Replacing OST caps on Microstar KT6 Delta motherboard

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Can someone please help me find a replacement output choke and a couple of mosfets since i'm getting ready to make an order for parts

(reason for replacing these parts is that i broke the choke and the mosfets have diode readings of 0.034v and 0.072v, on diode test mode)

If those readings are taken with the MOSFETs on the board, you probably have something else affecting them as these don't look very meaningful. But anyways, I guess it is better to grab a few spare MOSFETs than to ponder whether the ones on the board are currently shorted or not.

i can't find any replacement choke
HERE, that is 1.1uF / 25A
(although there is lots of toroidal inductors that look like it)

The choke is labelled on the vrm schematic as 1.1uH / 25A

^Wrong link?
Whatever you linked to does not list 1.1 uH, 25A choke. In fact, the lowest I see for through-hole inductors is 5.6 uH, which is a bit too high. SMD does have 1.1uH, but the current is way too low. However, I did find an SMD inductor rated for 1.2 uH and 25.4A here:
http://nz.element14.com/bourns-jw-mi...-20/dp/1180547
That should work, though, you may have to solder leads to it (preferably very thick ones) to attach it to the board (actually, it would probably be easier to solder the thick leads to the board and then the inductor to them). Kind of ghetto, but should work.

Also, I suggest you count the number of turns of wire on the good inductor and post it here. Should help with finding a proper one easier... or worst case scenario, use that info to pull something from another dead motherboard. Whenever you see inductors on PC hardware that might say something like "aRb", that means a.b uH (where "a" and "b" are single-digit numbers). Similarly, "Rab" translates to 0.ab uH.

Or worst worst case scenario - run the board with just half the VRM phases - i.e. with the broken choke (and associated MOSFETs) removed from the board. But a high power CPU could overload the other working VRM phase... so yeah, best saved as a last resort.

Both mosfets have this printed on them:
D432AN
FDB
6670AL
i think that it is a fairchild mosfet

So, would it be OK to use THIS,
as, in the component description, there is a note saying that this is a replacement for it.

^Yup, that's a Fairchild MOSFET. Part number is FDB6670AL. The MOSFET you linked to is a bit under-specced on the DC continuous current, but it has a higher pulse current, and same Rds(on) resistance, so it should work just fine. However, do note that what you linked to is a TO-263-3 case and not simply a TO-263. That means, the middle pin on the MOSFET is not cut off, so you probably will have to cut it yourself. But that shouldn't be a problem.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thanks

Those readings were taken with the mosfet legs desoldered.
They both switch on and off OK using the multimeter diode test mode.
But when i tested the mosfet diode reverse bias, those were the reading that i got - so was suspicious, because one was double the value of the other one, and the other two mosfets (15N03L) in the VRM both give reading of around 0.5V

That was the right link.
It was provided it to show where i was looking for the choke and that there was none that had a similar inductance/current rating, and that perhaps i had interpreted the inductance wrongly.
Shirt outa luck HERE also

Have provided a photo of the choke
It has 5 groups of 3 windings

i don't have any other broken motherboards with similar chokes
Just a gigabyte board with different sized chokes and windings (2nd thumbnail)

Was surprised that there was no chokes anywhere near the stated 1.1uH value - which usually means that i'm doing something wrong.
Thought that they would be reasonably common - but am now suspecting that they would have being custom made

Short of using that surface-mount choke, would it be better just to put this board away until i find a broken motherboard with a matching choke?

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Was surprised that there was no chokes anywhere near the stated 1.1uH value - which usually means that i'm doing something wrong.
Thought that they would be reasonably common - but am now suspecting that they would have being custom made

Most likely, very often they are custom made.

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Those readings were taken with the mosfet legs desoldered.
They both switch on and off OK using the multimeter diode test mode.
But when i tested the mosfet diode reverse bias, those were the reading that i got - so was suspicious, because one was double the value of the other one, and the other two mosfets (15N03L) in the VRM both give reading of around 0.5V

Did you discharge the Gate before testing the MOSFET diode? If shorting all pins with your finger doesn't do it, try shorting Gate to Source with a piece of metal, then try measuring again.

That was the right link.
It was provided it to show where i was looking for the choke and that there was none that had a similar inductance/current rating, and that perhaps i had interpreted the inductance wrongly.

Well, the schematic calls for 1.1 uH, so you did interpret it right.
Given that most inductors have 20% tolerance, going up a notch to 1.2 uH shouldn't be a problem at all.

