#2 5160 64-256k Motherboard memory question

[RDuckey]

I have built the dramduino tester and have gone through the memory on my 2nd 5160. I found 3 definitely bad chips. The (maybe) good ones were populated in banks 0 through 2 and the switches set to exclude bank 3. I have the Landmark diagnostic BIOS installed. I am aware that it is not the best of the two possible diagnostic ROMS, but it is here and being used. After startup i had to listen to the beeps for the keyboard and floppy controller errors, so I gave in and plugged them in. My 360k half height is not working, but I was pretty sure it was problematic when I put it in the chain.

I'm getting a Slow Refresh error at address 2098E Bit 2 and a Parity error at 28000. Everything else except the floppy read test passes. Am I correct in assuming that the "suspected" chip is in bank 2 and would be the 3rd chip? I going to further guess that the parity error is that the bank 2 parity chip doesn't agree with my suspected bad chip.

I know that I could just yank the chip and drop in one of the 6 remaining probably good ones, but I would like to understand why I'm doing something rather than just monkey fisting around with the chip lifter.

 

[modem7]

I'm getting a Slow Refresh error at address 2098E Bit 2 and a Parity error at 28000. Everything else except the floppy read test passes. Am I correct in assuming that the "suspected" chip is in bank 2 and would be the 3rd chip?

Address (pun intended) the slow refresh error at "address 2098E Bit 2", then see if the parity error goes.

Address 2098E hex = 133518 decimal = 130 KB. So, yes, on an IBM 64-256K motherboard, that corresponds to the third bank, bank 2.

As for "bit 2", see the diagram at [here].

 

[RDuckey]

Well that is progress. I swapped the "2" bit dram in the 3rd bank and the diagnostic liked the ram and everything else shy of the floppy read. I let it run for 5 passes.

I swapped in the normal bios chip and got a "dead" system. With the Landmark chip back in the board, I'm now seeing a failure at U28, which Landmark lists as System Bios. Everything, again shy of the floppy, passes testing.

 

[modem7]

I swapped in the normal bios chip and got a "dead" system.

Perhaps a faulty IBM BIOS chip in socket U18.

With the Landmark chip back in the board, I'm now seeing a failure at U28, which Landmark lists as System Bios.

The reference to U28 is in regard to an IBM 5150 (the diagnostic ROM not being able to determine if it is in a 5150 or 5160).
Per [here], in the IBM 5160, the 5150's "U28" equates to part of the U19 ROM.

 

[RDuckey]

Sorry for the delay in responding. I've decided to invest in some blank rom chips and a programmer. Unfortunately it will probably be a few weeks before delivery. Then I get to figure out how use the new equipment. If new roms don't fix the issue, I'll be back for more advice..

 

[modem7]

I've decided to invest in some blank rom chips and a programmer. Unfortunately it will probably be a few weeks before delivery. Then I get to figure out how use the new equipment.

Certainly in the long run, I think that you will find the purchase worthwhile.

For example: If one of your 5160 motherboards fails again, then depending on the symptoms, we may direct you to program some test code from [here] into an EPROM/EEPROM for socket U18, getting you feed back the results to us.

For example: Create a 'Ruuds Diagnostic ROM'.

If new roms don't fix the issue, I'll be back for more advice..

If that is unsuccessful, some next steps are:

- If in use is an MDA or CGA card, verify that switches 5 and 6 on motherboard switch block SW1 are set appropriately.

- Via the Supersoft/Landmark Diagnostic ROM (SLDR), verify that the aforementioned switches are being read correctly by software.

- Try Ruud's Diagnostic ROM (RDR), because it may detect a RAM addressing failure. SLDR does not look for RAM addressing failures. The POST in the IBM BIOS ROM will detect some, and when it does, simply halt, giving you a 'motherboard appears dead' symptom.

 

[RDuckey]

For example: Create a 'Ruuds Diagnostic ROM'.

This is one of the many reasons that I decided to add more toys to my collection.

- Via the Supersoft/Landmark Diagnostic ROM (SLDR), verify that the aforementioned switches are being read correctly by software.

I'll drag the machine back on the bench and try to figure out how to do this. I'm familiar with only the initial screen. I have done the switch testing on my first/other 5160 when it was still riddled with bad ram chips.

 

[modem7]

This is one of the many reasons that I decided to add more toys to my collection.

