Faulty RAM board?

[Cristiano Paris]

Hi,

I bought a Compupro Godbout Econoram Xa (32KB) for my IMSAI 8080 (with 8080 IMSAI CPU board) and I tried to put it to test.

The board has two 16KB banks and I mapped them to +8KB (Bank A) and to +24KB (Bank B) since I have another 8KB RAM board that I have in my system.

I tested it through the front panel and it seems that Bank A has some "stuck bit" issue. Specifically:

1 - I tested three different addresses, two consecutive ones (same IC) and one far away. For the consecutive addresses, a few bits keep their state regardless of what I put in the locations, but the stuck-bit pattern is the same for both addresses. If I test the far-away address, the faulty pattern is different.

2 - To ensure it was not some miscommunication when interfacing the front-panel, I also tried to write a program that blinks the programmed output leds using a pattern read from a specific location and, indeed, when I target a location inside the faulty bank, the program doesn't work, shown a different pattern, with some bits on and others off. So, the behaviour is consistent between front-panel and CPU.

Before dismissing the board as faulty, I have two questions:

1 - The switches are setup as follows:

Switch 1 - All off, as I'm not using extended addressing. I also removed the octal comparator in U12, as suggested by the manua.

Switch 2 - 1 abd 3 off, all the others are on, i.e. both blocks are enabled, no write protect, MWRITE enabled, disabled PHANTOM line

Switch 3 - 2, 3 and 7 are on, all the others are off. This is the memory mapping switch.

Are they coorect? The manual is here:

https://retrocmp.de/northstar-hrz/econoram-xa_ocr.pdf

2 - In case you think Bank A is actually faulty, is this fixable somehow, like replacing the IC on the board or whatever?

Thanks a lot for any help you can throw at me.

Cheers,

C.

[Peter Sterrantino]

Hi Christiano, 

I wouldn't call the whole board faulty, but it looks like you may have some issues. Can you provide some more information on your statements below?

1 - I tested three different addresses, two consecutive ones (same IC) and one far away. For the consecutive addresses, a few bits keep their state regardless of what I put in the locations, but the stuck-bit pattern is the same for both addresses. If I test the far-away address, the faulty pattern is different.

The actual addresses would help. Your memory array is 64 4Kx1 memory chips. That's a lot of chips and a lot of power. The good news is you're seeing single bit failures, which would point to individual chips. The hardware that collects those bits and presents them to the data bus appears to be working for the other bank, so without  a deep dive into your schematic I would say you have a number of bad memory chips (or bad connections).

• I would check the output of all the 7805 regulators to make sure you have all the power you need. I can't tell from the schematic if these are all in parallel or if they power seperate sections, but it would be good to start there.
• Then check the +5 on the individual chip to be sure they are getting power. 
• When you narrow down to the chips based on the addresses, I would start swapping memory chips to see if the bit errors move with the chips (bad chips)
• And with that many connections and that many years a good reseating of all the chips (and possible cleaning of pins where needed) can eliminate a lot of problems. 

To make things simpler, I might recommend pulling your 8K board out of the system and resetting your Xa switches to occupy 0000h to 7FFFh to keep the addressing simple. Definitely check power and reseat chips first... That may clarify things. 

Pete

--
You received this message because you are subscribed to the Google Groups "Altair 8800" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Altair-8800...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/Altair-8800/CAApis8ydW7WYz-t0yaMRMBeXVSTcop16AiT_0Fvz-_Jfjpvf%2Bg%40mail.gmail.com.

[Cristiano Paris]

Dear Peter,

thanks, this is much appreciated. I'm not very good at electronics, but I know people who can help me. One question: would the layout of Bank A be so that each IC appears as a contiguous block of 512 bytes in the address space? This would help me diagnose where the problem is, as I can test each chip individually and rule out if it appears fault or not. If all the chips appear faulty, then one of the regulators could be the issue.

Thanks again.

BR

C.

--

GPG Key: 4096R/C17E53C6 2010-02-22
Fingerprint = 4575 4FB5 DC8E 7641 D3D8  8EBE DF59 B4E9 C17E 53C6
 

[Walt Perko]

Hi, 

Do you think it's just the RAM chips that might be faulty?  IF so, then you could swap chips around on the board and see where the stuck bits move, then replace the faulty chip(s).  

.

[Cristiano Paris]

Sure, that I will do. But instead of going blindly about it, I'd do it with knowing which IC I'm address with a specific address, that's why I'm asking wheter the 512 bytes of each IC are seen as contiguous in the address space (like 0x0000-0x01FF for the first IC).

Thanks!

C.

To view this discussion visit https://groups.google.com/d/msgid/Altair-8800/23cb882a-901c-49c7-873c-519c237f9a58n%40googlegroups.com.

[Walt Perko]

Hi,

 Moving the chips around will help you map out the chips on the board itself … at least partially until you find the faulty chip … but you could then mark that chip and continue moving it around on the board to complete mapping out all the RAM chips.  

 

.

[Patrick Linstruth]

Unless you know how the board works, randomly swapping ICs is usually never a good idea.

According to the schematic, it looks like each chip provides 1 bit for a 4K block of RAM. So when you read or write 1 byte of RAM, you are utilizing 8 chips to form a complete byte.

Which bits are stuck and and at which addresses?

I would definitely remove any other RAM boards, disable block B, and put the block A start address at 0000. That makes it easy to know which chips are providing which address.

I would then do the same for block B by disabling block A and putting the block B start address at 0000.

If you know which bit(s) are not working (assuming all enable logic and buffers are working), it will be the chip in the location for that bit (left to right) and 4K block (top to bottom).

Also, if you’re using a front panel, I think MWRITE should be disabled, since that signal should be provided by the front panel.

From https://www.retrocmp.de/northstar-hrz/ns-hrz_p01_restoration.htm

This board is organized as two blocks of 16K each: A + B. Each block can be set to start on any 4K boundary in the currently addressed 64K page. The corresponding bits are vertical organized: left is bit 0 and right is bit 7. Each row are 4K, beginning with A0 at the top (A0, A1, A2, A3, B0, B1, B2, B3).

To view this discussion visit https://groups.google.com/d/msgid/Altair-8800/915841c7-6abb-4c52-83b4-488120989ae1n%40googlegroups.com.

[Cristiano Paris]

Thanks, this is kind of clarifying why certain bits are stuck across addresses. I will do some testing, as you suggest, and report back.

Cheers,

C.

To view this discussion visit https://groups.google.com/d/msgid/Altair-8800/84FDB066-8DFC-4F50-AA95-D6700AC90722%40deltecent.com.