CBM 3040 Unable to read disks

[Wreck]

Hi

In a previous repair attempt of a CBM 3040 I got stuck with three red LEDs. I haven't made any progress on that drive, but I got my hands on another 3040, which has a good mainboard, as I can use the various disk commands to access both drives.
Problem now is, I am unable to read anything. Also, I cannot format any disks. Both drives react the same; rattle "home" to track 0, then move back to roughly the centre of the disk.

What would the next suggested course of action?

Thanks for all inputs

Joe

 

[Hugo Holden]

I'm not much of an expert on drives, but have some experience with them, in that they have many similarities to other machines I am very familiar with such as video tape machines with similar servo control systems. And I have repaired a few disk drives.

If the drive is out of radial alignment, which they can be if people have been previously working on them, and the drive otherwise working, you can still format new disks and play those back in the drive, it is just that they won't be compatible with standard drive alignments or work in other drives.

If you cannot get the drive to work at all, obviously the first things to check are the power supplies, then the servo systems that position the heads and rotate the disk and the function of any sensors. (you didn't say if the disk was spinning only mentioned rattling and the head position). When you are checking the power supply rails, check the power supply distribution throughout the board. For example there are often series inductors, sometimes resistors, these can go spontaneously open, or fuse if the Tant capacitor they feeed shorts out.

If all these appear ok and the drive is still not working, one starts to wonder if the heads are recovering any data from the magnetic media. Probably, the most likely part to fail in a vintage drive is not the IC's, not the capacitors in many cases, but the heads, because they are wound with very fine wire and it corrodes through near the solder terminations inside them. The same thing happens to many devices like this for the same reason, for example digital delay lines, inductors with very fine wire etc.

After doing the above checks I would then put in a known good disk in the drive with something on it and put the scope on the output pf the head pre-amplifier IC looking for evidence of any recovered data or signal as the head moved across the disc surface. Even as it crosses tracks you should be able to see a triangular shaped wave envelope that look similar to to modulated RF. If the heads are open or defective, you won't see any signal, just noise. If there was adequate signal recovery then its a matter of following the signal pathway.

If there is some sort of logic fault going on in the control system, the error lights might have some meaning and be of some diagnostic use, this is where the manual also comes in handy.

Fault finding a drive is little different than most other computer repair work, except for the electro-mechanical considerations of the drive transport and heads & head positioning & sensors. Always start with the basics first. The manual and the scope is your friend.

You will probably get more specific advice from somebody very familiar with this drive model, that always helps.

 

[Wreck]

Dear Hugo,
Thank you very much for your detailed answer. I have done the below work on the drive so far, and it looks quite promising!
- General cleaning.
- removal of the filter cap in the AC power supply part (the drive is from Germany).
- Replacement of C31 near VR2, which was blown already when I opened up the machine for the first time.
- Checking voltages of VR1,VR2,VR3.
- cleaning of the heads.

Only then did I connect the drive to a trustworthy CBM8032 and "fired" it up (no smoke).

The three LEDs come on, and go off after a few seconds

I have one new Commodore CBM test/demo disk (P/M4040037), fairly recently acquired, plus several older disks.

Firstly I tried to insert an older disk into drive 0, and because the 8032 "knows" BASIC4, I tried the following;

load"$0",8

searching for $0
?file not found error
ready.

This was followed by an attempt to format;

header"new" ,d0

are you sure ?y
? bad disk

ready.

The above format command made the drive (0) spin up, the head went to its stop and "rattled" 10 times, then went back to the center of the disk, drive spin stopped and the "bad disk" message appeared (the red center LED on the drive stays off).

When trying the load"$0" command, it results in the "file not found message", with the red center light staying on (which, I guess, is normal).

Now narrowed the problem down to the following; either all my floppy disks are bad, or both drives are bad. As I have another 3040 (which doesnt work), I tried swapping the mechanical drives around; essentially I got the same messages "?bad disk" and "?file not found".

Also, I tried the "print ds" function, which returned 21 (read error). The "print ds$" returned 21,read error,52,11.

My conclusion is, that it is very unlikely that all four disk drives are faulty, but at the same time it seems strange that all floppy disks are faulty as well (I have at least ten different disks around, including the "new" demo disk).

So the next step will be to check with my scope to yield some data from the head pre-amplifier IC. I can see TP3 and TP4 on the analogue board being used for the alignment process, which is probably not what I am looking for.

Any thoughts where to start looking for the head signal?

Thanks

Joe

 

[Hugo Holden]

The head signal itself is very low level and difficult to see even with a scope that goes down to 5mV/cm, plus usually the actual head signal is a differential signal for noise immunity, and the two inputs pass to the + & - inputs of a head amplifier OP amp arrangement. But, at the output of that amplifier, or at least the stage following it, the signal should be easy to see on the scope, and it becomes ultimately ground referenced by the amplifier system. Its a matter of checking the schematic to work out the best test point. Can you post that part of the schematic ?

 

[Wreck]

I think that would be TP1 and TP2 at UA5 (LM592)?
Out of curiosity I swapped the analogue board from one drive to the other, and it is now able to read disks (also formatting). As I would like to understand more about what is happening, I will continue with the troubleshooting

 

[Wreck]

I had an hour or two to get acquainted with the analogue board of the CBM3040. As mentioned before, I have two of these boards, one not working (produce read errors all the time), and the other only "half" working (drive0). I tested both mechanical drives, they both work (reading & writing).

Now I am trying to fix the "not working" part of the analogue board (drive1).
By looking at the schematics, I can see transistors (Q1-Q8). Q1-Q4 being used for drive1, and Q5-Q8 for drive 0.

I found Q2(N4403) showing no voltage on base (actually 0.54V), whereas the "mirror" part Q5 having 12.0V.

The signal comes from UC1, output PIN8. It branches to UB2, again PIN8 output to UB1(7406), and once more output on PIN8 via R6 to Q2. I hope I haven't made any thinking error so far.

As far as I can tell, it looks like part of the 7406 at UB1 is faulty (PIN8 output).

I will try to replace the 7406 and see if there is any improvement

Any inputs appreciated

Joe

 

[Wreck]

small update; I re-checked the board, and replaced Q2 (4403 Transistor), which now has an 11V output. When trying to access drive1, I can format a disk until an "error#21 read error on 18.04" appears. Strangely the "rattle" of drive1 is more subtle than of drive 0, but who knows. Drive 0 works like a champ.

Next step is to re-check all voltages and input/outputs of the various ICs on the analogue board, maybe I can find something out of the order.

Joe

 

[Wreck]

OK, another day, another update, good news overall; Further troubleshooting revealed Q3 (4401 Transistor) also being defect. I replaced it, and both drives now work as they should; formatting, copying, reading etc all work

Now since I fixed one analogue board, I am getting ready to fix the other one; I will also post my progress here in this thread.

Joe