MSI K7T Pro2 Ver1 - capacitors have higher values?

[Acerman123]

Trying to fix this board, MSI K7T Pro2.

All Tayeh capacitors showed either as damaged or with uf values which were faf from what they should be so I have replaced them and all other capacitors except those big ones near CPU.

The motherboard powers on but doesn't post. My PCI bios code reader will most of the time dsplay some some kind of code if there has been no power to board for a while but I don't think those codes mean anything in this situation. If I power off the board and try powering it on the second time the code reader usually doesn't show anythin. Unplug power supply and try again and usually there is some kind of code.

I removed those three Yageo capacitors and my tester claims that all of them have higher uf values than they should. I thought that capacitors lose capacity with age? Are those Yageo caps likely to be faulty? They should have 3900uf but I recall one of them showing up as 4800uf or something like that?

I have not been able to CPU but ram and graphics card I removed from this setup were all fine.

[stj]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

caps that read high are screwed, internal resistance is fooling your meter.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Okay, I removed those four other big caps and all of them also show more than 4700uf, three of them over 5000uf and one was 4900uf...

I replaced two G-Luxons before and when measuring they also showed more than rated capacity so I guess none of these electrolytes were good. Tayeh caps were definetely bad.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

This is the tester I use.

I do have another Socket A board so I could test that Duron, but that other board has Athlon 1400 and its core is in quite bad shape, reinstalling the big heatsink it has could destroy it.

[momaka]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Electrolytic caps have +/- 20% tolerance for their capacity, meaning they can show 20% lower or 20% higher than the stated capacity on their sleeve and still be considered OK.

Of course with crap cap brands such as Tayeh or 2nd tier brands like Teapo, I'd say anything reading over +15% higher of the rated capacitance on the sleeve should be considered suspect and replaced, regardless of the ESR shown. Teapo SC in particular have a bad habit of deteriorating with high internal charge leakage... which usually results in them showing higher capacitance on a meter.

On the other hand, the green United Chemicon KZE should still be OK. KZE series are very good. Their reading of 5036 uF is only about +7% of stated (4700 uF) capacitance, so they are well in spec.

Probably best to focus on the rest / any remaining Teapo and Tayeh caps on the board, including small ones under 470 uF.

Also do this test: put POST card in motherboard and remove RAM, but do install CPU (with heatsink and thermal compound, of course - this is a socket A CPU and it will burn out without one!) Then power on the motherboard and see if it beeps at you with a memory error. POST card should hopefully get stuck on the same POST code too.

I'm assuming you already tried removing CMOS battery. Most boards from that era have no problem POSTing without a CMOS battery, but nevertheless, do reinstall a good CMOS battery back in there, just in case.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Those bigger ones are the only electrolytes I haven't replaced so far since I don't have replacements yet.

I have tried various configurations, without either cpu, ram or gpu. I have tried with different ram sticks / slots and different gpu. Power supply is known good one since it runs that Athlon 1400 board just fine. It is still true that I haven't tested that Duron and I haven't tested another cpu in this MSI board. Externally that Duron looks fine.

Only time when the code reader displays something is when the board is powered on while it has been completely disconnected from power source for a while. Codes are sometimes corrupt so I don't think they itself mean anything but the fact that board needs to be without power for a while for some type of code to appear might mean.

EDIT: I have not tried to reflow solder joints in general to see if some of them would be bad, on AGP slot etc.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Disconnected CPU fan, powered on the board for 5 minutes while gripping the heatsink and I don't think it got warm at all.

[momaka]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Some socket 462 boards had a feature that wouldn't run the CPU VRM and/or sound an alarm through speaker if the CPU fan is disconnected. So for s462 boards, CPU fan always needs to be connected just in case. If you want to see if CPU gets warm, just remove the CPU fan from the heatsink, but leave CPU heatsink attached to CPU/board. Then see if heatsink is warm. If not warm, measure CPU VRM output voltage. Should get about 1.5-1.75V depending on CPU model.