Have provided a photo of the choke
It has 5 groups of 3 windings

The 3 wires are all grouped together to form one thicker wire. Looks like a 5 turn inductor.

i don't have any other broken motherboards with similar chokes
Just a gigabyte board with different sized chokes and windings (2nd thumbnail)

The chokes on the Gigabyte look like 5-turn as well. If the choke toroid is the same size (well, more or less), this may work, especially since that board looks like it's from the same era.
But IMO, that's more of a crapshoot than using the 1.2 uH coils. Plus, if you are buying the replacement MOSFETs, you might as well grab two of those coils (and I say two so that you can replace both on your motherboard, just in case something doesn't work out).

Thought that they would be reasonably common - but am now suspecting that they would have being custom made.

It may or may not be custom made. However, I often find that many motherboard manufacturers usually make one VRM design, and then they use it on almost every motherboard that they make in that given era. So it's very unlikely that the design on your specific motherboard model is unique. Most likely, other Microstar motherboards from that era use the same coils.

Short of using that surface-mount choke, would it be better just to put this board away until i find a broken motherboard with a matching choke?

It's unlikely that you will find a perfect-matching choke (actually, chokes may sometimes look a bit different, but they could still have the same inductance and current handling capability). The surface-mount choke should work... the motherboard just won't look quite "factory made". But there's beauty in that too .

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thankyou for that
i looked at that surface-mount inductor, and it looked quite different - so was skeptical.
But now i feel confident, and combined with your last statement, inspired, that is OK - unfortunately minimum order is 77

Sorry for those mosfet readings - must have being getting a bit tired when doing that.
It wasn't ohms, and also wasn't the diode reverse bias - i was talking shit

OK,
i shorted the gate to drain, and they all return a reading from 0.5v to 0.6v for "forward voltage drop of mosfet body diode" (voice in mosfet testing vid) - which is similar to the other two.

What i initially should have said, was that after turning the mosfets on (+ve to G and -ve to S), and then returning the +ve lead to D, i got the following readings:
0.084v
and 0.039v
The other two (which, when the pins are soldered to the mobo, would have their drain connected to the input caps) give readings of around
0.509v
and 0.549v

Does that indicate any cause for concern?
i now notice that there is a consistant difference between both couples of mosfets of about 0.04v, and perhaps that is due to the gate still being soldered to the board?

i'm kinda getting the feeling that these mosfets are testing OK

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

I guess it is better to grab a few spare MOSFETs than to ponder whether the ones on the board are currently shorted or not.

The mosfets are not shorted

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

It still doesn't hurt to have spare parts, especially if getting more quickly will be inconvenient or pricey. When I buy parts I typically buy more than I need.

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thankyou for that
i looked at that surface-mount inductor, and it looked quite different - so was skeptical.
But now i feel confident, and combined with your last statement, inspired, that is OK - unfortunately minimum order is 77

Lol, now that's a downer.
I'll look through my box-o-coils and PM you if I find any suitable ones. Maybe easier if I just mail you a few .

OK,
i shorted the gate to drain, and they all return a reading from 0.5v to 0.6v for "forward voltage drop of mosfet body diode" (voice in mosfet testing vid) - which is similar to the other two.

Sounds reasonable.
In these tests, I suggest shorting G to S, though, and not D. Shorting G to D may not always discharge the Gate properly.

Does that indicate any cause for concern?
i now notice that there is a consistant difference between both couples of mosfets of about 0.04v, and perhaps that is due to the gate still being soldered to the board?

Yes, quite possible.

i'm kinda getting the feeling that these mosfets are testing OK

Well, I do recall your motherboard was turning ON (or at least running the fans and the PSU, that is). If a MOSFET was shorted, that would usually shut down the PSU as well.
I have spare MOSFETs as well, though they are the smaller TO-252 ones and not TO-263.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Yes, i definitley did short G to S
(at the time of typing that, i was thinking of the schematic, rather than the physical pins, and got mixed up - sorry)

Thanks

Will be on to this some time next week
Might even temporarily just use the old broken choke - as i'm thinking that it should still work, even if a section of the bottom of the ferrite? ring is missing.

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Might even temporarily just use the old broken choke, as i'm thinking that it should still work even if a section of the bottom of the ring is missing.

I wouldn't, in that application it could throw the circuit stability out of whack and cause more problems.

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Might even temporarily just use the old broken choke - as i'm thinking that it should still work, even if a section of the bottom of the ferrite?

No, you can't use that choke with the broken core. Any time you introduce or increase the air gap in a core, that changes the inductance. If you had all of the peices from the broken choke and you glued them back very tightly, the inductance may or may not change too much, and the motherboard *might* just work, but I still would discourage you from doing that (unless you feel experimental ). Though I doubt anything major is going to happen if you mismatch the inductance - most likely the voltage regulation will be off and the motherboard will refuse to boot.

Anyways, I checked my stash... and all I have is 4-turn chokes from Xbox 360s. Don't know what inductance they are, but the turns are made from three thick wires, so at least they should be able to handle the current.
I also have 1.2 uH chokes from an ASUS LGA1156 motherboard (they are on CPU VRM), though I don't know their current-handling capability and they look a bit small.

