Hello all,
Again, great ideas to try.
I am brand new to taking any scope measurements. Please let me know if I did something wrong or if I should change anything. My probe was set to 1x, and I used auto range for the divisions.
Reading off of the OSCO pin on the KBC socket you can see the first photo attached. The clock looks really messy, but I think that is correct. There is both a 6mhz and 12mhz signal on the pin (again, I think).

Reading the clock off of the POST card on the clock LED is my second photo. Again, super messy and this one makes less sense to be messy. I see 8.28 Mhz, but the signal is all over the place.

Onto trying the MMX, I promise no MMX were hurt in the taking of this photo. The system was powered on for maybe 3 seconds. I left the heatsink off so as to show it was an MMX and I did verify on VCC2 there was 2.33volts and on VCC3 there was 3.43 volts. The split voltage jumpers are just below the socket. Unfortunately, no change to the post code either using Post1a or the original BIOS (the original BIOS has that C sticker on it and the post1a bios is at the bottom of the photo).

I put the bestkey KBC back into the socket as that doesn't seem to be making a difference. Also, while looking around for a good ground for the scope, it looks like the board's 5v rail is 4.45volts. Does this lead anyone to ideas? Could this be a real problem or is 4.45 volts not low enough to be an issue?

Is the 5V rail from the PSU at 5V without the motherboard connected? Would seem like something is shorted perhaps?

I verified that we are at 4.98 volts when plugging this psu into another test board. Something is pulling down the voltage on our problem board. Capacitor?

One more thing. The resistance between the grounding plane and the 5v rail is 1600ish ohms rather than the 100k ohms I would expect

Maxvintage871 wrote on 2024-04-28, 18:07:

I put the bestkey KBC back into the socket as that doesn't seem to be making a difference. Also, while looking around for a good ground for the scope, it looks like the board's 5v rail is 4.45volts. Does this lead anyone to ideas? Could this be a real problem or is 4.45 volts not low enough to be an issue?

4.45 is too low. Maybe there is a faulty voltage regulator somewhere.

since you have a scope of some kind put it on a 5V rail, is it steady or very spiky?

The 5v rail seems steady, but low. I am still getting 4.45 consistent on my multimeter, but the scope shows closer to 4.

If you suspect capacitors, I can do some recapping. However, don't electrolytics most often fail open? None are leaky or bulging as far as I can tell. There are 3 tantalums and then a million surface mount ceramics. Maybe a shorted ceramic would pull the 5v rail down, but I have to find a way to hone in on at least the area of the shorted cap.

I will take any other measurements you think makes sense.

Maxvintage871 wrote on 2024-04-29, 12:09:

The 5v rail seems steady, but low. I am still getting 4.45 consistent on my multimeter, but the scope shows closer to 4.

If you suspect capacitors, I can do some recapping. However, don't electrolytics most often fail open? None are leaky or bulging as far as I can tell. There are 3 tantalums and then a million surface mount ceramics. Maybe a shorted ceramic would pull the 5v rail down, but I have to find a way to hone in on at least the area of the shorted cap.

I will take any other measurements you think makes sense.

Caps can fail hard and still look like brand new.

What are the voltages coming out of Q2 and Q3? MMX core is 2.8V and 3.3 input/output.

3.45 is fine but 2.33 is low and that could prevent it from working. It wouldn't be a first if the Vcore mosfet failed.
You could desolder it and check +5V again.

I think they are both driven (set V by) the 2 small 8 lead Q26 and Q29 just above. Can you write what's written on top of those 2 small SOP-8?

Where did you measure this 5V? is that right on the AT connector?
btw I just noticed https://theretroweb.com/motherboards/s/epox-pronix-ep-p55-tf doesnt have a picture of your board. Would it be possible for you to remove all components and make a nice high resolution photo from the top and bottom? as high resolutuion and quality as possible, ideally so we can soom on individial chips and have readable markings and small traces. This will also help with debugging.

Its possible PSU is tired/malfunctioning, and its just poor lucky the other board workts with it. That scope reading looks nice and flat at 4V, it should be, Supply should have undervoltge protection and cut the power. Weirdness all around. Do you have another supply?

Nexxen wrote on 2024-04-29, 12:45:

...3.45 is fine but 2.33 is low and that could prevent it from working. It wouldn't be a first if the Vcore mosfet failed.

