Oh, I forgot to mention that I've tried two CPUs in it. Ans also the cache RAM gets very hot.

Beware - cache SRAM(s) getting VERY hot could mean they are inserted in the wrong direction!

I would probably remove the cache altogether for the time being if dealing with a dead board, just worry about it later if you can get it working. Bad sram, incorrectly inserted or wrong voltage could be a factor.

Not finding 3.3volts on pci would be normal in this era

I took the cache out. It was inserted correctly, though. I see activity at A0-A7 at the BIOS, although I've no idea if this looks normal:

LeeDerbyshire wrote on 2026-10-22, 17:43:

crushed clips in the ISA bus: I/O CH CK/ and DATA 7 were pushed flat into GND and RESET DRV.
POST codes. In the PCI slot, the -12, +12, +5V, CLK, RDY and FRAME LEDs light up; the +3.3V and RESET LEDs do not.

both ISA and PCI reset is active low so in theory reset shorted to ground would not fry anything, with reset low all the other pins should be inert and also safe from short, in theory 😀

now reset not going high on power on with short cleared is a problem and at least one of the reasons board is dead

LeeDerbyshire wrote on 2026-10-22, 17:43:

ABit AB-PB4 Rev:1.5

can you post good quality picture of the whole board from the top?

LeeDerbyshire wrote on 2026-10-22, 17:43:

There were some crushed-together legs in one corner of the ALi M1487. The affected pins are: RTCAS, IOWJ, ZSMEMWJ and CPUSPD. I don't really know these do.

broken IOWJ would cause no POST codes, other wouldnt affect at least the very early boot steps. But lack of Reset makes all that irrelevant

LeeDerbyshire wrote on 2026-10-22, 17:43:

I have some low-temperature solder paste

how low temp? and why would you need it in the first place?
pins you can solder with normal soldering iron

LeeDerbyshire wrote on 2026-10-22, 17:43:

that I will try with a hot air gun

highly unadvisable on old hardware unless you bake it first

LeeDerbyshire wrote on 2026-10-22, 17:43:

Is it worth persevering with it?

do you have an oscilloscope? a logic analyser? anything to see if there is any activity going on?
Edit: replied before seeing last post 😀
picture is soo small cant tell what you are showing
try pressing/shorting Reset button and looking at POST card reset led

The reset pin does not appear to be going high, assuming I'm testing the correct pin. From the solder side, I'm testing the 4th from the left, 3rd up from the bottom; which I think would make it 3rd from the left, 2nd up at the actual chip pins. I have a couple of cheap scopes, and the ubiquitous cheap 8-channel logic analyser. I can see a clock signal at the CPU which looks more like a sawtooth than anything, and I'm assuming it's okay. I can see some activity at the BIOS chip A0-A7, but it looks like it's all at A4.

The reset LED on the POST board stays off at all times, even when shorting the reset pins.

I wanted to use low temp (138C) solder paste because my hands are not very steady, and I've seen how delicate the pins on the M1487 are when I was trying to straighten them. I'm hopeless at very fine soldering, and I thought a hot air gun at, say, 150C would not do too much damage. I don't normally get shaky until I'm trying to do something under a microscope, then I'm all over the place.

I'll try again with the trace. Uploading it somehow seemed to shrink it. I'm probably doing something wrong.

From the top, this is A0-A7 at the BIOS at power-on. It does this until I switch off. Only A4 seems to really do anything.

LeeDerbyshire wrote on 2026-10-23, 15:47:

The reset pin does not appear to be going high, assuming I'm testing the correct pin. From the solder side, I'm testing the 4th from the left, 3rd up from the bottom; which I think would make it 3rd from the left, 2nd up at the actual chip pins.

I got lost somewhere in the middle 😀
isa https://en.wikipedia.org/wiki/Industry_Standa … SA_Bus_pins.svg
pci https://en.wikipedia.org/wiki/Peripheral_Comp … onnector_pinout
CPU C16 https://flylib.com/books/en/4.57.1.22/3/#:~:t … et%201%20pinout or https://bitsavers.trailing-edge.com/component … ov89.pdf#page=5
chipset https://theretroweb.com/chip/documentation/m1 … 3505.pdf#page=8

Its super hard to get a particular Pentium board diagram, so the best I can do is m1487/9 based industrial board https://www.versalogic.com/wp-content/themes/ … 6-1.pdf#page=73
and Am486 Microprocessor PCI Customer Development Platform https://acer.manymanuals.com/motherboards/486 … -49935/download

Afaik on this Ali chipset there is no reset input and boards manipulate Powergood signal to make chipset reset whole system:
"The switching power supply sends a POWERGOOD signal to the M1487 to generate systemreset and reset the chipsets to initial state."
Check powergood M1487 pin 13, it should flip when pressing reset

LeeDerbyshire wrote on 2026-10-23, 15:47:

I have a couple of cheap scopes, and the ubiquitous cheap 8-channel logic analyser.

good

LeeDerbyshire wrote on 2026-10-23, 15:47:

I can see some activity at the BIOS chip A0-A7, but it looks like it's all at A4.

worrying, but not until we know if reset is performed

LeeDerbyshire wrote on 2026-10-23, 15:47:

The reset LED on the POST board stays off at all times, even when shorting the reset pins.

put a scope or logic analyser on Reset pin on ISA, PCI and finally CPU pin and observe as you press reset. Since you probably have only two hands I recommend soldering wires to those pins 😀

LeeDerbyshire wrote on 2026-10-23, 15:47:

I wanted to use low temp (138C) solder paste because my hands are not very steady, and I've seen how delicate the pins on the M1487 are when I was trying to straighten them. I'm hopeless at very fine soldering, and I thought a hot air gun at, say, 150C would not do too much damage.

that junk solder will, its only good at desoldering components provided you immediately clean off the pins with proper solder
and hot air + old chips is not a good idea without long baking

LeeDerbyshire wrote on 2026-10-23, 15:47:

I don't normally get shaky until I'm trying to do something under a microscope, then I'm all over the place.

make a tripod for the iron and a place to rest the wrist, use big tip loaded with solder, dont push in the tip, blob pins in solder and wait till it takes (preferably with flux), use solder wick to remove the blob

>attachment

thank you

Im looking at the picture and your board now and its missing U16 74f32, U18 ?, U19 RTC, U20? - between ISA slots, BUT Im not sure if those got removed or this revision never meant to mount them
edit: ok I know now, Ali M1987 B1 stepping integrates RTC and probably some more logic.

rasz_pl wrote on 2026-10-23, 17:39:

Im looking at the picture and your board now and its missing U16 74f32, U18 ?, U19 RTC, U20? - between ISA slots, BUT Im not sure if those got removed or this revision never meant to mount them

I can see the images for those on the silk screen, but it looks like they've never been there. I see no evidence of work there, just solder flowed through the vias. I haven't actually found a full image for this board rev 1.5.

I removed part of my post about chipset revisions, I posted before I looked into datasheet, it actually documents integrated RTC, weird
Yes no rev 1.5 pictures, thats why I asked for yours 😀

I checked the various reset pins (one at a time to maximise my dexterity), and none of them were lifted off 0v. If the M1487 is involved in all of this, then here is what it looks like in the corner. I know it looks rough now, but you should have seen it before. These are the pins that I have not dared to resolder yet, so they may not make contact, although I tried to push them down a little - the tracks have come loose, and it's all a bit delicate down there. I incorrectly identified the affected pins before (I offer no excuses, other than my own stupidity). If VCC6 is meant to connect to that via, then it looks a bit dodgy from some angles.

The pins that seem most affected are the bottom 6: GA15 GA16 GA17 GA3 GA2 VCC6

Well, I do get 5V at the chip end of the leg next to that via. If I understand the pinout for the M1487 correctly, pin 20 is CPURST, and I'm assuming that should be outputting 5V, but it never gets above about 1V. Unless I should be looking at SYSRST instead? But that stays at 0V.

>I checked the various reset pins (one at a time to maximise my dexterity), and none of them were lifted off 0v.

now measure what happens on powergood M1487 pin 13, and trace what generates it all the way to reset header

>GA15 GA16 GA17 GA3 GA2
cpu address bus, but doesnt matter for now since there is no reset

edit:
>pin 20 is CPURST, and I'm assuming that should be outputting 5V, but it never gets above about 1V. Unless I should be looking at SYSRST instead? But that stays at 0V.

buzz where those pins connect to, do they lead to CPU reset? maybe CPU sreset C10?

rasz_pl wrote on 2026-10-23, 19:50:

>I checked the various reset pins (one at a time to maximise my dexterity), and none of them were lifted off 0v. […]
Show full quote

>I checked the various reset pins (one at a time to maximise my dexterity), and none of them were lifted off 0v.

now measure what happens on powergood M1487 pin 13, and trace what generates it all the way to reset header

>GA15 GA16 GA17 GA3 GA2
cpu address bus, but doesnt matter for now since there is no reset

There is 5V at one side of the header for the reset button, and 5V at powergood at pin 13. I can't really follow the traces into the board, but there seems to be only a 100 ohm resistor between the two, which must be okay, I guess.

rasz_pl wrote on 2026-10-23, 19:50:

edit:
>pin 20 is CPURST, and I'm assuming that should be outputting 5V, but it never gets above about 1V. Unless I should be looking at SYSRST instead? But that stays at 0V.

buzz where those pins connect to, do they lead to CPU reset? maybe CPU sreset C10?

Assuming I'm looking at the correct CPU pin at C16:

CPURST at M1487 pin 20 has a resistance of about 10K between it and C16
SYSRST at M1487 pin 63 has a resistance of about 150 ohm.

Neither of them makes it to SRESET at C10. Bothe have a resistance of about 7M.

Um, when I'm checking for reset voltage, should I be looking at SRESET at C10, or RESET at C16? Up until now, I've been looking at RESET C16.