I guess if you can't find anything, let me know. I can either pull out and send you those 1.2 uH chokes or I can send you the 4-turn chokes from the Xbox 360s and you can either re-wind them with 5-turns with the wire you have from your broken choke or you can use the 4-turn chokes as-is (in that case, you must replace both on your motherboard, becase it would be better if they match).

I also have MOSFETs and caps - again, Xbox 360 pulls, but they are all working. Could send them with the coils.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Thanks, momaka, for the advice, and the offer to send the parts.

Yes, i thought about re-gluing the broken section, but only after throwing it out.
Doh!

Will first overcome a reluctance to deal with humans, and fear of rejection, and go to the waste recycle depo and ask if i can have a scrounge.
That makes a lot more sense, than you posting parts.

Will post back here when i have put the new caps and chokes on.

Cheers

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Removed the other choke, and then tested the other (non-VRM) side of the chokes to ground:
27 ohms with no CPU installed
4 ohms with cpu installed

Looking at the pictures for this board, I don't see a separate buck circuit for the Northbridge, so it is possible this is one of those few boards that powers its Northbridge from the CPU V_core rail. Therefore, don't be alarmed if the "lower" MOSFETs appear shorted even with the CPU removed.

So, that 27 ohms is possibly quite normal.

Pulled a choke from an artifacting nividia video card, which had same size toroid and number of turns, but slightly thicker wire, and used that to replace the one that i destroyed.
Tested the caps, and they all had good capacitance and ESR - so put them back in.
Re-soldered the VRM mosfets, since they were not shorted, and tested in agreement with the schematic (also noticed that the output mosfets are paralleled with another pair that are not on the schematic)

Put the CPU and heatsink back on, and connected power supply.
Pushed power button and the CPU fan was spinning and the keyboard flashed it's lights as normal.
CPU got hot
Also the green front panel power LED lit up
Checked the voltage to the CPU it was 1.68V which seems about right power for a Barton 462 processor (Spec = 1.65V)
But that was it, - no beep, or boot into BIOS
Powered off with power button

Tried again several times - No change
So turned it off, cleared the CMOS with the jumper (just in case that was causing the problem)

Now the motherboard doesn't respond when the power button is pressed (no CPU fan and CPU cold)
Checked that the CPU was OK in another motherboard - yep OK
Removing the 12V CPU plug makes no difference

Something is stopping the power button from starting the PSU
There is 3.3V on the power button positive pin.
On the ATX female connector, there is 254 ohms to ground on the 3.3V rail.
i compared with a couple of other MSI socket A boards and there was no connection to ground whatsoever on the 3.3V rail.
So perhaps this is a hint as to what is causing the problem?

Odd that it was half working, until i reset the BIOS.

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Now the motherboard doesn't respond when the power button is pressed (no CPU fan and CPU cold)
Checked that the CPU was OK in another motherboard - yep OK
Removing the 12V CPU plug makes no difference

Something is stopping the power button from starting the PSU
There is 3.3V on the power button positive pin.
On the ATX female connector, there is 254 ohms to ground on the 3.3V rail.
i compared with a couple of other MSI socket A boards and there was no connection to ground whatsoever on the 3.3V rail.
So perhaps this is a hint as to what is causing the problem?

Odd that it was half working, until i reset the BIOS.

Do any of the fans in the PC twitch or move for a second when you press the power switch?
- If yes, you have a short-circuit somewhere. That 254 Ohms to ground on the 3.3V rail could be it.
- If no, rest CMOS again. maybe even try without CMOS battery. Also check your CMOS reset jumpers - if you forgot it in the reset position, the motherboard may not boot. Lastly, check voltage going to the Southbridge and LPC/Super IO chip.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

No, the fans do not twitch - they remain absolutely still.
Yes, have tried resetting the CMOS several times now - with, and without, the battery.
The jumper was returned back the "Keep Data position"

There is voltage on the PSU green wire of 4.97V
5VSB : 4.98V

Winbond W83697HF LPC chip:
VCC Pin 5 = 0V
VCC3V Pin 22 = 0V
VSB Pin 99 = 4.98V
VBAT Pin 102 = 3.60V
AVCC Pin 105 = 0V
(Just noticed that there is dots on the board that aid in the counting of the pins - that makes things a lot easier)

Think that i found the rail that powers the South Bridge
The MSI schematic shows that the VCC33 pins of VT8235 (the one on this 2.0 version board is VT8237) is connected to an electrolytic capacitor, EC44 - and there is no voltage present there

Also checked the voltage at the Bios Jumper, and get 3.60V to ground

P.S. What is meant by low pin count? - obviously, it's not referring to the number of pins.

