IBM 5170 motherboard issue

[romanon]

Hello, i obtaned this 5170 board, but it not work, no beeps, no signal on monitor, only keyboard blinks on power on. I tried 8 bit EGA card, 8 bit VGA card and 16bit VGA card but Nothing. PC speaker also not typicaly cracks on power on.
Also tried two power supplies. Any help?

 

[Stone]

Check the 12V and 5V lines for shorts. If there is a short on the motherboard e.g., capacitor, it would shut down the power supply almost immediately after you turn the power on.

 

[romanon]

Power supply is running after power on for any time.

 

[Stone]

The PSU fan doesn't count. It can sometimes run even after a PSU shutdown of the 5V or 12V supply.

 

[romanon]

Ok, i have one observation, maybe its nothing but for sure. I noticed that one pin on isa slot was wry and touched oposite one (on picture) . I did not notice it and powered it on. Later i did notice it and i corrected it, but nothing changes.

 

[Stone]

When I said to check 12V and 5V for shorts I meant to check with an ohmmeter. :smile:

You might also have a faulty ram chip.

 

[modem7]

I suggest the following:


STEP 1

The 5170 motherboard will not start unless there are at least two things from the power supply:

• +5 volts; and
• POWER GOOD signal

Using a voltmeter (or multimeter), verify that both exist.
The +5 volt line should measure +4.5 volts or above.
The POWER GOOD line should measure +3 volts or above.

Diagram at [here].

If both do not measure as expected, measure again after:
1. Pulling out all cards from the motherboard; and
2. Disconnecting hard drive and floppy drives from power supply.


STEP 2

Verify that the two BIOS ROMs are in the correct sockets (i.e. U27 chip in U27 socket, U47 chip in U47 socket).
Verify that the two BIOS ROMs are corectly orientated.
Re-seat the two BIOS ROMs (in case of poor electrical contact with their sockets).

Photo at [here].


STEP 3

Your motherboard has one bank of RAM (consisting of 18 chips). If any of those chips is faulty, the motherboard will appear to be 'dead'.

Note that rather than a RAM chip actually failing, what might have instead happened is that a RAM chip has developed poor electrical contact with its socket.
To see if that is the case, re-seat all of the RAM chips (wiggle each chip in its socket).


STEP 4

See if a POST card displays anything.

Note: If one or more motherboard RAM chips has failed, the POST card will display DD and some other numbers. See [here].

 

[romanon]

Ok, i have little progress here. After all, i tried Supersoft/Landmark Diagnostic ROMs for IBM 5170 from my friend.

This is result (attachment)

Its seems to by fine, but freeze on PROTECTED MODE CPU. I waited for 1/2 hour but nothing changes.

It is some CPU fault or?

 

[Scali]

I'm not sure
The 286 was notorious for being able to switch TO protected mode, but not being about to switch BACK
There are custom solutions for that, basically involving the chipset doing a 'warm reboot' of the CPU, then bringing the context back up.
Perhaps the mechanism fails on your system... or the software is not compatible with the solution on your board... or the software is not supposed to be able to switch back on a 286 anyway. Did you try it on another 5170? And if so, did anything happen after protected mode?

Either way, I think the issue will be small. For regular DOS, it shouldn't affect anything. Most software did not even use protected mode at all, and apparently regular 8086-compatible mode works, else you wouldn't have gotten that far.

It could still be memory-related as well. In protected mode, the CPU can address up to 16MB.
Perhaps enough of your memory works to run regular DOS applications, but as soon as it goes into the extended range, you still hit faulty memory?

 

[romanon]

I dont know if you fully understand me, this software is runnig from bios eprom chips, not from DOS. If i use IBM ROM chips, motheboard appears to be "dead".
And no, i dont have any 5170 MB or CPU to test with it

 

[modem7]

The Landmark/Supersoft diagnostic ROMs are known to work on an IBM 5170 motherboard.

Late last year, member bobba84 acquired a 5170 motherboard for which the same PROTECTED MODE CPU test failed, however 10 seconds into the test, he saw "FAILED".
The post about that is [here], and is followed by some discussion.

