Roman555 wrote on 2024-04-18, 18:37:

Any power problem can cause active RESET signal ! Usually ICH gets PWROK (power OK) signal from VRMs and sets PLTRST (platform […]
Show full quote

kotel wrote on 2024-04-18, 15:37:

Re-flowing the mosfets did not resolve the issue. Any ideas?

Any power problem can cause active RESET signal !
Usually ICH gets PWROK (power OK) signal from VRMs and sets PLTRST (platform reset) into inactive state as an answer. Sometimes RESET for PCI bus (PCIRST) is controlled by a SIO chip instead of ICH.

If I were you I would also check typical problems: USB data lines (using ohmmeter to measure resistance of the each line to GND rail) and loose contacts under CPU socket and/or MCH (to put something soft under a mainboard in a necessary place and press slightly a heat-sink before powering a mainboard on)

Intel 865 Chipset Customer Reference Board Schematics

All USB lines are okay (I get 1round 15k ohms on data an more than 2k on 5v). Tried putting a metal backplate but same results.

DudeFace wrote on 2024-04-21, 16:05:

ive had this issue with a couple of boards, i wouldnt have thought 2 buldging caps would prevent the board from working , i had an msi board with eleven 6.3v 1000uf caps, all buldging or leaking and the board still worked fine, tho caps can fail even if they look good so you can never be sure, those caps you have are close so i would try them out but i wouldnt run the board for long with them in. i think your problem is the bios, either a bad flash or its just corrupted for no reason which can happen. ive got a gigabyte board that does it every so often, ill plug it in and it turns itself on to a blank screen, it has dual bios probably for that reason, eventually i manage to get it to boot from the backup bios, then i reflash the bios. i've had other board do the same and its usually bios related, try a reflash or a new bios chip.

Later on in the post I replaced the 2 caps. Problem is with flashing the BIOS. I don't own any DIP32 EEPROM programmers and PLCC32 to DIP32 adapters. Not only that, but the BIOS is soldered on. I don't think a bad bios would cause the CPU to not go out of reset (and probably not get its power since it doesn't even get warm). Any ideas?

kotel wrote on 2024-03-04, 10:53:

Mosfet drivers have 2.97v. Without cpu and measuring with an Metex multimeter, on Q3B2, Q3B1, Q2B1, Q1B1 i get 242 ohms on source, 953k ohms on gate, 174k ohms on drain. On Q3B3, Q2B2, Q2B3, Q1B3 i get 1 ohm on source, 240 ohms on drain, 12m ohms on gate. Any ideas?

CPU VRM looks good from these tests, or at least the MOSFETs are.
If the PWM chips is getting the proper supply voltage and enable signal, along with the MOSFET drivers, then the CPU should be getting power.
Measure the voltage going to the CPU once again. Do this with a CPU and both RAM slots populated. Keep the black multimeter probe connected to a ground (a Molex connector on the PSU is a good place for that.) For the red probe, you can use the Drain pad on any of the MOSFETs Q3B3, Q2B2, Q2B3, Q1B3 for convenience.

kotel wrote on 2024-06-20, 09:26:

Later on in the post I replaced the 2 caps. Problem is with flashing the BIOS. I don't own any DIP32 EEPROM programmers and PLCC32 to DIP32 adapters. Not only that, but the BIOS is soldered on. I don't think a bad bios would cause the CPU to not go out of reset (and probably not get its power since it doesn't even get warm). Any ideas?

You are right. It's too early to tinker with bios if the reset signal is active.
As I understood the motherboard does autostart. Have you tried to turn it off by pressing power button for a short or long time? Have you measured voltage of a power button pin of the front panel?
Next: any burned on-board device (LAN, audio or 1394 chip) can pull reset signal down. So check if the additional devices are not scorching hot.

P.S. kotel, you didn't provide a hardware revision of the motherboard so you can check here if your motherboard supports Prescott CPU:
https://www.cpu-upgrade.com/mb-Intel/D865PERL.html

P.P.S. I googled a little bit and it looks like there is an easy method to make d865perl support Prescott if it doesn't out of the box. It's a resistor 0 ohm near C4B2 label that should be removed. But I have neither the mainboard nor a high-res photo around ATX12V conn to check it visually.

Roman555 wrote on 2024-06-21, 15:14:

You are right. It's too early to tinker with bios if the reset signal is active. As I understood the motherboard does autostart. […]
Show full quote

kotel wrote on 2024-06-20, 09:26:

Later on in the post I replaced the 2 caps. Problem is with flashing the BIOS. I don't own any DIP32 EEPROM programmers and PLCC32 to DIP32 adapters. Not only that, but the BIOS is soldered on. I don't think a bad bios would cause the CPU to not go out of reset (and probably not get its power since it doesn't even get warm). Any ideas?

