Commodore 8050 dual disk drive

[TonyDuell]

According to page 13 (and 14 for the stepper) of that .pdf service manual, the motors in drive 1 run off +12VA, those indrive 0 run off +12B.

On page 12 of the manual, +12B also provides +12F (F = 'filtered', I guess) which is the supply for the read amplifier.

So it's possible the spindle motor in drive 0 is putting enough noise onto that supply line to make reading unreliable.

Do the tests that Siriushardware suggests. If you see a lot of noise on the supply line then maybe the motor is wearing out, Maybe a capacitor fault though, C!6,C26, C28 are possible faults here.

[ScottishColin]

I've spent quite some time on this and I'm wondering if there's something wrong with the scope or software.

If I leave the division at 5V, it happily reads 11.9V on the scope. If I move it to 2V/div, it gives me a nice flat line but only reads 9.2V.

So I've spent some time chasing my tail wondering where the lost voltages went, but having flicked it to 5V this morning, they display happily.

If I use the drive then, the line doesn't fluctuate. The drive 0 reading is a tiny bit more messy that the Drive 1 reading which stays nice and flat, but drive 0 still works ok.

And then I moved to changing the scope to AC not DC and the drop-down box is greyed out and I can't see a way either in the software or the manual to change it to AC....

So all I have right now is a slightly messy 11.9V line on the scope which doesn't fluctuate when the drive is in use and a drive that works 90% of the time....

Colin.

Originally Posted by SiriusHardware

It is interesting that the drive 0 voltage is on average a little bit lower than the drive 1 voltage but the voltage for the '12V' is not usually too critical, not like the 5V, which has to be 5V +/- only a small amount. It could be that drive 0 is running a tiny bit slower than drive 1, due to the marginally lower supply voltage.

What we are more interested in for now is how 'steady' the 12V supply for drive 0 is. Ideally, when viewed on a scope it should be a steady flat line at 12.0V no matter whether the drive is idle or active, because these are regulated supply rails, the implication being that they shouldn't vary too much even as the load varies.

Start with the scope on DC input mode as usual, put the '0V' trace right down on the bottom line and set the scope to 2V / Div. If you then attach the scope probe to the 'drive 0' 12V rail and initiate some drive activity the trace should hold steady at the 12V mark (about 6 divisions up the screen), and should not vary very much no matter what the drive does. In practice it may sink a little bit whenever the drive is running. What you should not see are any big dips, wobbles, or variations on the 12V line.

Disconnect the scope probe from the 12V line, move the trace to the centre line of the screen, set the input mode to AC instead of DC and set the sensitivity to 0.5V / Div.

Reattach the scope probe to the 12V rail, the trace will momentarily bounce upwards off the screen and then it will re-centre. With the scope set up this way, the large 12V DC offset voltage on the supply rail will be ignored and the scope will only show you any additional variations / wobbles / spikes / noises which are happening on that rail. Try 'exercising' the drive again, does that cause any large blips on, or movement of, the trace?

[SiriusHardware]

The drive 0 reading is a tiny bit more messy than the Drive 1 reading which stays nice and flat, but drive 0 still works ok.

If you had been able to do the second part where we would zoom in on the 'mess' using AC coupling combined with higher gain, we might have been able to get a better look at the nature of the noise. The fact that the 12V B line does look 'more messy' is cause for interest because of course it should look no different to the 12 A line when that line is powering the other drive motor. It seems odd that your hardware does not support AC coupling mode when the software clearly does.

The workaround would be to put the scope back into DC coupling mode and place a non-electrolytic capacitor with a value of perhaps 0.47uF between the 12V test point and the tip of the scope probe - this will block the large DC standing voltage but allow any 'variations' or noise on the 12V B line to pass through and be observed at 0.5V / div.

As to the odd behaviour on 2V / div, try putting your scope on 2V / div and measuring various known voltages (like the main supply voltages) both with your meter and your scope and see whether they agree or disagree.

[SiriusHardware]

Just to confirm, having looked around, it seems that the hardware side of Colin's PC-based Hantek 6022BL scope really doesn't have an AC coupling facility out of the box. There are a couple of people who have modified them to enable that facility but honestly, it's just as simple to leave it in DC coupled mode and insert a series capacitor between the test point and the scope probe.

