Diagnosing an issue with C8PDF drive

[lowrybt1]

Hello, I'm trying to figure out why an original 8" drive (Siemens FDD-100-8D) will not track to home when I try to boot disks. (Yes, it's a 43 year old drive...why should I expect it to work) Here's what I know:

* Using the C8P with a GoTEK and D-13 and 50-pin to 34-pin FDD adapter works fine when I boot HFE disk images

* When I replaced the D-13 with the original A-12 and connected to an 8" drive that came with C8P (appropriately strapped to be Drive A and with terminating resistor installed), physical 8" disks will not boot

* If I first manually (and carefully) move the drive's jack-screw so the R/W head is in the trk 0 position and then press D, the disk boots great. However, if I try to home to track 0 from the kernel, 65D3.3 reports OK but the drive mechanism does not move the R/W head to track zero. Any further attempts to load code/data from other tracks yields a #5 or #9 error (mostly #5). To get the disk to boot again requires manually moving the jack-screw so the head is repositioned back at trk 0.

*Once rebooted, if I DIR TT to a higher track (like 40), the drive tracks just fine.

I figure the likely culprit is an old drive with either a bad track zero sensor or a bad component on the drive's PCB somewhere that processes the signal from the track zero sensor or both. But just to be sure, based on the points above, could there be some signaling issue that starts with the 505 board?

I do have spare (untested) Siemens FDD-100-8D PCB's and drive hardware which I'll try to test but before going there want to rule out any other possible causes.

Thanks, Tom

[Thomas]

Looks promising, that DIR TT is working to upper tracks.

As the Siemens drives have plenty of test points, you should check the Track 00 sensor.
Look for TPA2 (photo sensor output), when the drive is operating or
LTRK00 (or TPA4) on the LSI chip pin 1 (TRK00 input)

Another problem still can be the Stepper motor drivers Q10..Q12. If one fails, the drive might still move in one direction but not backwards. So here you should check if the 3 phases are working as intended.

[Mark]

Upon 'D' from the boot prompt, the OSI will step the head up one and then down repeatedly until track 0 signal is active. Since you can boot with the head moved to track 0 it's likely the problem lies with the stepper circuitry since TRK0 must be active to boot. Also when the head is not over track 0 the 'D' command will continuously send step down signals until TRK0 is active.

David Gesswein's floppy disk test program has a "Status" command that lets you manipulate head position manually up & down & view port bits which may aid you in solving the problem. (S)elect the drive & Step (U)p or (D)own.

There is a bootable disk image include in the archive on my OSI Tools page which you could convert to HFE, or you can load it from serial port.
After booting you can interact with it on the OSI or via serial.

Disk Test Status

DiskTestStatus.gif (6.58 KiB) Viewed 12056 times

One of my 8" drives had 1 output pin of the LSI chip spontaneously die (part of the select logic) which I was fortunate to fix with a couple TTL gates in its place.

Good Luck!

[lowrybt1]

Thomas and Mark,

Thank you for this insight. I'm familiar with testing transistors like the TIP31A using a multimeter and will give that a try tonight. As for checking the test points on one of these boards, that is new for me. Question: what mode should the multimeter be set to when checking TPA 2 and 4 and where should I position the common/black probe while checking the test point?

Tom

[Mark]

Looking at the FDD 100-8 schematics on page 36 It looks like TPA2 is under 5v, passively pulled high & switched lower by the photo transistor, and TPA-4 should be logic level hi/low... so multimeter black goes to an available ground like one of the grounded resistor leads or (perhaps TPE-3?) etc. Since they are both part of the same circuit the important one to check is if TPA-4 has a big swing between < 0.8v low and > 2.0V high when Track 0 sensor is open or blocked.

The LSI chip generates the logic level phase controls for the stepper motor. You can follow the signals from the LSI chip through the driver chip & finally to the transistors. Make sure the 12V zener VR5 is providing the 12v and the 24V supply is correct.
As the phase are sequenced the transistors should ground each phase one at a time, the order depends on the direction of rotation desired. There are additional ICs involved with step logic and step inhibit that seem could change depending on strapping/board options.

It seems strange you could step up correctly but not down if one of the phase drivers was bad, but who knows?

Let's see what you find out!

[lowrybt1]

Updates and corrections. In fact, (Mark you were right) after manually setting the r/w head to trk 0 and booting into 65D3.3, the drive was able to track up and down the full range of tracks except when commanded to HOME or INIT, at which point the drive mechanism would not move. A check of TPA4 and TPE3 showed 0.0V whether or not the track 0 sensor was open or blocked.

I then swapped in two spare PCB's, each tested on two drive mechanisms (all FDD-100-8D) and I encountered the same issue. Drives would boot if I manually positioned the R/W heads to trk 0 but could not HOME on their own. I've not checked the test points on these spares (yet). Gonna take a break. Symptoms point in the direction of all 3 PCB's having LSI/Trk 0 issues.

[Mark]

So now you have me believing it really is just the Trk 0 sensor as Thomas suggested. You can use a flashlight shining on the phototransistor to determine if its the LED or transistor.

I didn't review the code, but I'm thinking the HOME command steps down until TRK0 signal is active. If it is always active then it has nothing to do/already at track 0.

Normally if the light is blocked @ track 0, the TPA-2 signal is pulled high & TRK0 drive signal is active. When moved away from track 0 the LED is visible, the photo transistor conducts & TPA-2 voltage should go lower, which make TRK0 inactive.

If you pull TPA-2 low (with a resistor (like 1K or so/not critical) or jumper) It should simulate the photo transistor is not at track 0. The head should step down continuously upon 'D' from the H/D/M Reset menu. Be careful though, 'cause the head will hit the endstop as there is nothing to tell it to stop. Moving the head up to the other end should give you a bit of time to see if it moves towards 0 & either pull the jumper or reset the OSI before slamming into the low end.

Or you can use the attached code to view the status of disk PIA PORT-A. It just initializes the disk PIA, selects the drive & shows the port status TRK0 bit as a graphics character on the screen near the monitor HEX line. (Reset the OSI, Press M, Press L & send the DispPortA.lod file via serial, or enter bytes manually perhaps).

My guess is if trk 0 is always active, you will see the "racecar" / graphics character 0 no matter where the head is located.

Trk 0 active

On Trk 0.gif (989 Bytes) Viewed 12027 times

The )( character shows the Trk 0 signal is inactive. (Floppy signals are inverted/active low).

Trk 0 inactive

Not Trk 0.gif (1.05 KiB) Viewed 12027 times

Code: Select all

0300 78       SEI
0301 A0 00    LDY #$00
0303 8C 01 C0 STY $C001
0306 8C 00 C0 STY $C000
0309 A2 04    LDX #$04
030B 8E 01 C0 STX $C001
030E 8C 03 C0 STY $C003
0311 88       DEY
0312 8C 02 C0 STY $C002
0315 8E 03 C0 STX $C003
0318 A9 FF    LDA #$FF
031A 8D 02 C0 STA $C002
031D AD 00 C0 LDA $C000
0320 29 02    AND #$02
0322 8D D0 D0 STA $D0D0
0325 4C 1D 03 JMP $031D

[Mark]

Just a followup,
On my FDD 100-8 TPA2 goes from 4.7v at track 0, to 0.13v at any other track.
Track 0 active does not prevent the drive from trying to step lower if commanded to.