• Please review our updated Terms and Rules here

Fixing Five Festive 5150 Boards over the Holidays (BOARD #3 Thread)

VeryVon

VeryVon

Miner 2049'er
Staff member
Joined
Oct 21, 2021
Messages
326
Location
Land of the Cheesesteak
I recently received five 5150 boards from another collector in various states of repair. They've obviously put a lot of effort into the restorations as 3 out of the 5 boards have bank 0 socketed. To my knowledge they weren't using a diagnostic ROM. I'm really not sure if I'll be able to repair any of these boards!! but I do have a diagnostic rom, and a working 5150 of my own so here we go...

Here are the initial results of my tests using the minuszerodegrees Minimum Diagnostic Configuration:

1702425918544.png

Board Overview (each board has a label affixed in the lower-left corner as positioned in these pictures)
1702426007759.png
1702426041594.png

BOARD #3

I'm starting with board #3 because it's the most promising, and I'm trying to gain some confidence of course haha. Results of diagnostics with all ram & parity chips removed is confusing, because it shows the even bits as good without any chips in them! I can add a parity chip, but once I add an additional ram chip to any of the bits, the board is DOA. I've been staring at schematics, I though I was on to something with "odd & even" on U94 S280, but that just appears to be flipping parity checks...

Any ideas why the "Bad bits" would look that way, even with the board unpopulated?

1702426345500.png
 
VeryVon said:
F*ing 5150 Boards
Click to expand...
Your spell checker changed 'IBM' to 'F*ing' !
How weird. :unsure:

VeryVon said:
I'm starting with board #3 ...
Click to expand...
And for the other boards, create a unique thread for each one. Why? Because multiple parallel threads running in this thread will confuse the hell out of me (and I, and possibly others, will make mistakes).

VeryVon said:
Any ideas why the "Bad bits" would look that way, even with the board unpopulated?
Click to expand...
Yes. What you are seeing is like what the POST in the IBM BIOS would report (if it reported an 'early RAM' failure on-screen). It is because the RAM test uses various bytes against each address, and it aborts on the first failure. So the test byte would have been AA or 55.

You are using an early version of Rudd's Diagnostic ROM. Upgrade to the latest one ([here]. The newer versions do not fail an address until all of the test bytes (FF,00,55,AA,01) have been used. In that way, you get a 'complete picture'.
 
I put "ruuds_diagnostic_rom_v4.3_8kb.bin" on a freshly erased 2564 using the GQ-4X "2564(TEST)" profile as I did with the older image. Verifies fine, but when I put it in U33 now no beeps or video. Tried another 2564, same result. No other changes in the setup, does this new version have different ram requirements? hmmm.
 
VeryVon said:
I put "ruuds_diagnostic_rom_v4.3_8kb.bin" on a freshly erased 2564 using the GQ-4X "2564(TEST)" profile as I did with the older image. Verifies fine, but when I put it in U33 now no beeps or video. Tried another 2564, same result. No other changes in the setup, does this new version have different ram requirements? hmmm.
Click to expand...
For version 4.3, the RAM requirements are in the 'Requirements' section at [here]. No motherboard RAM is required. If an MDA card or CGA card is present, Ruud's Diagnostic ROM will use some of the RAM in that.

I grabbed a spare 2564JL-35 EPROM, and put ruuds_diagnostic_rom_v4.3_8kb.bin into it. I then put that EPROM into the U33 socket of a 5150 motherboard of type 64KB-256KB. As expected, Ruud's Diagnostic ROM displayed on my CGA monitor.
 
Thanks for confirming that, legs look good on the chip. I'm going to burn the old Ruud's version to the 2nd 2564 I have and report back.
 
Are you going to keep all five of these boards, once repaired, or will you be looking for good homes for them?

I also like the idea of a single motherboard project to a single forum thread.

(Is F Star a LAN topology I'm unaware of?) :unsure:
 
I'll be giving all the boards back to the original collector regardless when the project is completed. At that time I can let him know about your interest though!
 
I put the older Ruud's version (2022-12-12) on another chip, and confirmed the board comes up same as screenshotted above with the odd bit's X and the even O. When I swap out for Ruud's 4.3, do not get any beeps or video. Using an MDA card. Back to the drawing board.

Took a quick logic probe down the CPU with Ruud's 4.3 in there. Status bits 010 = Write I/O port.

1702495163262.png

Update: I'm going to try a different video card
 
Last edited:
VeryVon said:
I put the older Ruud's version (2022-12-12) on another chip, and confirmed the board comes up same as screenshotted above with the odd bit's X and the even O. When I swap out for Ruud's 4.3, do not get any beeps or video. Using an MDA card. Back to the drawing board.
Click to expand...
The one thing that hasn't been verified yet is the actual 2564 that you are using for Ruud's Diagnostic ROM (RDR) version 4.3
I don't think that just because an EPROM programmer successfully reads back the content, that one should declare the EPROM to be 100% functional.