You are right. It's too early to tinker with bios if the reset signal is active.
As I understood the motherboard does autostart. Have you tried to turn it off by pressing power button for a short or long time? Have you measured voltage of a power button pin of the front panel?
Next: any burned on-board device (LAN, audio or 1394 chip) can pull reset signal down. So check if the additional devices are not scorching hot.

P.S. kotel, you didn't provide a hardware revision of the motherboard so you can check here if your motherboard supports Prescott CPU:
https://www.cpu-upgrade.com/mb-Intel/D865PERL.html

P.P.S. I googled a little bit and it looks like there is an easy method to make d865perl support Prescott if it doesn't out of the box. It's a resistor 0 ohm near C4B2 label that should be removed. But I have neither the mainboard nor a high-res photo around ATX12V conn to check it visually.

Indeed, the mobo autostarts. I can turn it off and off using the power button by holding it for long. Power button voltage is 3.25. None components seem to be running scorching hot. According to the link you gave me, my board doesn't support the CPU, but switching it out for a supported one (celeron 2.50ghz 400 fsb 128kb from another dead board that displays nothing on post card) ends up with same results. Revision is AA C27646-213.

momaka wrote on 2024-06-21, 00:35:

CPU VRM looks good from these tests, or at least the MOSFETs are.
If the PWM chips is getting the proper supply voltage and enable signal, along with the MOSFET drivers, then the CPU should be getting power.
Measure the voltage going to the CPU once again. Do this with a CPU and both RAM slots populated. Keep the black multimeter probe connected to a ground (a Molex connector on the PSU is a good place for that.) For the red probe, you can use the Drain pad on any of the MOSFETs Q3B3, Q2B2, Q2B3, Q1B3 for convenience.

Measuring as you said (only with CPU in, didn't want to risk damaging my precious RAM sticks) I get nothing on the pads of Q3B3, Q2B2, Q2B3, Q1B3. From my understanding, there could be an broken trace going to the mosfets you said. Any ideas?

Measuring Q3B3, Q2B2, Q2B3, Q1B3 I'm only getting 12v on gate. Rest of the pins don't have any voltage. I don't think the mosfets would be damaged, but who knows what the previous owner was doing with it. Any ideas?

kotel wrote on 2024-06-25, 09:39:

Measuring Q3B3, Q2B2, Q2B3, Q1B3 I'm only getting 12v on gate. Rest of the pins don't have any voltage. I don't think the mosfets would be damaged, but who knows what the previous owner was doing with it. Any ideas?

Are they top (high-side) or bottom (low-side) mosfets? It's from a i865 reference schematic so labels (or even more) of mosfets may be different:

P.S. Have you checked voltage on VCC pin of a PWM controller ADP3168 ?

nevermind.

Change of behavior when i apply pressure onto ADP3168. Got it once to E1 (as I remember) and was beeping about ram, or at least what I remember. Then there was weird clicking noise when i apply pressure to it, almost as like a relay switching. Now, after switching the PSU on, I hear the same switch noise, then the mainboard shuts down and becomes unresponsive. Narrowed the noise down to the PSU itself. Any ideas? Maybe its a short?

Did some tests. Narrowed the (what I suspect) short to the cpu VRM. Guess I maybe solved 1 issue, but made some more...

Roman555 wrote on 2024-06-25, 11:24:

Are they top (high-side) or bottom (low-side) mosfets? It's from a i865 reference schematic so labels (or even more) of mosfets […]
Show full quote

kotel wrote on 2024-06-25, 09:39:

Measuring Q3B3, Q2B2, Q2B3, Q1B3 I'm only getting 12v on gate. Rest of the pins don't have any voltage. I don't think the mosfets would be damaged, but who knows what the previous owner was doing with it. Any ideas?

Are they top (high-side) or bottom (low-side) mosfets? It's from a i865 reference schematic so labels (or even more) of mosfets may be different:

P.S. Have you checked voltage on VCC pin of a PWM controller ADP3168 ?