I see "B-C-E" printed on the transistors, so we´re safe to say these are bipolar transistors (BJT) rather than MOSFETs.
Of course the still can be faulty.

majestyk wrote on 2024-04-29, 15:43:

Nexxen wrote on 2024-04-29, 12:45:

...3.45 is fine but 2.33 is low and that could prevent it from working. It wouldn't be a first if the Vcore mosfet failed.

I see "B-C-E" printed on the transistors, so we´re safe to say these are bipolar transistors (BJT) rather than MOSFETs.
Of course the still can be faulty.

I make a wholesale use of mosfet (and got caught 😀). I didn't have time to check.
But I had that issue recently.

https://pdf1.alldatasheet.com/datasheet-pdf/v … ILD/D45H2A.html

So many new things to try! First, I have some photos of front and back of the board. I will submit these to theretroweb some point soon.

Second, I have odd (at least not helpful) news. I cannot recreate the 4.45v situation again. I have tried everything and after I de-populated the board for photos, no combination of ram/video card/ cpu (mmx or not) , post card, etc has that same low volt reading. POST card still reading its same codes after repopulating.

Next, I do have 2 modern ATX psu. Photos attached. We shall call the Venus PSU1 and the Hercules PSU2.
With nothing populated on the board, psu1 reports exactly 5.00volts at the AT plug. At the turbo LED, it reports 4.8volts. With the P133 installed, we see 4.85volts at the AT plug and 4.67volts at the turbo led.
With the P133 installed, PSU2 reports 4.92volts at the plug and 4.71 volts at the turbo LED. With the MMX installed, PSU2 reports 4.86 at the AT plug.

When configured for MMX (there really only seems to be the one voltage split jumper settings according to the jumper manual from theretroweb. Either VCC2==VCC3 or VCC2 is the lower voltage), we see 2.33volts at VCC2 and 3.42 for VCC3 for both PSU (voltage seems very stable for both PSU on the socket after the regulators).

I can try to unsolder either or both voltage regulators for the socket, but I can't attempt that before Thursday due to schedules.
Let me know if there is anything I missed in this response.

Maxvintage871 wrote on 2024-04-30, 11:51:

photos of front and back

sweet

Maxvintage871 wrote on 2024-04-30, 11:51:

Second, I have odd (at least not helpful) news. I cannot recreate the 4.45v situation again.

where were you measuring that 4.45 before? its possible you measured in wrong spot

Maxvintage871 wrote on 2024-04-30, 11:51:

>4.67volts at the turbo led.

like this, bad spot to measure 😀 any Led output goes thru limiting resistor and will always read lower
if you want to measure in another spot to verify proper power propagation thru the board measure on
- BIOS chip, Winbond one, last pin - right of the ridge
- on Q2 Q3 transistors hmm afaik the tabs (part with a screw) are collectors so there

Maxvintage871 wrote on 2024-04-30, 11:51:

I can try to unsolder either or both voltage regulators for the socket

no need

Hello. I'm sorry for the delay. Lots going on.

Measuring the voltage to the tab of either Q2 or Q3 results in the 3.57v output of the transistor (same measured on both or 2.49 volts for the vcc2 transistor with mmx settings). Measuring to the emitter leg gives 4.96 volts with no cpu. 4.81 with either cpu type installed.

I can't find where I found the low voltage before. Lost to either bad measurement or actual problem that I can't recreate.

Any suggestions for a test?

Maxvintage871 wrote on 2024-05-02, 23:28:

>emitter leg gives 4.96 volts

ah pnp yes

So 5V is not a problem. Dropping more than 5% should have triggered supply shutdown, so that 4.45 was looking bogus. Sadly that leaves us with no clues as to weird post codes 🙁

I was poking around and decided to scope the address lines on the RTC. All address lines on the RTC seem to float flat at 1.26v. This is how it looks immediately on start up. My probe is using 1x. I could easily have the scope settings wrong. Any possible suggestions based on this weird behavior?

https://www.youtube.com/watch?v=3Sfahr9JElw

were you actively resetting while taking measurements? if CPU is crashing early you wont see any activity on X bus (separate bus with rtc/keyboard controller and bios)

I have measured right at the moment of reset and I do see normal activity for an instant. I don't know how to use my scope well enough to get a picture though.
Any other suggestions to try? We might be reaching the end here.

I dont know if VT82885 address lines are connected in parallel with bios.
Try directly on BIOS chip address lines, specifically the lowest ones A0-2 and all data pins. Should be constantly flipping, especially with Deunans post1a bios. We know post1a does "something" as it also produces bogus Postcard output, so in theory it should run at least a little bit.