[ortek_service]

Or maybe add the ac-coupling capacitor at 'scope plug end, as a 'scope's AC-coupling switch would do. That way, it should have the same High-Pass Filter frequency, irrespective of whether a 1x or 10x probe is being used, as will always be into 'scopes 1M input resistance.
And doing does allow the probe tip to be used normally.

You can get inline BNC Plug to Socket module boxes, for building this sort of thing into. But if there is actually provision in the 'scope's hardware to add it, that would probably be the cheapest option.


Getting an incorrect DC level reading on 2V/div, compared to 5V/div sounds a bit like an Amp / ADC saturating too soon.But this shouldn't normally happen when the trace is still visible on the screen (unless DC offset control has been adjusted to stop it going off the top - You might expect that to subtract from input voltage, but then that wouldn't read correct into the ADC unless it can also read the offset voltage).

Moderately priced Tektronix Colour-TFT display DSO's seem to handle all this quite well, but I do recall their Real Time Spectrum Analysers complaining about ADC overload as soon as you got near the top of the display Reference Level).

[SiriusHardware]

Perhaps then the way to get an accurate 2V / div reading is to set as follows:

Scope channel sensitivity = 0.2V / div
Switch on scope probe set to 'x10'.
'Scope Probe Type' setting on channel to 'x10'.

I've looked at the 'add AC coupling' mods online and there are no handily missing components with empty pads just waiting for them to go onto. The mod involves a bit of track cutting and bodging.

[SiriusHardware]

Scope channel sensitivity = 0.2V / div
Switch on scope probe set to 'x10'.
'Scope Probe Type' setting on channel to 'x10'.

Alternatively:-
Scope channel sensitivity = 2V / div
Switch on scope probe set to 'x10'.
'Scope Probe Type' setting on channel to 'x10'.

It's hard to think through the possible convolutions but if we 'inform' the scope that a x10 probe is in use with the scope on 2V / div, it may automatically multiply the actual value seen at the input x10 before displaying it.

[ScottishColin]

Does it matter which leg of the capacitor is connected to the 12V line and which to the probe?

Colin.

[SiriusHardware]

Not if you are using a non-polarised capacitor.

If you choose to use a polarised (ie, electrolytic) capacitor then in this case I would take the +ve end to the 12V line and -ve end to the scope probe tip.

The value I suggested (0.47uF) was purely off the top of my head, for this purpose a 0.1uF ceramic disc capacitor will probably do just as well.

[ScottishColin]

I'm just checking the basics and I seem to have a wandering earth.

When I switch the drive unit on, Pin 7 of the 74LS32 in U11 has a zero resistance reading to earth on the chassis.

When I use the drive, it goes up to 100+Ω and stays there.

Does anyone know how the digital board actually earths? I may have to take the drive apart to see if there's anything up with the 240V earth connection. I am confused right now.

Oh - and Drive 1 disk motor is permanently spinning too.

I'll get back to the other tests when I'm happier with this (although I'll be away until Monday so don't hold your breath).

Colin.

[SiriusHardware]

Measuring resistance in the presence of voltage is never advisable, ever, because the ohms range on a meter works by passing some current through the part under test but if the circuit is live then the chances are there is already some current flowing through the component under test, possibly in the opposite direction to the current being imposed by the ohm meter. This can give bizarre results and can also really upset the operation of the circuit under test, so this is why we always say, only ever do resistance measurements with the power off.

Use a voltmeter instead: If pin 7 of the 74LS32 and chassis ground have a good zero-ohm connection between them then the voltage difference between them, with power on and under all operating conditions, should be negligible. If you see a significant voltage appear on the 74LS32 (7) under certain load conditions that suggests you have a resistive path between 74LS32(7) and main ground.

The most immediate problem you need to sort out is that permanently active drive 1 motor. (Always fix the obvious, permanent faults before trying to fix the more elusive ones).

Since you had CR26 out for testing purposes earlier, and disconnection of that diode would make the motor run all the time, check the soldering on CR26 and if that's OK, with the drive in 'idle' mode where neither motor should be running, measure and report the voltages on

U11 (6)
U11 (4)
U11 (5)

U10 (12)
U10 (13)

Enjoy your weekend...

[ScottishColin]

Evening all.

Life eh?

So - I've fixed the motor on drive 1 permanently spinning with the advice below.

