Can someone help me date my 5150?

[modem7]

I tried using the MDA/parallel card instead of the CGA card and composite, the MDA doesn't trip the power supply protection (5 and 12 show up as expected) but it won't even beep on RDR with it installed.

Very very recently, it was discovered that in the last 8 versions of RDR, a bug existed that made RDR incompatible with the 'IBM Monochrome Display and Printer Adapter.'
That bug has been fixed, and the new RDR version is 4.8

can the 9 pin on the CGA card run the 5151 in b/w mode?

No. There are other MDA-to-CGA differences, like the scanning frequencies.

 

[MacGyverr]

Well, it's got all new caps anyway now.
I guess I'll have to pull the RDR to test it, thanks for keeping RDR updated BTW.

Knowing that, I pulled the RDR and put the BIOS back in, and got this from the MDA and monitor.
When I put the CGA back in I got an onboard RAM error (likely from the bad P8255A-5 telling lies about how much is configured).
RAM on the video card?

 

[modem7]

Knowing that, I pulled the RDR and put the BIOS back in, and got this from the MDA and monitor.

RAM on the video card?

Related are posts #7 and #8 in the thread at [here]. There, the change of a space (20h) to an exclamation mark (21h) is a change of bit 0. When I simulated RAM bit 0 on my IBM MDA being stuck high, basically, the whole screen (where spaces) changed to exclamation marks. Only some of the OP's screen are exclamation marks, and I put that down to a partially faulty bit 0 RAM chip.

Looking at your screen, it is interesting that in it too, not all of it (the spaces) is affected. Affected are 1024 characters. The change of a space (20h) to the ' character (60h) is bit 6 being high. Weird though is the "7777 777", which will be the 201 error. Why all 7's? Let me see what happens if I simulate RAM bit 6 on my IBM MDA card being stuck high.

 

[modem7]

Further regarding the unexpected characters on the IBM MDA.
Looking at the circuit diagram more:

U12 and U13: Character - bits 7 to 4
U14 and U15: Character - bits 3 to 0
U8 and U9: Attribute - bits 7 to 4
U10 and U11: Attribute - bits 3 to 0

Each chip does almost half of the screen. So, if we look at the U12/U13 pair, one of those does character bits 7 to 4 for the top 1024 character positions on the screen. The other does character bits 7 to 4 for the bottom 976 character positions on the screen. 25 lines of 80 characters = 2000 characters = 1024 + 976

I would have to examine the circuit diagram more to establish which does the top and which does the bottom. Although it is probably documented somewhere.

At [here] is the result of me tying pin 12 (connected to bit 6) of U13 to +5V. The reason why all spaces on the screen, rather than approximately half, have changed to a ' is because pin 12 of U12 and U13 are connected, i.e. I am affecting both U12 and U13.
At the top-left is a bit-6-stuck-high version of "1054 201", followed underneath by a bit-6-stuck-high version of " 301", followed underneath by a bit-6-stuck-high version of "601". E.g. 31h ("1") to 71h ("q").

Looking back at your screenshot, I do not have an explanation for the "7777 777" (your corrupted 201 error). Why all the same character? Also, I think the ' character is different to the one that I am thinking it is.

 

[MacGyverr]

Was going to buy a few of these TMS2114-20NL 200ns replacements and try to fix the RAM, but found a replacement (untested so I may be repairing one of them anyway) for way too cheap on ebay (less than I could buy the chips and get them shipped). I just don't trust my ability to desolder that many chips and not screw up the board. I'm betting it was my violent desoldering gun and me replacing the tantalum above U13 that caused the problem in the first place.

It was a bundle of 3 cards, I couldn't pass up having a spare floppy controller and the 64 spare RAM chips on the expansion card alone.

 

[modem7]

I would have to examine the circuit diagram more to establish which does the top and which does the bottom. Although it is probably documented somewhere.

I did that. I am very confident, based on the circuit, and an experiment that I did, that U13 and U15 are for the top 'half' of the screen, and U12 and U14 are for the bottom 'half'.

( And I confirmed John Elliott's information. The card decodes more than 4 KB of motherboard address space: 32 KB from B0000 to B7FFF. )

 

[MacGyverr]

I ordered a replacement desoldering gun (a Hakko FR301-03/P clone that doesn't use a bouncing solenoid), if you think those TMS2114-20NL 200ns replacements will work I'm will to give it a try. Should I be replacing U7 (74LS139) U16, U17, U18 (74LS157s) as well as U13, U15? I can't really read the schematic I can only understand that they're connected.
Do you have a preferred vendor that can do small chip orders, I can easily find vendors for the 2114s I just don't need 650 of them, I found these Jameco ones, but not sure if they are 100% compatible..

