Commodore 8050 dual disk drive

[SiriusHardware]

Start by using the scope to see if U10 pin 11 changes from high (its 'idle' state) to low, even momentarily, when you issue a disc command, presumably there is one to list the contents of a disc?

[TonyDuell]

Thnking about it, there's one thing that puzzles me.

These drives have no disk-inserted switch that is read by the digital board. No index sensor either. So if you try to read a disk I would expect it to start the spindle motor and keep it running long enough to get up to speed and to attempt to read the disk. And yet Colin says that the motor-on signal changes too fast to see with a multimeter.

What happens if you try to read from the working drive but with no disk in it? How long does the drive select LED stay on for? How long does the motor run for? Now try doing the same with the faulty drive, again no disk in it. How long does its select LED stay on for?

[ortek_service]

Originally Posted by SiriusHardware

It's really a solved problem Owen, the ROMs were read using an Arduino which, among other things, produced the 16-bit checksums of the two ROMs under investigation.

It was then possible to load the matching .bin files for these ROMs from online sources into a hex editor and observe that the .bin files had the same checksums as the ROM readouts produced by the Arduino.

The same Arduino sketch can also, alternatively, output the code as Intel Hex code which can be cut from the Arduino serial monitor, pasted into a plain text editor and saved as a file which can then be loaded by the majority of EPROM programmers.

I agree it is odd that the simple trick of pulling the active-high CS pin high with a resistor did not allow the STAG programmer to read out the ROMs, but maybe it is so clever that it first deactivates the CS signal to the device in the socket and checks to make sure that its outputs are tristate under those conditions.

Yes, I'd seen in this case that these were verified by matching their checksums. But I thought it would still be handy to be able to have the readout of the IC's in a standard format, as for some reason it hadn't been possible to get the Stag one to work.
So might be useful for the future to have the Arduino code do this, in case you want to readout an IC to backup, that there doesn't seem to be an existing good copy of.

[SiriusHardware]

So might be useful for the future to have the Arduino code do this

It already does. (See post #52, especially the part about the role played by the Arduino 'A0' input).

[ortek_service]

Yes, reading through the posts, it seemed it went from 3 to 4 flashes, with nothing changed. But just replacing two 2114's, has now resolved this, with no other digital errors reported.
And so it seems you've been quite lucky with this one, compared to your PET's, without a large number of IC failures.
And the 6530 RRIOT also seems to be OK?

Maybe the 6532 RIOT's you'd ordered (for this?) might still be needed for the printer. Although, now the digital part of this is running, then this could now be used to test the 6532's from that, if this has better diagnostic tests / feedback of fault.

Originally Posted by ScottishColin

This device has been giving different flashes at different times, but you;re right now; having replaced 2 of the 2114s, there's no flashes and therefore the digital PCB thinks all is good.

See post 107 for a summary re the right hand drive (Drive 0) not working however.

Colin.

[ortek_service]

Originally Posted by SiriusHardware

It already does. (See post #52, especially the part about the role played by the Arduino 'A0' input).

Thanks for info - I see it now does have the ability to do this, and so could have posted the dumped-contents of these, to compare contents / checksum from within HxD etc. rather than a pasted-from-window text file (But maybe that facility wasn't in the sketch, when it was originally first being used)

[SiriusHardware]

The 'Intel-Out' feature was already there, but I think we saw that the ROM checksums exactly matched those of the online files when the ROMs were read out in 'Human-Readable' mode and we thought 'we'll take that'. The Intel-Out mode of the Arduino sketch certainly would be good for making file copies of firmware which is not already widely available online, but that wasn't the case here, image files of the ROMs are already available on at least two different websites.

[SiriusHardware]

Getting back to the drive problem - are these drives hardware configurable to drive numbers other than the ones they are set to when new? I would assume so, in case you want to hang two (or more) disc drive units off the PET, for whatever reason. Is it possible that the non-working drive has just been assigned some other drive number, not the drive number it would usually be expected to have?

[TonyDuell]

The IEEE-488 address of the complete drive unit can be changed, eithr by software or by cutting tracks on the digital PCB.

By default the drive unit is device 8. If you have more than 1, you can power one of them up, send appropriate commands to it to make it device 9, then power the second one up (which will be device 8). More practically you can cut some tracks on the board to set the bottom 3 bits of the device address. This is shown in the service manual.

But this won't affect the numbering of drives 0 and 1 in a particular unit. You can't reassign the IEEE-488 address of one of the drives only

[SiriusHardware]

Fair enough.

Colin, when you are able could you please try the observations suggested in para 2 of Tony's post #122 and also have a look at the 'Motor On' signal input with your scope, as per post #121?

[TonyDuell]

I have just noticed that the motor-on inputs on the analogue board (e.g. pin 11 of U10) are driven by 7406 _open collector_ buffers on the digital board (see the very bottom of page 29 of the .pdf service manual).

As a result it's safe to pull them low with a bit of wire to test the motor control circuit on the analogue board. So :

Put all the diodes on the analogue board back in place. Connect everything up and power the drive unit up. Connect one end of a bit of wire to logic ground, touch the other end on pin 13 of U10 on the analogue board (take great care not to touch pin 14, which is of course a 5V supply pin). The 'good drive' motor should spin. Now stop touching pin 13 of U10 and touch pin 11 instead. Does the 'bad drive' motor run?

Also, when you tried to read a disk in the 'bad drive' what error did you get? I can't remember how to read the error codes but can find out if you don't know.