[Agent24]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

https://en.wikipedia.org/wiki/Low_Pin_Count

Do you get something between 3-5v on the PWSW+ connector at front panel?

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

3.28v

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Think that i found the rail that powers the South Bridge
The MSI schematic shows that the VCC33 pins of VT8235 (the one on this 2.0 version board is VT8237) is connected to an electrolytic capacitor, EC44 - and there is no voltage present there

VCC33 on the VT8235 in the schematic connects to VCC3. And if you look at the ATX connector on page 43, you will see that VCC3 is the 3.3V rail on the PSU. Since the PSU is not turned ON, then there would be no VCC33 going to the VT8235.

That said, the VT8235 has two other type of power pins: VSUS25 (which connects to VSUS2_5 on chematic) and VSUS33 (which connects to 3VDUAL point on schematic) - as given on page 18 and 19. Check what voltages you get for these two.

3VDUAL can be found on capacitor EC14 (supposedly a 1000 uF, 10V cap), and is generated either by Q16 when system is powered OFF/standby or Q25 when system is ON. (Page 42 bottom.)
VSUS2_5 is derived from 3VDUAL through Q39. (Page 39 bottom.) However, I am not sure if VSUS2_5 should be present without the PC being powered ON. Therefore, I think only 3VDUAL should have 3.3V on it at all times, regardless of what state the PC is in.
Same goes for VCC2_5 as for VSUS2_5.

[socketa]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Appreciate you checking that out, and the explanation.

VSUS2_5 is derived from 3VDUAL through Q39. (Page 39 bottom.)

Am assuming that you meant page 42?
(Just checking that we are looking at the same document)

EC14 has 3.3V on it
And EC36, which is connected to VSUS2_5 (page 42) has 2.15V on it

Same goes for VCC2_5 as for VSUS2_5.

VCC2_5 (same page) is connected to EC31, which has 0V on it.

Something that may, or may not, be worth mentioning:
When i was checking EC14 +VE to ground, i heard a noise like a small fly that had being hit with fly-spray when it's spinning on its back, and it stopped as soon as i removed the probe
When i tested it again, there was no noise
No smoke or smell.
I had the backside of the board vertical, and towards me, and the PSU and CPU fan on the other side, so they were obscured from view. Swapped the CPU fan just to see if that was causing a fault - but changing it made no difference.
No flys or fly-spray in the house.
Possibly just a fluctuation of the quantum vacuum?

[momaka]

Re: Replacing OST caps on Microstar KT6 Delta motherboard

Am assuming that you meant page 42?
(Just checking that we are looking at the same document)

EC14 has 3.3V on it
And EC36, which is connected to VSUS2_5 (page 42) has 2.15V on it

Ah yes, page 42 indeed. (I did it again! Thinking of one thing and typing another. Maybe I need to start taking )

Anyways... So with 3.3V on 3VDUAL, then the SB (VT8235) is at least getting power there.

Now, that voltage you got on VSUS2_5 seems a bit low. You should get +/-5% at the most. So that's 2.375V for the lower limit. I don't know if that is indicating problem with something pulling too much power from VSUS2_5, thus causing it to sag, or if the LM358 (U12A) is bad (or maybe some voo-doo black-magic causing your multimeters to be off?... even though we already checked that they work right in the ASUS P5VDC motherboard thread). I guess you should check the reference voltage (VREF1_25) on that LM358, pin #3. I think VREF1_25 is generated by that MSI-5 ACPI controller.

Also, while looking at that LM358 op-amp, I kind of wonder what voltage does "9VSB" have. It seems like another voltage generated by that MSI-5 IC. I can see, however, that when the PC is running, that "9VSB" becomes 12V because of diode D11.

All of this is on page 42, by the way (both for my reference and yours)

VCC2_5 (same page) is connected to EC31, which has 0V on it.

Well, I will only *guess* that VCC2_5 is meant to be off, since we do have 2.5V on VSUS2_5 (well, not quite, but it still counts, I suppose).

Something that may, or may not, be worth mentioning:
When i was checking EC14 +VE to ground, i heard a noise like a small fly that had being hit with fly-spray when it's spinning on its back, and it stopped as soon as i removed the probe
When i tested it again, there was no noise
No smoke or smell.
I had the backside of the board vertical, and towards me, and the PSU and CPU fan on the other side, so they were obscured from view. Swapped the CPU fan just to see if that was causing a fault - but changing it made no difference.
No flys or fly-spray in the house.
Possibly just a fluctuation of the quantum vacuum?

I'll go with that last one. I literally have no clue what could have happened there . Still, observations like that are always worth noting nonetheless, as that implies you are a careful tech with attention to detail.

I sometimes get quite annoyed (and eventually amused too) by having to fix something after other careless "tech" had worked on it. If I see missing or misplaced screws, that's usually a good hint of what to expect .