Early 5150 troubleshooting

[pkhoury]

I'm trying to troubleshoot my early 5150 (serial 0100485) that I picked up at Computer Reset back in 2019. Several years ago, I got a copy of the Supersoft diagnostics with adapter from someone in our local retro computer club. One of the issues with the machine was bad tantalums. I took the liberty of replacing all those tonight with replacements sourced from Mouser, obviously noting polarity. Tried to boot up again, and nothing. Put in the Supersoft ROM, and I started getting results. I know modem7/minuszerodegrees says the results aren't that accurate on a 16-64KB motherboard. In particular, it reports errors for two of the four ROM BASIC chips, and lots of errors in the first bank. Would the next step to try Ruud's diags? I was thinking of bringing this with me to VCF Southwest next month.

On that note - does anyone have a good EPROM writer they can recommend? I was using a BK Electronics writer, that I've since misplaced, but it only worked under 2000 and XP (I got it super cheap, though). Also, is it okay, for the most part, if the problem is in fact in Bank 0, to desolder the chips and put in sockets? I picked up some NOS 4116 on ebay about a year or so ago. Thankfully, I have a pretty decent Hakko desoldering gun that's served me well. As you can also note, by how early this board is, the CPU isn't socketed, and a lot of the chips have 1981 date codes on them.

 

[640KB]

Wow! You've even got the original 04/24/81 ROM set too. Did it have the original set of adapter cards and floppy drives? Pics of those?

So grateful Computer Reset was able to adopt out much of the collection so we can preserve and enjoy what is likely a very small number of surviving models like that!

 

[pkhoury]

Wow! You've even got the original 04/24/81 ROM set too. Did it have the original set of adapter cards and floppy drives? Pics of those?

So grateful Computer Reset was able to adopt out much of the collection so we can preserve and enjoy what is likely a very small number of surviving models like that!

Sadly, it didn't have any I/O cards in it. In fact, I don't even remember if it had a top cover or not, or one that was original, anyways. The goal is to restore this to semi-original. I say semi, because I'll want to still expand the memory to 544KB, and definitely opt for a genuine Hercules card instead of the crappy IBM mono card.

 

[modem7]

Put in the Supersoft ROM, and I started getting results. I know modem7/minuszerodegrees says the results aren't that accurate on a 16-64KB motherboard.

Well, per [here], RAM related results. And per the first bug listed there, on a 16-64KB motherboard, do not believe that a failure of the "16K CRITICAL MEMORY REGION" test must mean that a chip/s in RAM bank 0 has failed.

Also, is it okay, for the most part, if the problem is in fact in Bank 0, to desolder the chips and put in sockets?

... and lots of errors in the first bank. Would the next step to try Ruud's diags?

Yes, because, as stated above, the RAM problem may not be caused by a RAM chip/s in bank 0. It would be a pity that if in installing IC sockets, you damaged the PCB, only to discover later that the bank 0 chips are good.

... it reports errors for two of the four ROM BASIC chips ...

Be prepared to discover that those two ROM chips are indeed faulty.
When the RAM problem is sorted out, anticipate the symptom shown in the table at the bottom of [here].

 

[pkhoury]

Well, per [here], RAM related results. And per the first bug listed there, on a 16-64KB motherboard, do not believe that a failure of the "16K CRITICAL MEMORY REGION" test must mean that a chip/s in RAM bank 0 has failed.

Should I move forward to using the Ruud diagnostic ROM instead just to verify?

Be prepared to discover that those two ROM chips are indeed faulty.
When the RAM problem is sorted out, anticipate the symptom shown in the table at the bottom of [here].

But is the ROM required for the system to go into POST?

 

[modem7]

Should I move forward to using the Ruud diagnostic ROM instead just to verify?

Yes.

If there is a bank 0 problem, you will probably see what is linked to in the yellow table row of the web page at [here]. I used "probably" because of the possibility that note 1 on that web page discusses. Nonetheless, there will be some form of indication of RAM failure below the 16K address.