Also, after a lot of cleaning using IPA of the drive heads, I can get a slightly more successful set of reads and writes, but there's still a clear difference between Drive 1 (more successful) and Drive 0 (less so).

I have also bought some 96TPI DS/DD diskettes as the ones I have been using are not marked and are therefore probably not the right diskettes so I'll test them when they arrive (hopefully tomorrow or Friday). I guess using the right diskettes may be better....96TPI diskettes are not cheap.

To the specific measures below:

U11 (6) - 0.325V
U11 (4) - 0.234V
U11 (5) - 0.224V

U10 (12) - 0.235V
U10 (13) - 4.915V

I'm sure there's more in the current logic to be followed through but I was just wondering why I might be getting different '12V' voltages on both P3 on the main motherboard and the connected P6 of the drive board (12.02V vs 11.76V)? Happy to leave that alone if that's not really an issue.

Colin.

Originally Posted by SiriusHardware

Measuring resistance in the presence of voltage is never advisable, ever, because the ohms range on a meter works by passing some current through the part under test but if the circuit is live then the chances are there is already some current flowing through the component under test, possibly in the opposite direction to the current being imposed by the ohm meter. This can give bizarre results and can also really upset the operation of the circuit under test, so this is why we always say, only ever do resistance measurements with the power off.

Use a voltmeter instead: If pin 7 of the 74LS32 and chassis ground have a good zero-ohm connection between them then the voltage difference between them, with power on and under all operating conditions, should be negligible. If you see a significant voltage appear on the 74LS32 (7) under certain load conditions that suggests you have a resistive path between 74LS32(7) and main ground.

The most immediate problem you need to sort out is that permanently active drive 1 motor. (Always fix the obvious, permanent faults before trying to fix the more elusive ones).

Since you had CR26 out for testing purposes earlier, and disconnection of that diode would make the motor run all the time, check the soldering on CR26 and if that's OK, with the drive in 'idle' mode where neither motor should be running, measure and report the voltages on

U11 (6)
U11 (4)
U11 (5)

U10 (12)
U10 (13)

Enjoy your weekend...

[SiriusHardware]

Hi Colin, didn't notice you were back, sorry. Hope all is OK with you and yours.

It seems you've solved the problem of drive 1 constantly rotating, so that's one down.

Can you be a little more specific about where you are seeing these different 'flavours' of 12V - you mention P6 on the drive board and P3 on the mainboard, but not which pins of those connectors.

Each of the two drives has its own 12V regulated supply, 12V A and 12V B, and these are provided to the drives by separate +12V regulators so it would be surprising if the 12 A and 12 B supplies were identical to within 1/100th of a volt. The regulated output voltages from fixed voltage linear regulators are rarely so closely accurate, although in a perfect world, they actually would be. I would expect that the '12V' voltages on P6 (3) and P6 (5) would not be absolutely identical.

There is a 'case history' over on VCFED where someone who was having reliability problems with only one drive of two replaced its associated 12V regulator and that got it working reliably. However these are T03 (big metal can types) aren't they, and they won't be easy to source or cheap enough for a try-it-and-see. They may even be riveted to the heatsink?

Did you get to the bottom of the voltage measurement problem when on 2V / div on your scope? If not, then any future attempts to observe anything using that range may only prove confusing. It would really help if you could find a way to observe the noise levels on the 12V A and 12V B lines when their respective drives are running.

[ScottishColin]

Evening - thanks.

P3 pin 3 on the main motherboard connects to P6 pin 3 on the board mounted on the floppy disk drive. That measures with my meter at 12.01V at rest. That's the Drive 1 12V feed and that's the drive which is more reliable.

P3 pin 5 on the main motherboard connects to P6 pin 5 on the board mounted on the floppy disk drive. That measures with my meter at 11.77V at rest. That's the Drive 0 12V feed and that's the drive which is less reliable.

I was just wondering if 11.77V is enough of a drop to create any problems?

The VRs are riveted in so they're not going to be fun to replace even if I wanted to.

In terms of my scope - no; if I switch to 2V divisions, it affects the voltage reading and there's nothing I can seemingly do about that. If I use 5V then the voltages are read/displayed correctly but obviously there's less detail to be seen.

I have tried OpenHantek on my Linux laptop and exactly the same thing happens so I guess it;s some kind of hardware quirk.

