One last attempt to get this motherboard working...

[Divad]

I've got a 16-64K 5150 motherboard here which has been resisting all attempts to repair it. With previous help/advice from this forum I've got it working with the diagnostic ROM, but no further...

Here's what's been done so far:
12V capacitors replaced
Landmark ROM showed faulty RAM, so bank 0 removed and replaced with socketed 4116s (original bank 0 parity chip still fitted)
Landmark ROM then showed faulty U34 (8253), chip removed and replaced

Now Landmark appears to show all RAM tests as passed, but the board will not boot with the original BIOS chip. It powers on but no beeps or other activity.

I'm also getting the errors shown below. The first picture is with just the MDA card fitted, the second is with MDA and floppy card fitted (no drive). Both give a continous high-low beep when it gets to testing MDA memory.

Does anyone have any ideas what I should try next? The MDA card has been briefly tested in another system and appears to work OK.

With MDA card only:


With MDA and floppy card fitted:


Thanks

David

 

[modem7]

The fact that the SuperSoft ROM is executing means that the motherboard is at a repair state where the IBM BIOS ROM should also start (repeat: start) to execute.
What have you done to prove that your IBM BIOS ROM chip is not faulty?

 

[jh1523]

Was the MDA card tested with the landmark bios in the other system? It appears you have some memory errors on the video card (not just what the test bios reports, but it appears you have some corruption in the second picture).

 

[modem7]

...., but the board will not boot with the original BIOS chip. It powers on but no beeps or other activity.

Confirm for us that "no beeps" includes the 'minimum diagnostic configuration' discussed at [here]. Because, if that is the case, there is no point adding cards, because that would achieve nothing; the lower-level problem would need to be fixed first.

 

[Divad]

Confirm for us that "no beeps" includes the 'minimum diagnostic configuration' discussed at [here].

Testing with the MDC originally highlighted the faulty tantalum caps. Now, I get to this stage http://www.minuszerodegrees.net/5150_5160/MDC/no_beeps_2.htm and all voltages measure within limits except possibly the "power good", which is nearly 5V rather than "over 2.4V". I get a single beep immediately at power on but no other activity.

The fact that the SuperSoft ROM is executing means that the motherboard is at a repair state where the IBM BIOS ROM should also start (repeat: start) to execute.
What have you done to prove that your IBM BIOS ROM chip is not faulty?

I'm afraid there's not much I can do to verify if the BIOS is faulty, but I do have two chips (in fact, I got one from you!) and both show the same behaviour.

Was the MDA card tested with the landmark bios in the other system? It appears you have some memory errors on the video card (not just what the test bios reports, but it appears you have some corruption in the second picture).

Unfortunately the other system I have is an XT clone, and I only have the 5150 version (27C64) of the Landmark ROM. I think I would need the 27C128 or 256 version for the clone as the BIOS chip is a 256.

 

[Chuck(G)]

Perhaps trying a generic ERSO BIOS might yield a bit more information and give you a system that boots DOS. Just an idea.

 

[SpidersWeb]

My first thought was an address latch starting to fail - but I'm certainly not skilled/knowledgeable enough to be sure or confirm.
Just wanted to mention it in case it sparks a thought in somebody else.

 

[jh1523]

Unfortunately the other system I have is an XT clone, and I only have the 5150 version (27C64) of the Landmark ROM. I think I would need the 27C128 or 256 version for the clone as the BIOS chip is a 256.

Like this one here? http://www.minuszerodegrees.net/supersoft_landmark/5155_5160/Supersoft Landmark ROM - IBM 5160.htm

I could burn a 27c256 and send it to you for the price of the postage. It may take a while for it to get from the US to Scotland though; the last package I received from the UK took almost 3 weeks.

 

[modem7]

Now, I get to this stage http://www.minuszerodegrees.net/5150_5160/MDC/no_beeps_2.htm and all voltages measure within limits except possibly the "power good", which is nearly 5V rather than "over 2.4V".

POWER GOOD is a signal, rather than voltage that powers something, and the signal is at TTL levels. Per [here], for the PSU to generate a compliant TTL 'high' level on the POWER GOOD line, the voltage needs to be anywhere between 2.4V and 5V. That is why IBM will have indicated 2.4V or above.

