board - that will usually end up damaging other components. I'm sure you know this, but I have to mention it every time the topic arises, because I have seen people (even one here, though I can't remember which thread it was) remove bad caps and then try to power On their hardware - that's a big no-no.

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

Just as an update: I got some goodies from the US recently that I have been waiting for, including a cheap ESR meter, a cheap PCI test board, and enough capacitors for the motherboards I still think I can fix, like this one. I could have bought locally but the biggest distributor of electronic components in Montevideo charges more or less the same than what I payed at Newark including shipping.

So I went and recapped this board, and still get the same result. The old caps didn't really show any issues in the esr meter: capacitance was within 15% of the rated value, and ESR... well, I don't know if I should trust the meter, for the same capacitor it gave me like 4 Ohm and increasing in one measurement and ~.5 Ohm in another. Still the average was a bit higher than 1 Ohm for the old caps, where the new ones were consistently lower.

The PCI test board shows all voltages are OK, same for the CLK and RST signals. But I don't get any activity from the BIOS.

So before I get to poke the signals on the chips to see if they are there (and start playing with the hantek 6022 I bought some time ago) I'd like to try reflashing the BIOS, but unfortunately I don't have a programmer: according to the tracking info the one I bought entered the country in early March and still has not been delivered to me. And I absolutely refuse to buy it locally at ~50 USD when I already bought it at 10, so I guess I'll wait for it until Customs decide to release it, that might take some time though, I am waiting packages I bought last year that also had tracking info showing they had arrived to the country

But I digress, I only wanted to give an update since you were kind enough to guide me before


Regards

[carlitosbala]

Re: MS-6590 KT4 Ultra - short on vcore

Alright, I did another testing session today and I *think* I have identified a problem with the ATTP1 chip.

I approached the problem with the idea that it was a thermal shutdown. So I started looking at the signals involved, and searching for "THRM_PSON" led me to this thread, which I still need to read in detail as it has a very good description of how this part of the board works. Mine is not the same model, but almost all the components described connect to the same signals, and is similar enough to *also* not have the RT2 resistor in place, which I understand is supposed to be the actual temp sensor?

What I found is that when the board starts, there is no continuity between pin 7 (VREF) and pin 1 (PWR_DN), but after a few seconds they get shorted, and the board shuts down. Granted, this might be how it is supposed to work, and it might just be a symptom of something else, so I hope you can suggest something but given the number of components around this chip that are not actually in the board (see page 41 of the pdf), I don't really see how these two pins can get connected unless it is either inside the chip, or something else on the board shorting 5VSB and 3VDUAL, but that doesn't seem the case , as pin 8 (5VSB) stays at 5V even after the shutdown.

Anyhow, how can I test my theory? I would think removing the attp1 is probably not a good idea, is it?


Regards