I believe low side mosfets (from my understanding high side are the ones connected to 12v cpu power, low side is the one connecting directly to cpu socket). As I wrote in earlier posts, I get 3.2v on ADP3168. Any ideas?

kotel wrote on 2024-07-02, 08:29:

Roman555 wrote on 2024-06-25, 11:24:

Are they top (high-side) or bottom (low-side) mosfets? It's from a i865 reference schematic so labels (or even more) of mosfets […]
Show full quote

kotel wrote on 2024-06-25, 09:39:

Measuring Q3B3, Q2B2, Q2B3, Q1B3 I'm only getting 12v on gate. Rest of the pins don't have any voltage. I don't think the mosfets would be damaged, but who knows what the previous owner was doing with it. Any ideas?

Are they top (high-side) or bottom (low-side) mosfets? It's from a i865 reference schematic so labels (or even more) of mosfets may be different:

P.S. Have you checked voltage on VCC pin of a PWM controller ADP3168 ?

I believe low side mosfets (from my understanding high side are the ones connected to 12v cpu power, low side is the one connecting directly to cpu socket). As I wrote in earlier posts, I get 3.2v on ADP3168. Any ideas?

Really? I looked twice but didn't find. Anyway, don't you think VCC=3.2v is too low to supply ADP3168 ? I looked into the schematics - the pin 28 (VCC) is connected to some sort of 12V through a 10 Om resistor there. The same is in the datasheet.
It's better to check the ADP3168 VCC circuit of the mainboard. Good luck.

Roman555 wrote on 2024-07-03, 10:27:

Really? I looked twice but didn't find. Anyway, don't you think VCC=3.2v is too low to supply ADP3168 ? I looked into the schematics - the pin 28 (VCC) is connected to some sort of 12V through a 10 Om resistor there. The same is in the datasheet.
It's better to check the ADP3168 VCC circuit of the mainboard. Good luck.

Someone in earlier posts said that ADP1368 needs to get at least 3.10v to work, so I think the voltage i get is kinda low (as I understand it should be 3.3v). I don't understand how am I supposed to measure VCC circuit out of the mainboard. I also don't understand how applying pressure to the chip made the VRM shorted (powering it on without CPU power connector doesn't shutdown the PSU). Any ideas?

kotel wrote on 2024-07-03, 16:52:

Roman555 wrote on 2024-07-03, 10:27:

Really? I looked twice but didn't find. Anyway, don't you think VCC=3.2v is too low to supply ADP3168 ? I looked into the schematics - the pin 28 (VCC) is connected to some sort of 12V through a 10 Om resistor there. The same is in the datasheet.
It's better to check the ADP3168 VCC circuit of the mainboard. Good luck.

Someone in earlier posts said that ADP1368 needs to get at least 3.10v to work, so I think the voltage i get is kinda low (as I understand it should be 3.3v). I don't understand how am I supposed to measure VCC circuit out of the mainboard. I also don't understand how applying pressure to the chip made the VRM shorted (powering it on without CPU power connector doesn't shutdown the PSU). Any ideas?

Just try to understand where should the chip supply from. Trace the wire from the pin 28. Disconnect all power cables from the board and measure continuity between the pin 28 and power rails (+12V P4, +12V , +5V, etc and also GND) with a multimeter in ohm-meter mode.
I saw the Intel reference schematics and the ADP3168 datasheet so I suppose the VCC is 12V.

Traced the VCC to be connected t0 12V CPU connector (getting 10 ohms on it). Reading with nothing plugged in are the following:
to GND - 869 omhs
to 12V CPU connector - 10 ohms
to 5V - 2K ohms
to -12v - above 20M ohms (don't think it should be that high)
to 3.3V - 823 ohms
to 12V - 8.1K ohms
to 5VSB - 3.8K ohms
Measured with the good quality metex multi meter. I told you the wrong VCC voltage. No idea how, but I told you the voltage of pin 11, which was also wrong 😀 Nothing on pin 28, but that is expected since there has to be a short somewhere in the VRM (why else would it short out the PSU?). Any ideas?

kotel wrote on 2024-07-04, 07:52:

Traced the VCC to be connected t0 12V CPU connector (getting 10 ohms on it). Reading with nothing plugged in are the following:
to GND - 869 omhs
to 12V CPU connector - 10 ohms

Then voltage is going from 12vP4 through 10 ohm resistor as expected.

IMO, 869 ohm to GND is not short but too low. I think it should be kilo-Ohms or mega-Ohms. But I don't have the same model to be sure. Maybe some capacitors on the 12vP4 rail are bad...