But is the ROM required for the system to go into POST?

The BASIC ROM's are not optional for your 04/24/81 dated motherboard BIOS. The POST will start, then fail at the test of the BASIC ROM's. Per the Cassette BASIC related web page that I pointed to in post #4, you will hear {one long beep followed by one short beep. About 15 seconds later, a single short beep.} then nothing further happens.

If you upgrade the motherboard BIOS to the 10/27/82 dated one, the symptoms change. Expected it that the POST display 'F600 ROM' and 'FC00 ROM' and then continue on. Obviously, with Cassette BASIC compromised, it is unlikely that you would be able to boot to Cassette BASIC, but a boot to DOS should be possible.

Have you tried re-seating the Cassette BASIC ROM's ?

 

[pkhoury]

Yes.

If there is a bank 0 problem, you will probably see what is linked to in the yellow table row of the web page at [here]. I used "probably" because of the possibility that note 1 on that web page discusses. Nonetheless, there will be some form of indication of RAM failure below the 16K address.


The BASIC ROM's are not optional for your 04/24/81 dated motherboard BIOS. The POST will start, then fail at the test of the BASIC ROM's. Per the Cassette BASIC related web page that I pointed to in post #4, you will hear {one long beep followed by one short beep. About 15 seconds later, a single short beep.} then nothing further happens.

If you upgrade the motherboard BIOS to the 10/27/82 dated one, the symptoms change. Expected it that the POST display 'F600 ROM' and 'FC00 ROM' and then continue on. Obviously, with Cassette BASIC compromised, it is unlikely that you would be able to boot to Cassette BASIC, but a boot to DOS should be possible.

Have you tried re-seating the Cassette BASIC ROM's ?

I have not, but I can try that. If I put in the 10-27-82 BIOS from another board, do I have to change the cassette ROM, or can I leave it as-is for the time being, for testing purposes? I don't get any beeps from the machine, except with the Supersoft diag ROM present. When it tries to test cassette BASIC, I can hear/see the relay clicking during tests for each of the ROMs that fail.

 

[modem7]

If I put in the 10-27-82 BIOS from another board, do I have to change the cassette ROM, or can I leave it as-is for the time being, for testing purposes?

There is no requirement to upgrade Cassette BASIC from version C1.00 to C1.10

At [here] is shown a couple of examples of the BIOS upgrade kit that IBM supplied to get the motherboard BIOS to revision 10/27/82. No Cassette BASIC ROM's of version C1.10 were provided in that kit - the 10/27/82 BIOS does not require C1.10

I don't get any beeps from the machine, except with the Supersoft diag ROM present.

No beeps are expected from the POST of the IBM BIOS when one of the early tests fail. The diagram at [here] should explain things.

When it tries to test cassette BASIC, I can hear/see the relay clicking during tests for each of the ROMs that fail.

Expected behaviour.
See the first point within the 'Other Notes' section of [here].

 

[pkhoury]

There is no requirement to upgrade Cassette BASIC from version C1.00 to C1.10

At [here] is shown a couple of examples of the BIOS upgrade kit that IBM supplied to get the motherboard BIOS to revision 10/27/82. No Cassette BASIC ROM's of version C1.10 were provided in that kit - the 10/27/82 BIOS does not require C1.10


No beeps are expected from the POST of the IBM BIOS when one of the early tests fail. The diagram at [here] should explain things.

Thank you for the diagram link. I somehow missed that while perusing your site. Now it makes more sense.

Expected behaviour.
See the first point within the 'Other Notes' section of [here].

That makes sense as well.
I misplaced my parallel port EPROM writer, but think I know where my UV eraser is. So with the adapter I'm using for my Supersoft diags chip, I should be able to use if I burn my own copy of Rudd's diags, or a replacement BIOS? Also, would it be advisable to run test0599.zip? Lastly, whenever I figure out the problem, is there anything wrong with using an adapter and programming the original BIOS revision, if it turns out the ROM chip is bad? Call me crazy, but I want to keep most of the machine original, even though I only got it with the bottom chassis and motherboard originally.