Some of the tests at [here] use a "parallel (LPT) port POST code reader" as the output device. A photo of that is at [here]. They are pretty cheap. If I do a search on eBay using lpt post, a few show up, at about the $9 Australian mark.

Via the Supersoft/Landmark Diagnostic ROM (SLDR), verify that the aforementioned switches are being read correctly by software.

I'll drag the machine back on the bench and try to figure out how to do this. I'm familiar with only the initial screen. I have done the switch testing on my first/other 5160 when it was still riddled with bad ram chips.

Look for the window titled 'XT SW1'.

 

[modem7]

This is one of the many reasons that I decided to add more toys to my collection.

Some of the tests at [here] use a "parallel (LPT) port POST code reader" as the output device. A photo of that is at [here]. They are pretty cheap. If I do a search on eBay using lpt post, a few show up, at about the $9 Australian mark.

And a logic probe would be handy. See [here].

 

[RDuckey]

Two more toys to add to the tool box are on their way. I ordered a logic probe that seems to meet the same specifications as yours.

The LPT port post reader was interesting to find on eBay. It kept trying to sell me automotive battery terminals, but I won in the end.

 

[RDuckey]

I have received the programmer, the logic probe and the W27E257(-12) chips. I have successfully installed the Ruud's Diag rom and it is running. I'm hoping the LPT port post reader shows up soon. Only banks 0, 1 & 2 are populated at this point. Hopefully the ram upgrade kit shows up soon, it has been 6 weeks at this point since I ordered it.

I do not have a floppy controller installed, so I understand that error. The only other error is "Check ROM at F4000". I noticed that after 9 passes, the F4000 error stops being added to the list.

Attached (hopefully) is the display. Sorry for the purple display, we've discussed that before.

 

[RDuckey]

If I am reading the Ruud's Breakdown correctly at step 103, the ROM at U19 may be a problem. I believe that was the assumption derived from the Landmark ROM as well. I'll try re-seating it.

Nope, it still failed on the first 2 passes along with the floppy. I'm also waiting for another terminating resistor from Digikey so that I can use the former 2nd floppy from machine #1 in this one. They will both have XT-IDE's and one floppy.

I note that in the link for the U18 tests listed above, that the Winbond W27E257's that I have are not suitable in the 64-256KB board. I guess that I won't have the opportunity to use them in the board until the ram upgrade kit from Monotech stops wandering the planet and arrives.

 

[modem7]

If I am reading the Ruud's Breakdown correctly at step 103, the ROM at U19 may be a problem. I believe that was the assumption derived from the Landmark ROM as well. I'll try re-seating it.

Nope, it still failed on the first 2 passes ...

Suggesting a faulty U19.
You now have an EPROM programmer and W27E257 EEPROM's, and so you can now create a replacement U19.
From [here], you can determine which BIOS revision you have.
And knowing the BIOS revision, at [here] is applicable content for the W27E257.

 

[modem7]

I see that you have switch 1 of switch block SW1 set to ON.
Per [here], on an IBM 5160, that switch is usually set to OFF.

 

[RDuckey]

Good catch. Yes, I had that switch set incorrectly. I had turned it on when replacing the original ROM didn't work and failed to put it back. I also played with 40/80 column mode to see if it made any difference. It doesn't.

Unfortunately, repositioning it doesn't seem to make any difference.

 

[modem7]

Good catch. Yes, I had that switch set incorrectly. I had turned it on when replacing the original ROM didn't work and failed to put it back. I also played with 40/80 column mode to see if it made any difference. It doesn't.

Unfortunately, repositioning it doesn't seem to make any difference.

Me pointing out the incorrect switch setting was me predicting a future post from you where you got the IBM BIOS up and running, but discovered that the POST just kept endlessly restarting.

As established in another thread, the magenta (purple like) display that you see is due to something in your 5153 monitor. As for your motherboard, the only problem that I think remains is a faulty U19 chip of the IBM BIOS ROM pair. The W27E257 definitely works on the 64-256KB motherboard, and so you now have the means to create a replacement U19.

 

[RDuckey]

Well, I suppose I'll take your advice and make a new U19. I missed post#13.

 

[RDuckey]

Done. Works great. Thanks for the in depth help and guidance, it is really appreciated.

 

[modem7]

Done. Works great.

Good to hear.

I updated your '1246735' entry in the failure list at [here]. A page refresh in your web browser may be required.