I'll wait until the DS/DD 96TPI diskettes turn up until I do more testing I think but any views on the problems caused, if any, by the 11.77 V feed will be gratefully received.

Colin.

Originally Posted by SiriusHardware

Hi Colin, didn't notice you were back, sorry. Hope all is OK with you and yours.

It seems you've solved the problem of drive 1 constantly rotating, so that's one down.

Can you be a little more specific about where you are seeing these different 'flavours' of 12V - you mention P6 on the drive board and P3 on the mainboard, but not which pins of those connectors.

Each of the two drives has its own 12V regulated supply, 12V A and 12V B, and these are provided to the drives by separate +12V regulators so it would be surprising if the 12 A and 12 B supplies were identical to within 1/100th of a volt. The regulated output voltages from fixed voltage linear regulators are rarely so closely accurate, although in a perfect world, they actually would be. I would expect that the '12V' voltages on P6 (3) and P6 (5) would not be absolutely identical.

There is a 'case history' over on VCFED where someone who was having reliability problems with only one drive of two replaced its associated 12V regulator and that got it working reliably. However these are T03 (big metal can types) aren't they, and they won't be easy to source or cheap enough for a try-it-and-see. They may even be riveted to the heatsink?

Did you get to the bottom of the voltage measurement problem when on 2V / div on your scope? If not, then any future attempts to observe anything using that range may only prove confusing. It would really help if you could find a way to observe the noise levels on the 12V A and 12V B lines when their respective drives are running.

[SiriusHardware]

We'll let you try the inbound discs first, if that improves matters to the point where there is no difference between the drives I guess you'll be happy with that.

If not then one thing you could try, if possible, is to swap the wires going to P6 (3) and P6 (6) - it may be possible to unlock the contacts from the connector bodies, withdraw them and swap them over so that drive 0's regulator powers drive 1 and drive 1's regulator powers drive 0. If that transfers the problem to drive 1 then you will have reason to suspect the regulator which supplies 12V (B).

It is difficult to know whether that small difference in voltage is enough, it could well be making the drive 0 spindle run a little bit more slowly than drive 1.

In theory you can compare the difference in speed (if any) by measuring the frequency of the tacho output from each motor (when running) since this should be directly related to the RPM speed at which the motor is running.

Both motors ought to run at exactly the same speed when running, or more correctly, the drive 0 motor (which works intermittently) should run at the same speed as the drive 1 motor (which works consistently).

[SiriusHardware]

I forgot to mention that we still need to properly examine the 12V B supply for the level of noise on it and compare it with the level of noise on the 12V A supply, in both cases with the associated motor running. For that you are going to have to use a capacitor between the test point and the tip of the scope probe, as explained earlier, given that your scope does not appear to have an AC coupled input mode.

If there is a high level of ripple / wobble / noise on the 12V B supply then that 11.77V you are seeing there may only be the average voltage if you are using a meter to measure it. The peaks and troughs could be significantly higher / lower.

[SiriusHardware]

If not then one thing you could try, if possible, is to swap the wires going to P6 (3) and P6 (6)

In case you didn't catch it there is a typo in the above, it should be

"swap the wires going to P6 (3) and P6 (5)"

[ScottishColin]

The diskettes have turned up (along with another printer but that's a different story).

The results weren't much better really; the formatting (through the HEADER command) got further than before one every diskette, but only one completed successfully.

I then started to think about swapping things around, and I realised I have another PET here.....the older PET (3016) works much better with this disk drive than the newer PET (4032) which is a little surprising to me, but I can format many more diskettes and copy diskettes from one to the other using the 3016 which I could never do with the 4032, even by using the newer diskettes.

So when I have time I think I'm off to do some cleaning of motherboard and testing/swapping the 6520s and 3446s (which are all socketed on both PETs) to see what comparisons I can make. I'm using the same IEEE cable so it's not that.

I'll let you know.

Colin.

[ScottishColin]

I've just remembered that I replaced the 6520s in the new PET with 6821p devices. I wonder....

http://archive.6502.org/datasheets/c...821_mc6821.pdf

Colin.

[Mark1960]

The only issue I see in that applications brief is that inputs to the 6821 might need a stronger pull down to 0v. Are any of the inputs on the 6821 relying on pull down resistors?