P.S. You didn't have to do other measurements as you found the real source of voltage (+12vP4) for the chip.

kotel wrote on 2024-03-04, 10:53:

Karbist wrote on 2024-03-04, 08:20:

It would be better to take measurement without the cpu in the socket, there are some weird readings in your post, source of high side mosfets is connected to drain of low side mosfets and you say they have 2.4 and 4 ohms. also source of low side mosfets is connected to the ground and you say it has 1.9 ohms, maybe your multimeter is acting up

here's my last suggestion, check pin 3 on each mosfet driver, it must have at least 2.5v or the output of driver will be disabled.
pin3.jpg

Mosfet drivers have 2.97v. Without cpu and measuring with an Metex multimeter, on Q3B2, Q3B1, Q2B1, Q1B1 i get 242 ohms on source, 953k ohms on gate, 174k ohms on drain. On Q3B3, Q2B2, Q2B3, Q1B3 i get 1 ohm on source, 240 ohms on drain, 12m ohms on gate. Any ideas?

Did you measure it to GND? Does 12m ohms mean 12 mega-ohms? Then there're no shorts to GND in this part of the VRM.
But you measured 174k ohms on drains of the high side mosfets Q3B2, Q3B1, Q2B1, Q1B1 .
The drains are connected to 12vP4. How did it get so low as "to GND - 869 omhs". it's almost the same circuit.

Roman555 wrote on 2024-07-05, 14:32:

Did you measure it to GND? Does 12m ohms mean 12 mega-ohms? Then there're no shorts to GND in this part of the VRM.
But you measured 174k ohms on drains of the high side mosfets Q3B2, Q3B1, Q2B1, Q1B1 .
The drains are connected to 12vP4. How did it get so low as "to GND - 869 omhs". it's almost the same circuit.

Yes, I measured it to GND and 12m ohms means mega ohms. As to how I got 896 ohms, I have no idea. Guessing I'm gonna have to re-measure. Also, the 12VP4 only supplies power to the VRM, right? I also got an idea to de-solder the ADP1368 and power the mainboard back on to see if it shorts the PSU. From there I suppose we could see if the chip is the culprit, but I have no idea if running the board without it would cause even more damage...

kotel wrote on 2024-07-05, 17:24:

Roman555 wrote on 2024-07-05, 14:32:

Did you measure it to GND? Does 12m ohms mean 12 mega-ohms? Then there're no shorts to GND in this part of the VRM.
But you measured 174k ohms on drains of the high side mosfets Q3B2, Q3B1, Q2B1, Q1B1 .
The drains are connected to 12vP4. How did it get so low as "to GND - 869 omhs". it's almost the same circuit.

Yes, I measured it to GND and 12m ohms means mega ohms. As to how I got 896 ohms, I have no idea. Guessing I'm gonna have to re-measure. Also, the 12VP4 only supplies power to the VRM, right? I also got an idea to de-solder the ADP1368 and power the mainboard back on to see if it shorts the PSU. From there I suppose we could see if the chip is the culprit, but I have no idea if running the board without it would cause even more damage...

You don't have to desolder ADP3168 to watch it is shorted or not as you can just desolder the resistor 10 Ohm in the VCC circuit.
I'd like to warning you not to power any mosfets if their gates are not connected to control schematics (eg. a pwm-controller or a driver were desoldered) because the mosfets will burn.

Roman555 wrote on 2024-07-06, 09:05:

You don't have to desolder ADP3168 to watch it is shorted or not as you can just desolder the resistor 10 Ohm in the VCC circuit.
I'd like to warning you not to power any mosfets if their gates are not connected to control schematics (eg. a pwm-controller or a driver were desoldered) because the mosfets will burn.

Thanks for warning me what happens when I run the VRM without pwm controller. The only 10 ohm resistor I could find is near the ADP1368, so I suppose I should desolder that one?

kotel wrote on 2024-07-07, 18:10:

Roman555 wrote on 2024-07-06, 09:05:

You don't have to desolder ADP3168 to watch it is shorted or not as you can just desolder the resistor 10 Ohm in the VCC circuit.
I'd like to warning you not to power any mosfets if their gates are not connected to control schematics (eg. a pwm-controller or a driver were desoldered) because the mosfets will burn.

Thanks for warning me what happens when I run the VRM without pwm controller. The only 10 ohm resistor I could find is near the ADP1368, so I suppose I should desolder that one?

Before desoldering check that the resistor's pads are connected to the VCC pin and +12V P4 rail using a multimeter in ohm-meter mode.
Yes, I think it should be near ADP3168. It's a SMD resistor so it's better to use some magnifying glass or glasses to make desoldering proper.
Usually the SMD component is stayed soldered to one of its pads but it is just shifted or rotated to break chain. It's convenient as such tiny thing is easy to lose.