[ScottishColin]

So a successful LOAD"$1",8 will then show the directory with a LIST

If I try to do this on drive 0 (LOAD "$0",8), the central LED flickers between Red and Green and after 5 seconds-ish, I get a ?FILE NOT FOUND error. I can also hear a motor trying to be started but nothing spins. The central LED then goes solid Red.

a PRINT DS$ command gives me

74, DRIVE NOT READY, 00, 00

I can then successfully after this error go back to DRIVE 1 and work fine.

There's a full list of error messages on page 50 of this PDF:

http://primrosebank.net/computers/pe..._2031_text.pdf



Colin.

Originally Posted by TonyDuell

Thnking about it, there's one thing that puzzles me.

These drives have no disk-inserted switch that is read by the digital board. No index sensor either. So if you try to read a disk I would expect it to start the spindle motor and keep it running long enough to get up to speed and to attempt to read the disk. And yet Colin says that the motor-on signal changes too fast to see with a multimeter.

What happens if you try to read from the working drive but with no disk in it? How long does the drive select LED stay on for? How long does the motor run for? Now try doing the same with the faulty drive, again no disk in it. How long does its select LED stay on for?

[ScottishColin]

Drive 1 spins when pin 13 is connected, but drive 0 does not after connecting to pin 11.

A thought. Would it be worth me swapping over the two drives with each other - would that eliminate anything?

Colin.

Originally Posted by TonyDuell

I have just noticed that the motor-on inputs on the analogue board (e.g. pin 11 of U10) are driven by 7406 _open collector_ buffers on the digital board (see the very bottom of page 29 of the .pdf service manual).

As a result it's safe to pull them low with a bit of wire to test the motor control circuit on the analogue board. So :

Put all the diodes on the analogue board back in place. Connect everything up and power the drive unit up. Connect one end of a bit of wire to logic ground, touch the other end on pin 13 of U10 on the analogue board (take great care not to touch pin 14, which is of course a 5V supply pin). The 'good drive' motor should spin. Now stop touching pin 13 of U10 and touch pin 11 instead. Does the 'bad drive' motor run?

Also, when you tried to read a disk in the 'bad drive' what error did you get? I can't remember how to read the error codes but can find out if you don't know.

[SiriusHardware]

And this was with diode CR22 reinstated?

If so, it looks like you have a problem in the digital part of the drive 0 section on the analogue board.

Find a reliable way to keep U10 pin 11 connected to 0V, then repeat the voltage measurements Tony asked for back in #117.

Would it be worth me swapping over the two drives with each other - would that eliminate anything?

Not just now, as your tests appear to be narrowing things down to the digital part of the analogue board.

[TonyDuell]

I can't see the point in swapping the drives over. The only part of the drive mechanism that matters is the spindle motor and you've proved that will run on a 9V battery. You've also seen it will run with the motor control circuit on the analogue board when CR22 is lifted.

If the tachogenerator in the motor had failed then it would run too fast, but it would run.

It would appear the fault is one of 2 ICs on the analogue board, U10 or U11. Both are cheap and easy to get. The voltage tests I suggested will tell us which one has failed.

[ScottishColin]

I'll get to it - been a bit distracted taely; apologies.

Can you just confirm which pages of which documents I should be looking at for the schematics?

Thanks.

Colin.

[SiriusHardware]

The version of the manual we are referring to at the moment is the full one, which Tony posted a link to in #3. The current area of interest, the 'analogue' motor control PCB, is on manual page 11 / PDF page 13 of that manual.

We are asking you to force a 'low' on U10 pin 11 by connecting that circuit node to 0V, to simulate a 'motor on' signal from the digital board, and then measure the voltages on the points Tony mentioned in #117. To recap, those points were:

"pins 10 and 11 of U10 (74LS14) and pins 8,9,10 of U11(74LS32)". Since you will be tying pin 11 of U10 low for the purpose of this test, you need not include it this time.

Although you did already measure these points once, that was when the 'motor on' input was in the 'idle' state. We now want to see what happens to the voltages on those IC pins when the 'motor on' signal pin is held in the 'active' (low) state.

[ScottishColin]

OK - I connected U10/11 to U11/7.

This time, I get the following (the same voltages when it is idle as when I am trying a command for the failing drive):

U10/10 - 4.4V
U11/8 - 0.4V
U11/9 - 0.15V
U11/10 - 4.4V

Colin.

[SiriusHardware]

OK - I connected U10/11 to U11/7.

So you used U11's GND pin as your 0V to connect U10/11 to, that's OK.

U10/10 - 4.4V [Correct, as the input to this inverter gate is tied low]
U11/9 - 0.15V [Low, correct]
U11/10 - 4.4V [Correct, as it is on the same node as U10/10]
U11/8 - 0.4V [Incorrect].

For that last measurement, as Tony explained earlier the U11 10/9/8 gate is an OR gate, so if either U11/9 OR U11/10 is high then U11/8 should also be high - but it is not.

Power off and measure the resistance between U11 pin 8 and 0V, red on pin 8, black to 0V, what resistance do you see?

[ScottishColin]

3.75MΩ

Colin.

Originally Posted by SiriusHardware

So you used U11's GND pin as your 0V to connect U10/11 to, that's OK.

U10/10 - 4.4V [Correct, as the input to this inverter gate is tied low]
U11/9 - 0.15V [Low, correct]
U11/10 - 4.4V [Correct, as it is on the same node as U10/10]
U11/8 - 0.4V [Incorrect].

For that last measurement, as Tony explained earlier the U11 10/9/8 gate is an OR gate, so if either U11/9 OR U11/10 is high then U11/8 should also be high - but it is not.

Power off and measure the resistance between U11 pin 8 and 0V, red on pin 8, black to 0V, what resistance do you see?