Anyways, all the big caps need to be replaced before anything else (aside from the Chemicon KZE caps.) Otherwise, hard to say if issue is caused by the caps or if there is something else wrong with the board.

Once the caps are replaced, then see if anything changes with the POST card. If not, measure the various voltage rails on the board (i.e. CPU VRM, Northbridge Vcc, RAM Vdd, RAM Vtt, and etc.)

BTW, also try different PCI slots for the POST card. At least from what I have found with mine before is that I don't always get the same (or any) codes on specific PCI slots on some motherboards.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Heatsink didn't get warm with fan removed from heatsink but still connected.



Unfortunately I lack the knowledge how to measure mosfets etc., but I decided to try something... I did put multimeter in DC mode and place negative lead on the center pin and then tried poking left and right pins. I did get values approx values -3.3v and -5v with 2 left mosfets but nothing on the 2 right ones, picture attached.

[Per Hansson]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

It sounds then like the CPU is not getting any voltage.
I would recommend to measure the voltage at the bottom of the board at the capacitors:
A couple will be for VRM input (seems to be +5v in your case from your readings).
And then the capacitors closest to the CPU will have the VCore voltage (there may be nothing in your case).

[momaka]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Heatsink didn't get warm with fan removed from heatsink but still connected.

I concur with Per then - CPU may not be getting power... but let's check anyways.

Unfortunately I lack the knowledge how to measure mosfets etc.,

No worries!
Never too late to learn.
A multimeter is more than sufficient. And older motherboards like this are usually simpler to understand compared to newer ones. The usual MOSFETs you'd see on such are TO-252 / DPAK and TO-263 / DDPAK (their shape and case size/type.) These always have pins arranged in the same way. With MOSFET pins pointing towards the bottom when you look at it:
middle (cut) pin / tab (pins #2 and 4, respectively) = Drain
right pin (pin #3) = Source
left pin (pin #1) = Gate.

You only need to measure the voltage at the Drain (pin 4 is usually easier) and Source (pin 3) with respect to ground. To measure properly with respect to ground, place the black multimeter probe on one of the metal screws that's holding down the motherboard in the case. Or, if the motherboard is not in a case, use the pad around any of the screw holes on the motherboard. All of these are Ground points.

And also, just beware that some TO-263 and TO-252 devices on the motherboard may actually be linear voltage regulators / LVDO's and not MOSFETs. On older motherboards like this one, you can recognize these by the part number on the device - anything with *1084, *1085, *1086, *117, *217, *317, and *1117 will usually be a voltage regulator. These have different pinout configuration from what was described for the MOSFETs above.
Usually, but not always, pin 3 (right pin) = input and pin 2 & 4 (middle pin or tab) = output. These are really the only two pins you need to measure with respect to ground to see if the voltage regulator is working properly. Input should always have greater voltage than output. If not, the regulator might be bad or could be using a non-standard pinout. The 2nd most common pin config is middle pin / tab is tied to ground and output is the left pin.

but I decided to try something... I did put multimeter in DC mode and place negative lead on the center pin and then tried poking left and right pins. I did get values approx values -3.3v and -5v with 2 left mosfets but nothing on the 2 right ones, picture attached.

If you're getting negative voltages, then you're probably not using the correct ground for your measurement. See above and place black multimeter probe on a proper ground point.

I would recommend to measure the voltage at the bottom of the board at the capacitors:

+1
And this is a good alternative way to do it if feeling uneasy about poking at MOSFET pins that may be too close to each other when you want to measure voltages.

In fact, if you turn the motherboard upside down and measure the voltage on each cap, you can then use a marker / soft pen and write the voltages on the back of the motherboard. Then take a picture of that and the front of the motherboard and post it here. This should be enough to tell what voltages are where on the motherboard and if they are what they should be.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Brief update, I now used screw hole pad as ground and measured.


