P55T2P4 (AT) (untested... But from shop with warranty)
with Matrox Millenium (untested...)
K6-2 400 (untested)
64MB EDO RAM 70ns (Untested)

VGA (LCD) screen was tested though...

Power supply is a converted ATX bought from a shop, brand new - maybe they messed up the conversion?

Meanwhile, I will test the video card and ram on my working rig.

DraxDomax wrote on 2021-07-04, 13:47:

I hope it's one of those big 3-legged chips with the heatsink. If I am not mistaken, these would be some power converters which I probably have in my kit somewhere...
PS: that voltage jumper setting is a "secret" setting for 2.4v (K6-2 400)

I do not think you are supposed to jumper two of those voltage settings, where did you learn about this secret setting ?

Sorry to hear that.🙁
Visually I can't find any faults.
Check your PSU for correct voltages first.
If that is ok honestly I think you should take the motherboard to the shop and make use of your warranty. You'll most likely have to solder to troubleshoot your motherboard and I don't think the shop is going to take the board back if it has been tampered with. Judging from your question I think this repair is currently outside of your abilities. Although I absolutely encourage you to learn how to do component level repairs on your retro hardware a board without warranty is probably a better place to start.

Btw. did you start the board inside or outside the case? Is it possible anything got shorted on the back of the board?

fyi, cpu fan connector seems to be GND/12v/GND

Horun wrote on 2021-07-04, 14:16:

DraxDomax wrote on 2021-07-04, 13:47:

PS: that voltage jumper setting is a "secret" setting for 2.4v (K6-2 400)

I do not think you are supposed to jumper two of those voltage settings, where did you learn about this secret setting ?

Setting two of those jumpers to get lower voltages was a common approach on early socket-7 boards. It works fine if each of the jumpers enables one resistor for the upper resistor of the reference divider for an LM317/LT1086-like regulator. If you set two jumpers at the same time, you get two resistors in parallel, resulting in a lower total resistance which will create a lower output voltage. If the "secret" setting is indeed meant for that board, it is unlikely to be the cause of the problem.

However, since the board didn't have "official" K6-2 support, vendor might not be too willing to honor warranty. 🙁 But I've run k6-2 at 6x75 on these before and not had a failure... with younger caps though.

Yeah My guess is a bad cap or two which shorted the Mosfets......
I was just wondering where that is documented for this board to jump those two for the 2.4v, can't find anything.

http://www.thg.ru/howto/20000725/print.html

Well... If I must admit... The case I had only had 3 holes for standoffs, 2 near the IO shield and one in the center... I drilled and tapped the case for extra standoffs on the other side...
One idea I had was to use that hole the horizontal 3-legged-thingy (rightmost on the picture, labelled 9636 something) has - run a screw through that hole into the standoff... As both a support to that corner and a "heatsink" from the metal "plate" this has, via a metal screw, to a metal standoff and to the metal case...

Would that have shorted this component?

As for my general electronics knowledge, I did buy a lot of stuff (and I mean like a respectable nearly professional lab!) to study electronics properly and down to individual component analysis... Just still learning... I am at Ohm's Law for now 😀
I used to know a lot more, as I was on an electronics route in my high school... But I respect the field too much to assume I remember anything correctly and I am starting from zero.

As for the <2.5v "secret": http://www.thg.ru/howto/20000725/print.html

I think I'll start by removing the electrolytic caps and test for their capacitance? I guess if they are gone, the reading should be quite off, right?
I'll order these... Pretty sure they are transistors... I remember these models, they are quite common and nice to have around anyway

As for the CPU fan, I asked about this... The red this time means 12v, the yellow is some "7v speed control" or something... Which can be ignored and connected to GND...
Unless I have been given the wrong advice and grounding that yellow wire was wrong...

I turned it on inside the case. Before installing, I took a look at the underside, I was particularly interested in the standoff - they were all off the grid, from what I could tell.

I don't think it was a short from the back.

BitWrangler wrote on 2021-07-04, 14:53:

http://www.thg.ru/howto/20000725/print.html

Thanks ! Darn, my TP4 is rev 2 not a rev 3 and is missing parts.

DraxDomax wrote on 2021-07-04, 15:01:

One idea I had was to use that hole the horizontal 3-legged-thingy (rightmost on the picture, labelled 9636 something) has - run a screw through that hole into the standoff... As both a support to that corner and a "heatsink" from the metal "plate" this has, via a metal screw, to a metal standoff and to the metal case...

Would that have shorted this component?

If I understand you right this would have most likely created a short circuit. The tab of the transistor is internally connected to its drain. There's most likely either Vcore or Vio there. You connected the tab to ground. It would explain the situation you described. 🙁

Doornkaat wrote on 2021-07-04, 16:05:

DraxDomax wrote on 2021-07-04, 15:01:

One idea I had was to use that hole the horizontal 3-legged-thingy (rightmost on the picture, labelled 9636 something) has - run a screw through that hole into the standoff... As both a support to that corner and a "heatsink" from the metal "plate" this has, via a metal screw, to a metal standoff and to the metal case...
Would that have shorted this component?

If I understand you right this would have most likely created a short circuit. The tab of the transistor is internally connected to its drain. There's most likely either Vcore or Vio there. You connected the tab to ground. It would explain the situation you described. 🙁

I had the same idea because once I accidentally shorted the heatsink of a VRM mosfet on an S7 board with the case and got a similar 'pooff' effect.

yeah I am not very consistent with listening to myself.
You think I didn't see how this can be a very stupid thing to do? - I did!
And thought "nah it's so cool I am doing it anyway"

My problem has never been lack of ability to learn but a very shaky reasoning process.

I hope it's just local parts gone. I'll order the electrolytic caps for the entire board and those MOSFETs... I'll take things out, test and replace.
I will consult with my engineer friend to make sure I do this right.

As long as the traces and IC's are fine, I am going to fight this 😀

Most likely it will be fixed by replacing (some of) the three transistors with the black heatsinks. The NDP603AL basically just got bypassed and it's probably connected directly to a larger internal power plane (and not a small trace) so one of the three transistors gave up before any copper did. Also if my theory is right all overvoltage that could have been passed when whatever failed failed should have just been grounded.
But since you have an engineer friend you're going to get a much better assessment from him. 😎👍

Aaand since you voided the warranty you have a first board to fix and a place to start learning component level repairs after all. 😁😉

Lol, I keep wanting to describe my relationship with the seller (a large business selling PC parts)... Long story short: the fact I installed and messed up this board is unrelated to the fact they still owe me a "Tested. Working P55T2P4 rev 3.10". It's a long story with lots of defences why I am not the douche here, of course 😉
They've basically been sending me junk for $80. If I get one working board from them and manage to salvage one of the other boards they tried to get away with, I don't see it that I've made undue profit.