Replace the tantalum caps. It's almost totally normal that at least one has gone short on these boards over time.

Hmm, I dont think there is a short. Look on the lights on the POST card, all seems to be working...

What are the two non-lit LEDs on the POST card for?

Good job so far. Since you have a TL866, it has a logic IC test function. If you can find/build an adapter that fits _over_ the ICs you may be able to test them without desoldering. A sort of inverted socket. I have a similar adapter that I can place over eeproms so I can read/reprogram them without desoldering. Looks similar to this (not my picture):

stamasd wrote:

Good job so far. Since you have a TL866, it has a logic IC test function. If you can find/build an adapter that fits _over_ the […]
Show full quote

Good job so far. Since you have a TL866, it has a logic IC test function. If you can find/build an adapter that fits _over_ the ICs you may be able to test them without desoldering. A sort of inverted socket. I have a similar adapter that I can place over eeproms so I can read/reprogram them without desoldering. Looks similar to this (not my picture):

Indeed, thats a very good idea! I think I can build such adaptor with 2 empty IC sockets.

I already ordered some 74LS158, maybe I will test them with the piggyback-method first...

derSammler wrote:

What are the two non-lit LEDs on the POST card for?

They are from l to r:
Reset
Frame
Irdy
CLK

When the PC is running they should blink. The current behavior is i.e. caused by a missing/damaged CPU or BIOS.

Another candidate for checking is the 8288 chip, the bus controller. Unfortunately you won't be able to check that one with the TL866 AFAIK.

Predator99 wrote:

When the PC is running they should blink. The current behavior is i.e. caused by a missing/damaged CPU or BIOS.

Be careful here. A normal POST card does not work on a 5150 mainboard since the BIOS simply does not send the POST code to port 80h. So all the card tells you is that the voltages are apparently present and ok.

If you don't know this site already, start here:
http://www.minuszerodegrees.net/5150_5160/MDC … diag_config.htm

If CPU, lower 64 kb of RAM and BIOS are ok, the system should do something at least. If not, you need to check all traces for continuity.

derSammler wrote:

Be careful here. A normal POST card does not work on a 5150 mainboard since the BIOS simply does not send the POST code to port […]
Show full quote

Predator99 wrote:

When the PC is running they should blink. The current behavior is i.e. caused by a missing/damaged CPU or BIOS.

Be careful here. A normal POST card does not work on a 5150 mainboard since the BIOS simply does not send the POST code to port 80h. So all the card tells you is that the voltages are apparently present and ok.

If you don't know this site already, start here:
http://www.minuszerodegrees.net/5150_5160/MDC … diag_config.htm

If CPU, lower 64 kb of RAM and BIOS are ok, the system should do something at least. If not, you need to check all traces for continuity.

That's a 5160, but essentially correct. The POST codes go to port 60h, not 80h. POST cards don't monitor port 60.
If you have a logic analyzer you could monitor port A of the 8255 http://www.minuszerodegrees.net/5160/post/516 … point_codes.htm

No, it's a 5150. 5160 has 8 ISA slots with less distance between them and no cassette port.

ps: topic title is "IBM 5150 Mainboard repair", btw. 😉

That's true. My bad.

Short update: I built such "over-IC" adapter but unfortunately it doesnt work while the ICs are in circuit.

I also "piggybacked" the rusty 74LS158s but no change.

I am aware that I dont have to expect any numbers on the post card, but some lights should flash on it while Post-sequence is running and there should also be some noise from the speaker when Supersoft-ROM is installed.

I recently was given a 5150 with a 16K-64K motherboard labeled "1840401 456 XM "

Seems dead. I know that a regular diagnostic card isn't fully useful, but when I put it in, I see the LED lights flash for a fraction of second and then go black.

Looking over the board, there was a rusty lump sitting on the side of the P8259a chip.

Could the repair be just as simple as replacing this chip with a generic replacement ?

Here's a picture of the chip after giving it a brushing--

I'm going to try the repair.

If it isn't successful, did any vendors make 3rd party boards that fit a 5150 case?

Looks like this one might have the right slot spacing: https://www.ultimateretro.net/en/motherboards/10191

Looks like the rheostat looking thing is by the power is broken, since 1the top /2 of it fell off.

A 5-30 pF clock chip trimmer ?

http://www.minuszerodegrees.net/5150/misc/515 … ch_settings.htm

Looks like I only need it if I am doing CGA, yes?

Looks like it acts as a calibration for all the clocks:
http://www.minuszerodegrees.net/5150_5160/mis … _generation.jpg

From the schematic you can see it's in series with the main 14.318 crystal:

So if it's failed open then the clock won't run.

snufkin wrote on 2022-03-24, 12:08:

So if it's failed open then the clock won't run.

So if I wanted to over clock, would I go for a higher capacitance or a lower capacitance ?

I don't know much about the analogue side of clocks, but my guess is that the capacitor isn't going to move the frequency very far from the crystal 14.318. I think it's more for making small adjustments. I'd guess that small values will slightly increase the frequency, but go too small and it'll stop oscillating. If you can't find a variable capacitor of the right value then -0° suggests using a 20pF fixed capacitor instead, which might be worth trying just to see if you can get any life at all from the board.

the adjustment exists for the purpose of using that clock as an NTSC colour carrier