The Sinclair QL Forum

KINGCHEF wrote:As the disc i/face is now not working with 2 of the QLs and only with Minerva I thought I would look at that first.

This says the board is border line on the signals and might be loading the QL outputs or not giving enough on the QL inputs so putting it on a straight through backplane would make things worse. Perhaps the mod was done because the board always had the problem and might be pointing to the problem area.

KINGCHEF wrote:It may be that the board was sold with 256/512 options or that a previous owner had added memory later.

Would the board work one of the outputs so you know it works if it had no memory chips that would reduce its power consumption.

KINGCHEF wrote:On the subject of modifying the backplane or the expansion board there is a paradox.

Not really for the above reason. Keep the backplane straight through and simple.

KINGCHEF wrote:Checking the backplane connections were done from pin to socket using a dummy connector to get to the socket pins.

We assume end to end you checked for shorts by going around the three adjacent pins on every pin.

KINGCHEF wrote:I'm more interested in the +12v and -12v readings which are much lower than that. How/where is it regulated and what is it used to power?

Watch out for the AC output on the power supply I didn't realise one was AC that gives a low reading on DC measurements on the meter.

KINGCHEF wrote:Does it actually power anything in the peripheral board?

Check the medic circuit. Some old ram used funny voltages.

Long shot on this with one additional thought.

Did you say that you are using a QEPIII and a Medic disc interface? There was a limitation in older QL ROMs with QDOS 1.03 or earlier (certainly AH and JM, think it was fixed in JS, MG or Minerva) which meant you could only have one add-on card with ROM on it. This meant that "passive" backplanes could only use one add-on card. So, you may be able to use a disk interface without memory and a memory expansion, since memory expansions don't have onboard ROMs.

Some "active" backplanes such as the CST Q+4 card had onboard code which corrected the startup issue of not being able to recognise more than one.

More on this in Andy Pennelll's QDOS companion, page 138, including some example code.

Hi.

Thanks Dilwyn. The disk i/face and Qep3 worked for years in the original setup which is JS ROM so I guess the ROM issue was fixed by then. But any suggestions are really welcome.

Right. I'd missed that the 12v was AC. I'll recheck that.
I did side to side shorts but not the adjacent. Should have thought of that.
That's what comes of using equipment designed for 3 score years and ten, nearly 71 years later.

Minerva is now showing memory fault on the disk i/face at 400BC/400B0/4007C/40070/ which , I think is the first of the expanded memory. It does not move on to a second address and ignores the memory on the interface.


Dave

Update

I removed all the memory ICs from the Medic board and cleaned the sockets with isoclene. I very lightly sanded the IC pins and reassembled at what I thought was 128K at a time. I tried the card after each 128K but Minerva didn't see anything or make any comment although it did see the i/face. It finally found the memory after all were replaced and then said there was 144K total.
A short while later it found all of it.
Tried using the backplane but QL didn't see the card at all.
Decided to recheck all connections and all seemed well.
At this point I had only been checking the integrity of the backplane so as one last try I checked what I could from the connections on the inside of the QL at the expansion socket.
A couple of connections were not being made up at the pin 30s end.
Had a thought and propped up the backplane so it was straining a bit upwards against the QL connector and all was there at startup!
Loosened off and it failed again.
Added the QEP3 and the QL sees all of it.
Have yet to re-assemble the keyboard and drives but I'm a bit more optimistic now.
It seems that over time the socket on the backplane has become slack and at times some pins fail to make the connection. This would explain the 'works one minute not the next' situation.

It might not all be over yet!

Thanks
Dave

KINGCHEF wrote:It might not all be over yet!

Sounds like you have found the problem. The QL connector is in mid air but I assumed extension cards would rest on the outside edge of the case or have some kind of support. Looks like you need a block of wood under the attached board, a bit of 10000 BC tech.

Guess wot!

Used a block of wood

Would be nice to know the original keywords of the Medic I/F and the differences to TTs QFLP driver.

Hi,

Tel me where to look and I'll do my best. I might have a copy of the ROM to look at. Would that help?
Dave

Well, I do not know, where in memory your Medic ROM sits, perhaps, someone here can give a hint, so you can SBYTES it. I also do not know if it is a 8k, 16k or 32k Eprom. I also depends, if your Medic I/F is the first one in the expansion bus or if there are additional I/F with their own software.

Hi,
I have now read the MEDIC rom and it is a 2764.
Obvious words in text are
Medic Disk Upgrade v1.18
FILES STILL OPEN
READ/WRITE FAILED
FLP_USE
FLP_SEC
FLP_TRACK
FLP_START
PAR_USE
RAM_USE
MDV0 , SER, PAR, OEMS, HDRD,RTD, RAM0,QL5A,

Possibly one or two others.
Is there something I should be looking for?
Is this any help? I don't have TTsQFLP driver.

Dave