I'm afraid there's not much I can do to verify if the BIOS is faulty, but I do have two chips (in fact, I got one from you!) and both show the same behaviour.

I have lost track of what I have sent to people. I remember posting to you the Landmark/Supersoft ROM (in a 27256). I must have included a copy of the IBM 5150 BIOS ROM as well, in a 27256. That would have been the final version, 10/27/82 (no point in providing an earlier version).

You will be using that 27256 hosted IBM BIOS ROM via the same adapter that you are using for the 27256 hosted Landmark/Supersoft ROM.

So, you see:

Original IBM-supplied BIOS ROM: "single beep immediately at power on but no other activity"
27256 hosted IBM 5150 BIOS ROM in adapter: "single beep immediately at power on but no other activity"

What is odd about ""single beep immediately at power on" is that the POST in the IBM 5150 BIOS ROM (breakdown of POST at [here]) only produces a single beep at the end of the POST, something like 50 seconds after power on. So, that suggests that the beep you are hearing is not generated by the POST.

 

[Divad]

POWER GOOD is a signal, rather than voltage that powers something, and the signal is at TTL levels. Per [here], for the PSU to generate a compliant TTL 'high' level on the POWER GOOD line, the voltage needs to be anywhere between 2.4V and 5V. That is why IBM will have indicated 2.4V or above.


I have lost track of what I have sent to people. I remember posting to you the Landmark/Supersoft ROM (in a 27256). I must have included a copy of the IBM 5150 BIOS ROM as well, in a 27256. That would have been the final version, 10/27/82 (no point in providing an earlier version).

You will be using that 27256 hosted IBM BIOS ROM via the same adapter that you are using for the 27256 hosted Landmark/Supersoft ROM.

So, you see:

Original IBM-supplied BIOS ROM: "single beep immediately at power on but no other activity"
27256 hosted IBM 5150 BIOS ROM in adapter: "single beep immediately at power on but no other activity"

What is odd about ""single beep immediately at power on" is that the POST in the IBM 5150 BIOS ROM (breakdown of POST at [here]) only produces a single beep at the end of the POST, something like 50 seconds after power on. So, that suggests that the beep you are hearing is not generated by the POST.

Thanks for clearing up the PWR Good voltage.

I'm honestly not sure myself now what I received from you. What I have are two original IBM 5150 BIOS ROMs (latest revision), one that came with the motherboard and the other I thought came from you. Regardless, both give the same behaviour - single beep (sometimes slightly distorted) at power on, so seemingly pre-POST.

I also have a 2364 adapter from http://store.go4retro.com/adapters/ and a 27c64 with the Landmark ROM burned. This fails to boot at all.

And finally, I have a working 2364 24-pin EPROM with the Landmark ROM burned. This is the one I'm using and get the display shown above.

Anyway, the impression I'm getting here is that there is some fault with the board itself, possibly aggravated by a fault with the MDA card. It also seems unlikely that both IBM BIOS chips are faulty.

Given that, it might be worth getting another 2364 or 27c64 with a generic ERSO BIOS image and see if that generates any beep codes as ChuckG suggests. Would you expect an XT class BIOS to function in a 5150 board?

Thanks

David

 

[Trixter]

It appears you have some memory errors on the video card (not just what the test bios reports, but it appears you have some corruption in the second picture).

The errors are quite consistent (they are entire sections of text meant for another screen location), so I think the corruption is caused by something else. Faulty CPU? Flaky address lines? I don't think it's card memory errors because those always (in my experience) show up as random flipped bits in what is displayed.

I'd be curious to see if the screen corruption changes on every reboot, and/or if it changes (or is fixed!) by moving the card to another slot.

 

[modem7]

I'm honestly not sure myself now what I received from you. What I have are two original IBM 5150 BIOS ROMs (latest revision), one that came with the motherboard and the other I thought came from you.

That is quite possible. I do have some spare genuine IBM 5150 BIOS ROMs (of version 10/27/82). Maybe you requested an original.

Regardless, both give the same behaviour - single beep (sometimes slightly distorted) at power on,

I would have tested the IBM 5150 BIOS ROM that I sent to you. So, I am happy in ruling out the IBM 5150 BIOS ROM.

I am also presuming that you are not doing anything silly like inserting the ROMs in the wrong orientation (been there, done that, many times).