On a segue, I was planning to outfit it with dual 5.25" drives from a 5161, controller, two memory expansion boards (one probably an AST MegaPlus or Six Pack), and a Hercules board.

 

[modem7]

Also, would it be advisable to run test0599.zip?

The SuperSoft Diagnostic ROM is indicating a RAM problem (of some sort). Had no problem at all been indicated by that diagnostic, then that would be when TEST5099 would be used. TEST5099 detects a particular RAM subsystem failure mode that neither the SuperSoft Diagnostic ROM nor Ruud's Diagnostic ROM detects.

We are expecting that Ruud's Diagnostic ROM will give you an indication of a RAM problem within the first 16 KB.

( Hopefully later on, Ruud's Diagnostic ROM will be modified, so that TEST5099 is no longer required. )

Lastly, whenever I figure out the problem, is there anything wrong with using an adapter and programming the original BIOS revision, if it turns out the ROM chip is bad? Call me crazy, but I want to keep most of the machine original, even though I only got it with the bottom chassis and motherboard originally.

Well, I must be crazy as well. But, lucky for me, I have spare 10/27/82 IBM BIOS ROM chips. Had I not, then I would probably by okay with using a 2564 EPROM replacement, so as not to require an ugly looking adapter.

 

[pkhoury]

The SuperSoft Diagnostic ROM is indicating a RAM problem (of some sort). Had no problem at all been indicated by that diagnostic, then that would be when TEST5099 would be used. TEST5099 detects a particular RAM subsystem failure mode that neither the SuperSoft Diagnostic ROM nor Ruud's Diagnostic ROM detects.

We are expecting that Ruud's Diagnostic ROM will give you an indication of a RAM problem within the first 16 KB.

( Hopefully later on, Ruud's Diagnostic ROM will be modified, so that TEST5099 is no longer required. )


Well, I must be crazy as well. But, lucky for me, I have spare 10/27/82 IBM BIOS ROM chips. Had I not, then I would probably by okay with using a 2564 EPROM replacement, so as not to require an ugly looking adapter.

I only didn't want to install the latest BIOS ROM because I wanted the machine to be as close to original as possible, despite having replaced all the tantalum caps. I already have other machines with the latest ROM, including a 5160 connected to a 5161 with 300MB SCSI hard drive as the boot device. Now you can call my a hypocrite though, because I'd be using an LCD monitor with MCE2VGA, because I never liked CRTs (and it's a little too much to carry). Plus, I don't know if I have a viable 5151, anyways.

I'll see if I can pick up an EPROM programmer, and maybe I'll be lucky and have some spare 27C64's here.

 

[modem7]

I only didn't want to install the latest BIOS ROM because I wanted the machine to be as close to original as possible, despite having replaced all the tantalum caps.

One of my 5150 motherboards has the 10/19/81 revision BIOS ROM. I have no spares for that BIOS revision. If that ROM fails, I will most likely use 10/19/81 in the replacement EPROM, rather than the final revision.

I'll see if I can pick up an EPROM programmer, and maybe I'll be lucky and have some spare 27C64's here.

I get the impression that you read my '2564' as 2764. Different animals.

 

[pkhoury]

One of my 5150 motherboards has the 10/19/81 revision BIOS ROM. I have no spares for that BIOS revision. If that ROM fails, I will most likely use 10/19/81 in the replacement EPROM, rather than the final revision.


I get the impression that you read my '2564' as 2764. Different animals.

I probably did. Are there any EPROMs I can program with a modern programmer, that are readily available (ebay) that I could use sans adapter? Same would go for cassette BASIC ROMs as well.

