The Sinclair QL Forum

Hello, my name is Ronaldo, and I'm from Brazil. I've got a Sinclair QL from a friend, sold to him as non-working. I don't have a power supply for it, so I managed to connect an ATX power supply to it. The board is missing the power connector, suggesting someone has already been working on it. As I try to turn it on, this is what I get:

https://youtu.be/UuVmEFT3oXM

I have never used a Sinclair QL, so I don't know where to start looking for the problem. I have a multimeter and an oscilloscope, in case it helps. Any suggestions?

Thanks

As I told you, I don't know how the QL works, but I've noticed the screen is stable as I keep the reset button pressed, suggesting the video generator circuitry is OK (the processor stops and the image isn't updated anymore). I suspect there may be something wrong with the processor, RAM or ROM, and the moving patterns I see moving on screen are rubbish being executed by the processor.

Hi there! Welcome to the forum!

Traditionally, white screen suggests faulty memory. It could also mean a bad 8301. Luckily, all of those are fairly easy to get your hands on.

The first thing you need to do is to get a copy of Minerva. This is a replacement OS for the QL that does more advanced memory checking. In the event of bad memory, it can often show you the location and type of memory problem.

If you have the QL power supply, I can send you a Minerva ROM, tested 8301 and the QL power connector (and a couple of other useful bits like a couple of RAM chips) for $10 plus the cost of actual shipping (probably $14 for shipping) to you in Brazil, if you're interested.

You only need to supply the -12v for using serial - so you don't need to connect that if you're not using it. +12v is used by serial and the MC1377 that provides RGB output.

Useful info:

When the QL starts up, one of its first actions is to copy the contents of the ROM into the rest of memory, then read it back as a crude memory test. This means when things start normally you should see a white/red/green/black dot pattern fill the screen, remain for a few seconds, then disappear and be replaced with a prompt to choose F1 or F2.

The memory is organised as two banks of 64K, each IC providing 1 bit of the byte. IC1-8 is bank 0 and IC9-16 is bank 1. You might be able to probe the RAM signals and see if they're correct. The DRAM's refresh is provided by the 8301, which also reads the RAM to create the video image - so as a video image is being produced you know the 8301 is at least 90% working

Hope this gets you started.

Oi Ronaldo! What city in Brazil are you?
If you use Facebook, you can join our QL-Brasil group. We are few, but maybe we can help you locally:
https://www.facebook.com/groups/257954401019500/

BTW, YOU´RE MEMBER 500 OF THIS FORUM, CONGRATULATIONS!

Dave wrote:Hi there! Welcome to the forum!

Thanks!

Dave wrote:The first thing you need to do is to get a copy of Minerva. This is a replacement OS for the QL that does more advanced memory checking. In the event of bad memory, it can often show you the location and type of memory problem.

Does the memory checking work with no memory installed? In case a full chip is bad, at the lower portion of the memory, no RAM would be available.

Dave wrote:If you have the QL power supply, I can send you a Minerva ROM, tested 8301 and the QL power connector (and a couple of other useful bits like a couple of RAM chips) for $10 plus the cost of actual shipping (probably $14 for shipping) to you in Brazil, if you're interested.

That'd be great! Though I don't have an original power supply, I think I managed to make the proper connections to get +12, -12, +5 and ground from an ATX power supply. As I only have an old microdrive and no cartridges to test it I'm not worried about providing 9V to the system, yet.

Dave wrote:When the QL starts up, one of its first actions is to copy the contents of the ROM into the rest of memory, then read it back as a crude memory test. This means when things start normally you should see a white/red/green/black dot pattern fill the screen, remain for a few seconds, then disappear and be replaced with a prompt to choose F1 or F2.

I get nothing but random repeating patterns on screen, as if the program had crashed very early into execution and the processor was continuously filling up the memory with data.

According to what you said I would be almost sure the problem is one or more faulty RAM chips, but what puzzles me is that a short reset seems to get only a black screen. Could that be a symptom of another problem?

Dave wrote:The memory is organised as two banks of 64K, each IC providing 1 bit of the byte. IC1-8 is bank 0 and IC9-16 is bank 1. You might be able to probe the RAM signals and see if they're correct. The DRAM's refresh is provided by the 8301, which also reads the RAM to create the video image - so as a video image is being produced you know the 8301 is at least 90% working

I tried probing the memory chips in search of an obvious problem, but the fact they share the same data lanes makes it somewhat pointless, as I'm getting both the IC output and the output shared by the other parts connected to the same line. I couldn't find anything obviously wrong, though I wasn't very thorough probing all pins.

Szk wrote:Oi Ronaldo! What city in Brazil are you?
If you use Facebook, you can join our QL-Brasil group. We are few, but maybe we can help you locally:
https://www.facebook.com/groups/257954401019500/

BTW, YOU´RE MEMBER 500 OF THIS FORUM, CONGRATULATIONS!