Picture attached, ones with X I get 5v on the center pin but nothing on any other pin, ones with 0 there's absolutely nothing on any pin.

On another forum someone suspected that caps might be "drifting" and mosfets will not activate because of that? The poster apparently either has similar case at the moment or has had where the caps looked fine externally but are/were still the problem.

[momaka]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Brief update, I now used screw hole pad as ground and measured.


Picture attached, ones with X I get 5v on the center pin but nothing on any other pin, ones with 0 there's absolutely nothing on any pin.

That confirms the CPU VRM (CPU V_core voltage rail) is not running.

This could indeed be caused by the bad caps. But I'm more inclined to think either dead CPU or something else wrong on the motherboard, which can prevent the CPU VRM controller from getting and Enable signal, thus why it's not outputting anything.

Do the following troubleshooting steps next:

1) Try another CPU (with heatsink, thermal paste, and fan, of course) and repeat these voltage measurements on the MOSFETs. Report what you get.

2) If you get same voltage readings like last time (5V on the MOSFETs with X and nothing on the ones with the "O"), disconnect PSU from motherboard, remove the CPU, and measure resistance with your multimeter between the tab (pin 4) of the MOSFETs marked with "O" and ground (screw hole). If your multimeter is manual range, use lowest (usually 200 Ohms) scale. Report what resistance you get. If it's auto, report what resistance you get with the units - i.e. Ohms or KOhms or MOhms.

3) If you can, do as Per suggested and measure the voltage on each cap with the board turned upside down, then report these voltages here. Before the CPU VRM can get an enable signal, first the Southbridge and LPC / SuperIO need to have proper voltages up and running.

Hopefully these three should help us narrow down the problem. But it would also help to replace the Teapo caps on the CPU VRM, along with any other 2nd tier brands (United Chemicon is 1st tier Japanese brand, so the green KZE caps can stay.) And post more pictures of your board too (top and bottom side.)

On another forum someone suspected that caps might be "drifting" and mosfets will not activate because of that? The poster apparently either has similar case at the moment or has had where the caps looked fine externally but are/were still the problem.

Yes, caps can look absolutely fine and still be bad, especially with older 2nd tier brands like Teapo (even more so the SC and SM series.) "Drifting" is not an inaccurate term, but really what is happening is their electrolyte can break down over time, causing either high internal leakage current or higher ESR / impedance.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Brief update: the CPU is working, tested in Asus A7V133-C so problem is in the board.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Brief update, the three Yageo caps (circle) are all 5V and Chemicons (square) are 0V.

Two caps next to ram slots both show 3.3v which I think is the correct voltage for SDRAM.
When measuring around AGP and PCI slots I get either 3.3v or 5v depending on cap.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Update: Replacements for Yageo and Chemi-Con are installed but the situation did not change, 4 caps which surround the CPU and the 2 mosfets show 0V.

So I guess this board has other problems than capacitors, atleast now since it definetely had bad capacitors in the beginning of this repair attempt.

[momaka]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

Do the tests I outlined in 2) and 3) in my last post above and show the results here.

There should be more than just 3.3V and 5V near the PCI slots. If you check all of the MOSFETs and/or voltage regulators on the board, you should also find at least one that is lower than 3.3V - typically 2.5V, 2.1V, 1.8V, 1.5V, or similar. It's usually for the southbridge Vcc and possibly LAN chip too.

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

On mosfets near PCI slots I did get 1.5v and 1.2v, above AGP slot there was 1.5v but I measured the legs of mosfets and not placing ground probe to screwhole.

Using 200 range, the resistance from screw hole to "O mosfets" settles to 32.5

[Acerman123]

Re: MSI K7T Pro2 Ver1 - capacitors have higher values?

I positioned the camera bit badly but when testing those "O mosfets" and holding probes it will stay at approx. 32.5.

https://youtu.be/RoqTv1km700