##EDIT##
Been doing some more reading/research. I see that the MCM68766 or equivalent requires a high voltage programmer, which I could purchase. Any thoughts on this EEPROM solution, since I could theoretically use the same chip for both Ruud and the original BIOS revision?
https://www.ebay.com/itm/155133430035

I also found this for a replacement ROM, and I asked the seller some more detailed questions:
https://www.ebay.com/itm/155133430035

Are there any significant differences between the first and second ROM revision? I don't remember seeing anything on your site, unless I missed something (which seems to have been the case often enough). I do have some spare 16-64KB boards. One may or may not have an earlier revision ROM that I could try, though I'd probably want to try the board in its current state and see if that yields any progress before swapping ROM chips.

 

[modem7]

Are there any significant differences between the first and second ROM revision?

According to the first post at [here], the 10/19/81 revision "Fixed a couple of minor display bugs."
I probably sourced that from one of the editions of the book, Upgrading & Repairing PC's.

Are there any EPROMs I can program with a modern programmer, that are readily available (ebay) that I could use sans adapter?

What I am aware of that can be used is shown at [here]. I do not keep up-to-date with what 'modern' programmers support.

As you will have read, you do not need to create replacements yourself. Some of the entities in the list at [here], can supply programmed MCM68766.

Same would go for cassette BASIC ROMs as well.

Replacement EPROM types suitable for the U33 BIOS ROM socket equally apply to the Cassette BASIC ROM sockets as well.

 

[pkhoury]

According to the first post at [here], the 10/19/81 revision "Fixed a couple of minor display bugs."
I probably sourced that from one of the editions of the book, Upgrading & Repairing PC's.


What I am aware of that can be used is shown at [here]. I do not keep up-to-date with what 'modern' programmers support.

As you will have read, you do not need to create replacements yourself. Some of the entities in the list at [here], can supply programmed MCM68766.


Replacement EPROM types suitable for the U33 BIOS ROM socket equally apply to the Cassette BASIC ROM sockets as well.

Thank you very much. Supposedly, the one seller who has an EEPROM replacement says it will work with the programmer I referenced, so I bought three chips and a programmer from another seller, so I should be able to replace the two bad ROMs and be able to program Ruuds as well. And just in case, reprogram the original BIOS revision if that ROM is also bad, as well. I'm very much looking forward to seeing this board revived once again!

 

[640KB]

Thank you very much. Supposedly, the one seller who has an EEPROM replacement says it will work with the programmer I referenced, so I bought three chips and a programmer from another seller

What types of EEPROMs were those? It was always my understanding that there were no pin compatible EEPROMs to the MCM68766/2564 and you'd always need a 28 pin adapter.

 

[pkhoury]

Pin compatible EEPROMs that are already on adapters. The fellow I bought some from stated that they're drop in replacements for the MCM68766 chips, and then I purchased a programmer from another source (an assembled project, but from someone else's open source project).

 

[modem7]

Pin compatible EEPROMs that are already on adapters.

I have just now (by accident) spotted your edit to post #13. Obviously, you are referring to the eBay item pointed to by the edit.

 

[pkhoury]

I have just now (by accident) spotted your edit to post #13. Obviously, you are referring to the eBay item pointed to by the edit.

Correct. I ended up ordering 3 EEPROMs on adapters that I'm told are pin compatible with the MCM78766, and a high voltage programmer. The seller of the EEPROMs said he'd flash cassette BASIC on two EEPROMs and ruud's diag on another (I figure I can reflash the ROM later when it gets here). Obviously not original, but I like the idea of reflashing an EEPROM over and over. I also overlooked the fact that two of the four BASIC ROMs are Motorola, and diags reported them good. Surprise, surprise I suppose. Do the Motorola's have bit rot problems, or is the problem more prevalent with Mostek?

 

[modem7]

Do the Motorola's have bit rot problems, or is the problem more prevalent with Mostek?

I cannot answer that, but maybe someone else can. I have not seen any kind of 'ROM failure statistics by make-model' list.

Even if the scope is reduced to ROM's fitted to the 5150 motherboard, the list at [here] suggests a low failure rate. I doubt that a statistically significant sample size of U33/BASIC ROM failures exists.