I see that in your thread of February, at [here], you had some kind of issue with 2564's - see posts #5 to #14. You didn't reveal what the problem was.

One thing you could try is to erase the 'proven' 2564, the one presently containing RDR version 2022-12-12, then put version 4.3 into that.
 
Your February thread shows that you have an oscilloscope and a cheap logic analyser.

Do you have a logic probe?

Do you have a parallel/LPT reader device, like the one shown at [here]? If not, you should get one. It is very cheap, and is used as the output device for some of the ROM based tests at [here].
 
Great memory about the issues with the 2564's. Indeed I'm currently running 4.3 on the "proven" 2564, and the old version on the semi-flakey one. Still unsure what the issue is with the 2nd one. It appears to write & function successfully but fails verify.

I do have a logic probe.

Is there another source for those parallel/LPT reader devices? The most inexpensive 4-digit one I see on ebay is this one: https://www.ebay.com/itm/302662276561
 
VeryVon said:
Is there another source for those parallel/LPT reader devices? The most inexpensive 4-digit one I see on ebay is this one: https://www.ebay.com/itm/302662276561
Click to expand...
Here are a couple:

www.ebay.com.au

Laptop Mini Pci Lpt 4 Digital Tester Post Code Card ap | eBay

Mini PCI&(LPT+USB) 2 in 1. Mini PCI&LPT port 4 bit diagnostic card for Desk pc and notebook. The value of four-bit-code PC Analyzer's 'initiative code' is specified and it is '0000' or 'FFFF'. 'initiative code' is clear at a glance.
www.ebay.com.au
www.ebay.com.au

Laptop MINI PCI LPT 4 digital Tester POST CODE Card | eBay

Mini PCI&(LPT+USB) 2 in 1. Mini PCI&LPT port 4 bit diagnostic card for Desk pc and notebook. The value of four-bit-code PC Analyzer's 'initiative code' is specified and it is '0000' or 'FFFF'. 'initiative code' is clear at a glance.
www.ebay.com.au
 
Ah ok, nice! Searched one of those up on eBay us where I'm located and found one for 8.60USD shipped from china shipping "speedpak standard"

Got my working 5150 up on the bench, dropped the 4.3 eprom in there and no dice, so the eprom is not written correctly. going to try and erase / write again...
 
VeryVon said:
Got my working 5150 up on the bench, dropped the 4.3 eprom in there and no dice, so the eprom is not written correctly. going to try and erase / write again...
Click to expand...
A 'flaky' batch of 2564's ?
If you are not powering your GQ-4X using an AC adapter, perhaps try that.

These days, I use Winbond W27E257 EEPROM's together with a suitable adapter.
 
Yep, definitely powering the burner with a 9V adapter, all the lights are on, and I managed to write and verify the 4.3 image on both of these 2564's, but neither will beep or boot in my working 5150. Swapped back to the bios quickly and I can see it boot up, so my bench setup is working.

Yes definitely 'flaky' ! and I think I'm done with these 2564's. I just ordered one of the 2364 Adapter's mentioned on your site as RETRO INNOVATIONS is reasonable in price & shipping.
 
modem7 said:
I don't think that just because an EPROM programmer successfully reads back the content, that one should declare the EPROM to be 100% functional.
Click to expand...
Yes, after some recent "bad ROM experiences" with refurbished UVEPROMs from the far east, I agree with this remark 100%.
 
Last edited:
modem7 said:
I don't think that just because an EPROM programmer successfully reads back the content, that one should declare the EPROM to be 100% functional.
Click to expand...
Agreed, Over the years i've had a few Eproms that could be written / read / erased and written to again perfectly fine but not work when put in circuit, I also own a GQ-4x but the older model i bought new back in 2010 and i have to say it's not the best for reliably burning older Eproms.
 
Malc said:
Agreed, Over the years i've had a few Eproms that could be written / read / erased and written to again perfectly fine but not work when put in circuit, I also own a GQ-4x but the older model i bought new back in 2010 and i have to say it's not the best for reliably burning older Eproms.
Click to expand...
The thing is that there is nothing wrong with the GQ-4x programmer itself.

It is just that, in reality circuits task the eprom differently, with high read speeds that sometimes data bits can't be extracted from the eprom without errors.

Typically these are "fast read" applications, when say video character data is read out of the eprom.

I think that all you are seeing is that the GQ-4x does a relatively slow data read of the eprom file.

So it will report the eprom file is fine, but if it is a dodgy eprom, that eprom won't work in some specific application.

It is in fact the fault of the Eprom itself, not the GQ-4x, which obviously was not designed to detect marginal and poor quality eproms.

But, the GQ- 4x works perfectly for good quality "normal " eproms for programming and reading.

The trouble is that there are a lot of fake, refurbished, re-labelled eproms flooding the market and this is where all the troubles reside, not in the programmers.
 
You must log in or register to reply here.
Back
Top