I understand that the POST in the IBM BIOS does a similar test (stopping the motherboard from going any further).

 

[Scali]

I dont know if you fully understand me, this software is runnig from bios eprom chips, not from DOS. If i use IBM ROM chips, motheboard appears to be "dead".
And no, i dont have any 5170 MB or CPU to test with it

Ah I see...
So with those ROMs it fires up (which means the CPU and memory are at least mostly functional), but there is a slight issue with pmode.
With the IBM ROMs, it may or may not fire up, and may or may not lock up during POST, which may or may not be related to the pmode issue that the diagnostic ROMs have uncovered.

I guess your next step would be to obtain a POST card? Then you can track how far the IBM BIOS routines manage to go.

 

[romanon]

I obtained nothing, only freeze screen with Landmark/Supersoft diagnostic and PROTECTED MODE CPU without result. Nothing more, no beeps no posts, nothing. Is there a chance to buy a new CPU to resolve this problem?

 

[Scali]

I obtained nothing

No, I mean, if you want to debug this further, I think you will want to use a POST card, so you can keep track of how far the BIOS gets.

Is there a chance to buy a new CPU to resolve this problem?

Personally I don't think the CPU is the problem. CPUs rarely break down, and if they do, they usually break down completely, so you probably wouldn't get as far as getting the Supersoft diagnostics running, and passing so many tests (just getting some text on screen already requires most of the CPU ot work). I think it's something on the motherboard that is failing.

 

[romanon]

What do you mean wtih "using POST card"? I dont understand

 

[Krille]

What do you mean wtih "using POST card"? I dont understand

See this.

 

[dkedrowitsch]

I've had the BIOS ROMS fail on 5170s before. Since it shows life when you put the diagnostic ROMS in, I'd be willing to bet one of your ROMS are bad. Do you have spare blanks and a device programmer?

 

[modem7]

I've had the BIOS ROMS fail on 5170s before. Since it shows life when you put the diagnostic ROMS in, I'd be willing to bet one of your ROMS are bad. Do you have spare blanks and a device programmer?

If only one (or two) of the IBM BIOS ROM chips was bad, then the SuperSoft ROM diagnostics (which sit in the motherboard in place of the IBM BIOS ROM chips) would have executed successfully.

 

[romanon]

One more think, my PC speaker was broken so i couldnt hear Supersoft/Landmark Diagnostic error beeps, but i am little confused now. I hear first seconds 6xhi-lo 1 short (cant initialize monitor), then agin 6-1, but then, 6-2...
Where is problem? Problem is, that this combination is not described in Supersoft/Landmark Diagnostic manual. When i install original IBM ROM chips, there are no beeps as i wrote..

 

[modem7]

When i install original IBM ROM chips, there are no beeps as i wrote..

That is expected. The situation you are in is shown at [here].

One more think, my PC speaker was broken so i couldnt hear Supersoft/Landmark Diagnostic error beeps, but i am little confused now. I hear first seconds 6xhi-lo 1 short (cant initialize monitor), then agin 6-1, but then, 6-2...

I have been confused many times by the Supersoft/Landmark ROMs. I am sure that the beeps will be intended for those people who do not have a suitable video card. I think that you should ingore the beeps.

Where is problem

The Supersoft/Landmark ROMs show a failure of the 'PROTECTED MODE CPU' test. If you were to buy a POST card, and see what that displays when the IBM BIOS ROMs are in place, I think the POST card will show a code that indicates failure of the same test.

It might (repeat: might) be the CPU. Have you tried re-seating the CPU? Do you know someone who can loan you a working CPU ?

As you will have read, there is other circuitry on the motherboard that can cause failure of this test. Specifically, the circuitry under suspicion is involved in resetting the CPU (done as part of taking the CPU from protected mode to real mode). That circuitry is shown in the diagram at [here] - a reset initiated via the keyboard controller. That diagram is for a type 1 motherboard, not the later type that you have, but the circuitry will be similar. If you have a digital oscilloscope (and knowledge of to use it), I can illustrate what to look for.

Best to eliminate the CPU as the cause first.