 

[modem7]

Do you have a preferred vendor that can do small chip orders, ...

No. Others here might.

if you think those TMS2114-20NL 200ns replacements will work I'm will to give it a try.

Those are expected to work.
FYI: On a second IBM MDA that I have, the RAM chips are AM9114EPC.

I'm betting it was my violent desoldering gun and me replacing the tantalum above U13 that caused the problem in the first place.

Very suspicious. If you have not already, you should be investigating the possibility that your tantalum-above-U13 replacement activity did damage to the PCB, e.g. severed a trace. For example, are all of the chips in the area getting +5V on their Vcc pin?

Should I be replacing U7 (74LS139) U16, U17, U18 (74LS157s) as well as U13, U15?

Looking at the circuitry, I can see that BA0 (buffered A0) is used for the odd/even address selection (even=character, odd=attribute). I can see that when CCLK is low, the 6845 addresses the RAM. When CCLK is high, the address from the ISA bus addresses the RAM. Etc.
I cannot think of a cause for the "7777 777" - all the same character. An addressing issue perhaps. After verifying that U13 (and others in the area) is getting +5V on its Vcc pin, I think that you should replace U13, then see what MDA related symptom/s remain.

 

[MacGyverr]

So I got the replacement 8255 in the mail a D8255AC-2, the original was a P8255A-5 just want to confirm it is a proper replacement before I start trying to put it in (the data sheet seems to say it's the same just faster).
And it turned on 4 or 5 times correctly on the old chip that is still installed with the RAM set to 64K and the last few times it throws this 10D4 error, which by my math (and your cheat sheet) means either the 3rd or non-existent 4th bank, chips 2, 4, 6 and 7? Or is it like I thought and 10D4 is just after 64K (between 64 and 80K) and therefore just the failed 8255 chip telling the BIOS the RAM is larger than I set it to?

If you say it's a good replacement I'll try and swap it out.

 

[modem7]

So I got the replacement 8255 in the mail a D8255AC-2, the original was a P8255A-5 just want to confirm it is a proper replacement before I start trying to put it in (the data sheet seems to say it's the same just faster).

There should not be a problem using that chip. In my opinion, use of an IC socket is prudent.

Or is it like I thought and 10D4 is just after 64K (between 64 and 80K) and therefore just the failed 8255 chip telling the BIOS the RAM is larger than I set it to?

The "10" is pointing to the 4K block at address 64K, RAM that does not exist in your test configuration.

Yes, this is sure to be due to the combination of:
- The nature of the failure of your 8255 (i.e. incorrect switch settings reported); and
- The BIOS bugs described at [here].

 

[MacGyverr]

So after about 2 hours of me working to remove and replace that failed 8255, it finally works correctly.
It correctly shows all RAM cards and chips and the dip switches do what they're supposed to do.
Still having issues with now both my monochrome cards. (The ebay one doesn't work either, surprise)
I did an ok job I think, it does work, but I was worried I was never going to get the 40 pin IC out of there.
In the end I had to prop it up vertically and use an iron on one side and the desoldering gun on the other to get the sockets clear enough to nudge the chip out.

I'll try and run some diagnostics on my old card with the voltages you mentioned.
I can tell you that the error is consistent, it will display the same every time. The 2 pictures show 2 different power cycles and the longer on is where I typed a-z and A-Z and 0-9 at the command prompt.
The full screen nightmare is the newer ebay card, it's worse and is just random, I'll try replacing the 3 legged tantalums and see if it fixes it.
Thank you again for all your help.

 

[modem7]

So after about 2 hours of me working to remove and replace that failed 8255, it finally works correctly.
It correctly shows all RAM cards and chips and the dip switches do what they're supposed to do.

Good.

After verifying that U13 (and others in the area) is getting +5V on its Vcc pin, ...

I'll try and run some diagnostics on my old card with the voltages you mentioned.

Suggesting to me that you did the +5V verification that I suggested.

I'll try and run some diagnostics on my old card with the voltages you mentioned.
I can tell you that the error is consistent, it will display the same every time.

And let's see what remains after U13 replacement. There are other chips that affect only the first 1024 characters.

The full screen nightmare is the newer ebay card, it's worse and is just random, I'll try replacing the 3 legged tantalums and see if it fixes it.

- And maybe a good clean of the ISA edge connector.
- A thorough visual inspection.