Thank you, I've sent a request. I'm from Rio Grande do Sul.

realpeterjack wrote:Does the memory checking work with no memory installed? In case a full chip is bad, at the lower portion of the memory, no RAM would be available.

It would try and fail immediately. So, here's a brief outline.

The 68008 is the CPU.
The 8301 is the Master Chip. It does video generation from main RAM. When it does this, it uses the three ICs IC19, 20, 21 to isolate the CPU and ROM from the RAM. It also produces the 7.5MHz CLKCPU.
The 8302 is a peripheral controller. It operates the microdrives, and TX of both serial ports. It also produces the reset signal.
The 8049 is an Intelligent Peripheral Controller. It operates the RX half of the serial ports, keyboard matrix and joysticks. It is also the baud clock generator from the 11 MHz crystal.

realpeterjack wrote:That'd be great! Though I don't have an original power supply, I think I managed to make the proper connections to get +12, -12, +5 and ground from an ATX power supply. As I only have an old microdrive and no cartridges to test it I'm not worried about providing 9V to the system, yet.

I can through in a couple of blank microdrives for you, but that might push the package over 6 oz, which would increase shipping costs. You might do better making friends with Szk, who is also in Brazil. As I seem to recall, you don't have a single standard for power there and it could be 120V or 220V, almost randomly?

realpeterjack wrote:I get nothing but random repeating patterns on screen, as if the program had crashed very early into execution and the processor was continuously filling up the memory with data.

The QL can do that without the CPU even installed, because the 8301 runs independently and generates the video signal AND the CLKCPU 7.5MHz clock. The video image shows either RGB all equal 1, or 0, and never any other pattern where R or G or B is missing. The pattern isn't really random as it creates vertical lines.

realpeterjack wrote:According to what you said I would be almost sure the problem is one or more faulty RAM chips, but what puzzles me is that a short reset seems to get only a black screen. Could that be a symptom of another problem?

/RESET halts the operation of the CPU and 8301. They will remain frozen until you release the button - then /RESET goes high and they will kick off their resets. They're in a bad place due to the problem, so how they reset doesn't have to be consistent.

realpeterjack wrote:I tried probing the memory chips in search of an obvious problem, but the fact they share the same data lanes makes it somewhat pointless, as I'm getting both the IC output and the output shared by the other parts connected to the same line. I couldn't find anything obviously wrong, though I wasn't very thorough probing all pins.

The snippet of the schematic I included above should give you some insight. You can download your exact schematic here: https://z00m.speccy.cz/ql/pics/sinclair ... iagram.png

Warm welcome to the forum for our 500th member!!!

Dave wrote:It would try and fail immediately.

But it would run, at least, wouldn't it? That would let me know the video generator and processor are OK, at least.

I've just started reading other forum posts and found something about Minerva. Do I have to make any changes to the PCB to get it running?

Dave wrote:I can through in a couple of blank microdrives for you, but that might push the package over 6 oz, which would increase shipping costs. You might do better making friends with Szk, who is also in Brazil. As I seem to recall, you don't have a single standard for power there and it could be 120V or 220V, almost randomly?

Thank you for the offer. Right now my biggest concern is having no spare parts to check it for problems. I want to have it booting to the prompt, at least, then I may pursue fixing the drives (it had only one installed). I think trying to randomly replace RAM chips would be a waste of time and possibly good ICs, so I basically have two options:

I could ask the friend who gave me this QL to test the socketed ICs on his working QL;

I could get a Minerva rom and perform more testing on my own.

So, it safe to swap the possibly bad parts into the working QL?

Regarding the power standards, it would be something like this: essentially the states' capital cities get 127v, everywhere else gets 220v. Both are 60 Hz. I happen to live in a 220v area.

Dave wrote:The QL can do that without the CPU even installed, because the 8301 runs independently and generates the video signal AND the CLKCPU 7.5MHz clock. The video image shows either RGB all equal 1, or 0, and never any other pattern where R or G or B is missing. The pattern isn't really random as it creates vertical lines.

I tried removing the CPU and I got a black screen, only. By pseudo random I mean it changes, consisting mostly of vertical lines, but it seems to repeat every second or so. And the more I let the QL running like this the less I get white, as I start getting more and more semi-random black pixels.

Dave wrote:/RESET halts the operation of the CPU and 8301. They will remain frozen until you release the button - then /RESET goes high and they will kick off their resets. They're in a bad place due to the problem, so how they reset doesn't have to be consistent.

So I assume this inconsistent behaviour is to be expected from a faulty system, basically?

Dave wrote:The snippet of the schematic I included above should give you some insight. You can download your exact schematic here: https://z00m.speccy.cz/ql/pics/sinclair ... iagram.png

I think I'll have to keep from investigating the problem any further until I get either the Minerva ROM replacement or authorisation to test my chips inside a working QL. Even if I replace all memory chips that would not ensure having it in full working order...

vanpeebles wrote:Warm welcome to the forum for our 500th member!!!

Thank you!