I also have a 2364 adapter from http://store.go4retro.com/adapters/ and a 27c64 with the Landmark ROM burned. This fails to boot at all.

* Did you use your EPROM burner to verify the contents against the image used to burn the EPROM ?

* Maybe the speed rating of your 27C64 is too slow for the IBM 5150. IBM simply indicate, "The ROM is ... and has an access time of 250-ns and a cycle time of 375-ns."

And finally, I have a working 2364 24-pin EPROM with the Landmark ROM burned. This is the one I'm using and get the display shown above.

It also seems unlikely that both IBM BIOS chips are faulty.

Agreed.

Anyway, the impression I'm getting here is that there is some fault with the board itself, possibly aggravated by a fault with the MDA card.

Remember, the Minimum Diagnostic Configuration test on my website explicitly indicates that no cards are to be inserted into the motherboard, for the very reason of ruling out interference.

But maybe you meant "caused" rather than "aggravated".

We know that your power supply is good. Because of that, if you follow (exactly) the first page of the Minimum Diagnostic Configuration test, and do not hear the expected three beeps, then you can be sure that there is something still wrong with your motherboard.

 

[Divad]

The errors are quite consistent (they are entire sections of text meant for another screen location), so I think the corruption is caused by something else. Faulty CPU? Flaky address lines? I don't think it's card memory errors because those always (in my experience) show up as random flipped bits in what is displayed.

I'd be curious to see if the screen corruption changes on every reboot, and/or if it changes (or is fixed!) by moving the card to another slot.

To the best of my recollection, the corruption stays exactly the same at every reboot, no matter what slot the card is in.

Unfortunately, I can't run the tests again just now. I tried removing all but bank 0 RAM to see if that made any difference to the board behaviour (it didn't). When I replaced all the RAM the Landmark test show 2-3 failed chips, so I will need to replace those despite earlier tests showing all RAM as passed

Once I get replacement 4116s I will run the tests again and get back to you.

 

[Divad]

I would have tested the IBM 5150 BIOS ROM that I sent to you. So, I am happy in ruling out the IBM 5150 BIOS ROM.

I am also presuming that you are not doing anything silly like inserting the ROMs in the wrong orientation (been there, done that, many times).

* Did you use your EPROM burner to verify the contents against the image used to burn the EPROM ?

Remember, the Minimum Diagnostic Configuration test on my website explicitly indicates that no cards are to be inserted into the motherboard, for the very reason of ruling out interference.

But maybe you meant "caused" rather than "aggravated".

We know that your power supply is good. Because of that, if you follow (exactly) the first page of the Minimum Diagnostic Configuration test, and do not hear the expected three beeps, then you can be sure that there is something still wrong with your motherboard.

I can confirm that all chips are being inserted correctly.

I do not have an EPROM burner of my own, so I had to get someone else to burn the images. The 27c64 and 2364 were burned by different people, I can't say for sure that the 27c64 was programmed correctly.

Minimum diagnostic configuration instructions were followed exactly, with no keyboard or cards connected and 5 & 6 on SW1 both off. Although for convenience I'm using a non-IBM ATX PSU with adapter, as the voltages check OK, I assume this will not cause any odd behaviour.

See above reply to Trixter, I have some potentially bad RAM that didn't show up until the chips were removed and reseated, I will need to replace these before I can do more testing.

I didn't realise that was your site, you've got a lot of good info on there!

Thanks

David

 

[Divad]

OK, diagnostic ROM showed around 2-3 failed RAM chips, I replaced these and am now getting the following:

16K Critical memory region PASS
Memory refresh FAILED
System memory to 10000 PASS
Slow refresh to 10000 PASS

Bank 0 & 1 have had all 4116 chips replaced with NOS, banks 2 & 3 I have tried swapping chips but just can't seem to get past the memory refresh error! I know the 16-64K 5150 board causes some odd behaviour with the Landmark ROM but I don't see anything referring to this failure on http://www.minuszerodegrees.net/supersoft_landmark/5150/16KB-64KB - Bugs.htm

Motherboard behaviour with the original BIOS remains the same.

Regards

David

 

[modem7]

The 'memory refresh' test is described on page 22 of the manual. As stated there, it is possible for this test to fail due to the RAM chips. Maybe the "NOS" RAM chips you acquired are unsuitable, or not 100%.