• Please review our updated Terms and Rules here

Help to ID multifunction card (RAM/RTC/Serial)

Chr$

Experienced Member
Joined
Aug 13, 2021
Messages
153
Location
Saxony, Germany
I just wondered if anyone had seen one of these before. It's a clone multifunction card with 384k RAM, RTC and serial/game port connectors. Solder side is marked 'WTF' (which was also my line of thought when I saw the battery damage!) and there is also 'WTF 821115'.

filedata/fetch?filedataid=63015

It has a lot of bodges due to Ni-Cd leakage, but it does seem to work. I get the 3 extra things shown on the initial boot screen when I plug it in.

I left the switches and jumpers as I found them and would like to know what they do. My m/b (here https://www.vcfed.org/forum/forum/ge...clones/1221340) shows 256k without the card - I think that's a limitation/bug of the m/b. It also shows 256k with the card installed in its current configuration.
 

Attachments

  • photo63015.jpg
    photo63015.jpg
    161 KB · Views: 1
  • photo63016.jpg
    photo63016.jpg
    103.3 KB · Views: 1
I believe that this is your card.

The 5150 will only check memory up to the limit specified by the motherboard switches.

It took me less time to identify your card from the TH99 archives than it took for you to write the message asking about it.

Great, thanks Chuck. I did of course try to find details, but was obviously searching for the wrong thing in the wrong place.
 
You have an XT-class motherboard. The norm (repeat: norm) for that is that the two RAM switches in the motherboard's switch bank do two things:
1. Disable/enable motherboard RAM banks; and
2. Inform the power-on self test (POST) as to how much motherboard (repeat: motherboard) RAM there is.

For RAM (conventional type) beyond that, the POST will simply test bytes beyond the final address of motherboard RAM.

You have 256 KB of motherboard RAM. I see from the card photo and the information that Chuck pointed to, that the card is set to start its RAM at address 256 KB, and enable all of its RAM banks. That is expected. Based on that, I expect that the POST would count up to 640 KB of RAM.

Note that if the BIOS behaves the same as the BIOS in the IBM XT (IBM 5160), a design flaw causes a problem if there is a faulty RAM chip/s, e.g. maybe there is one or more faulty RAM chips in your card's bank 0. You can read about the flaw at [here].

Maybe you just need to re-seat the RAM chips in your card's bank 0. You could try swapping RAM chips between bank 0 and the other banks on the card.

Try the procedure at [here] seeing if a read from address 256K shows one or more faulty bits.
 
Thanks, I'll look into that. I did also try the ERSO v2.42 BIOS to see if that counted differently, but at this stage I don't know if that potentially has the same bug.

The 2 switches on the m/b seem to allow you to chose between having 64k or 256k (according to image from manual, below). The silkscreen near the RAM does mention that it will accept HM50256 chips, but I don't know if that means in just one bank or in multiple.

I can certainly test the RAM chips though, simply by pulling them out of card and popping them in one of the m/b banks one at a time. I did make a start on that, but it's late here now.
 
Four 8-data-chip banks of HM50256 (256 kbit) would provide 1 MB of RAM.

The thread at [here] discusses another XT-class motherboard, one that came in two 'flavours':
- 640 KB of motherboard RAM
- 1 MB of motherboard RAM
Different types of RAM chips, however there may be other differences, e.g. wiring, support chips, BIOS.
 
I believe that this is your card.

The 5150 will only check memory up to the limit specified by the motherboard switches.

It took me less time to identify your card from the TH99 archives than it took for you to write the message asking about it.

Fountain technologies? Shouldnt the company be mountain technologies?

Pretty sure i have this card in my box of pc cards.
 
Well, you've got me wondering. Time to pull out the 1988 PC product guide.

Mountain Computer (Scotts Valley, CA) made a bunch of hard cards as well as tape drives (FileSafe series). No multifunction. So your Mountain Technologies is a mystery. OTOH, my directory does show Fountain Technologies, in Somerset, NJ has a bunch of cards (disk controllers, CGA card, memory, etc.). They show the MFT with 384KB memory, 1 serial, 1 parallel, clock. Price $149.
 
TH99 doesn't mention being able to use 256Kb chips, though I suppose it might be possible to use only the first 64Kb of one--I don't know.

Sorry for the confusion, I meant the silkscreen on the motherboard. It mentions HM50256.
 
Again, I don't see any jumpers that would enable that. Fountain did produce a later MFT3000 board for (they claim XT and AT compatibility) 3MB of DRAM. Maybe the label is just an erroneous carryover.
 
Because there was some previous battery damage I buzzed all the RAM for continuity and found no issues. I suppose one or more sockets could be iffy, but I buzzed gently from the tops of the chips. With the computer on I also checked a few RAM chips to make sure they had +5v, and they did. It's a wonder that the other features of the card work, so I should be grateful for that. Oh, I also swapped the first bank on the card for the 1st bank on the m/b and it also made no difference.
 
Oh, I also swapped the first bank on the card for the 1st bank on the m/b and it also made no difference.
The norm on RAM cards is that, of the card's multiple RAM banks, the lowest numbered bank is the one that appears the lowest in motherboard memory space.
But as been pointed out to me previously, there exists some cards where the bank numbering is in reverse.
So try swapping out bank 6.

And if you extend that, there is the possibility that the lowest appearing bank in motherboard memory space may be neither 1 nor 6, but one of the banks in between.

I think it would be good for you to use the technique at [here] for address 256K. If you see, say, one bit in error, then that would suggest a bad RAM chip somewhere on the card.

Another possibility that comes to mind: 'Dirty' switches on the card. Try moving them between ON and OFF a few times.
 
That's interesting, thanks. The RAM chips in the last numbered bank on the card were actually the ones that were most affected by the battery leakage (so much so that a few have completely lost their etched designations!), originally close to the first numbered bank. I put them in the last bank deliberately so that all the 'good' ones came first, so that's a good idea to try.

And when I've set up my other machine with the only other 360k drive I have, I will put some useful stuff like Checkit etc onto a few hopefully reliable 360k disks. I'll see how I get on with debug tomorrow.
Thanks very much for all the useful suggestions.
 
And not only can the bank-to-address mapping be 'all over the place', the arrangement of bits within the bank can also be 'all over the place' - see post #8 at [here].
 
Have tried various RAM chip swaps. No joy yet. I also fitted 256k x1 bit chips to bank 0 of the m/b and without the RAM card it still added up to 256k in total somehow. I will continue to experiment. I've commandeered an XT-IDE from a 286 so should be able to get some files across more easily for Checkit etc. I've also tried out the RTC. The program Clock.com detects the RTC chip and can send it the time, but it doesn't save it after turning off.... then I realised that the original battery was 3.6v and my coin cell is only putting out 2.4v after the added diode so I'll need to get some more juice to it!
 
Just mentioning that my Megaboard has 640k directly on the board itself. But your board clearly differs a bit. I don't know if any jumpers/jumper wires/logic chips would have to be changed to support that on yours.

If this multifunction board came with this machine originally, then I would expect it to work, but if someone just threw it in there then there could be some configuration issue. I would absolutely poke around in the memory with DEBUG and see if it is actually reading/writing anything at all to any of the bank locations.
 
Back
Top