Interpreting uPD765 FDC result bytes

[Bryce]

Just a question: Have you tried formatting a new disk in this setup and reading it back?

Bryce.

[arnoldemu]

from the diagram I can see:

    support for 4 drives
    double sided supported
    uses terminal count input; so this needs to be asserted correctly to stop the abnormal termination of command (assuming it's not a genuine error)
    looks like it's double density only

missing address mark can also come if it's attempting to read a double density disc when set to single density and reading a single density disc when set to double density.
Can you confirm the clock to the fdc is correct?


Another question: It seems you have a broken p2000c which will not boot and you are fixing it?

[JonB]

@Bryce: can't use the machine to format a disk because it won't boot at all. However I do have some disks that were supplied with it and are readable on a PC with special software (22DISK); they look like CP/M boot disks.


@arnoldemu: yep, it's a P2000C that won't boot. Also has a problem with its terminal board, but that is a separate issue, you can connect a PC via RS232 to the mainboard and forget the terminal board altogether (which is what I have done; I'll fix the terminal board once the thing is booting into CP/M).

It has an inbuilt ROM based monitor / diagnostic tool called IPL that allows you to read and write tracks to a disk without booting. It is this command (read track) that is failing with missing address mark, but write track also fails. I think I t can read lower density disks; I assume this because some software was supplied on 160k floppys (assumed SS/SD) and there is a CP/M utility that you use to transfer to 640k disk.

As previously mentioned, the service manual says we should get 500khz at TP2 when the CLR line (74LS221 pin 3) is grounded. As I do not get this (it's reading 303khz), and I can't get the adjuster (trimmer labelled 4712) to alter it, I think the circuit is broken.

I will check the circuit as requested later on (when I get home).


Thanks
JonB


Oh, a thought. Is it possible some discrete component has failed like a capacitor? Some of them are electrolytic.

[Bryce]

Well I've had a chance to look at the schematics and the datasheets now. The 293 seems to be working normally. Pin 4 should be dividing the signal on pin 11 by 4 which it's doing: 1200/4 = 300. So to get 500 on TP2 you'd need 2Mhz on pin 11 which you don't have. So now the question is - Why not?

MC4024 is a voltage controlled multivibrator. It's being used in the voltage controlled mode, so Pin 2 (Vin) decides the frequency that it outputs. First off check the other pins to make sure it's getting what it needs:

Pin1 and 14 should be at 5V
Pin 5, 7 and 9 should be to GND

All good?

According to the datasheet, the MC4024 can output from 1Mhz to 5Mhz by varying the voltage to pin 2 between 2.5V and 5V. So to get 2Mhz pin 2 would need to be at about 3.25V. Is it? If not it's time to go back a step further and take a look at the 741 circuitry.

Bryce.

[JonB]

Bryce, I have this so far (measurements already taken round the 741)



+12 and -12 volts are present. The output voltage is .6v (at Pin 6).


So if it needs to be higher, we might conclude that some stuff round the 741 is damaged. The 741 inputs are both at 400mv wrt ground.


If you can give me a list of test points, I can take measurements in one go. Unfortunately I have a tiny amount of time in the evening to do this, so it's better to try and do as much as possible in each session.  My train is already 45 mins late tonight so that cuts down my evening time even more.


And I am really keen to fix this machine. It's so frustrating....

[kevin.c]

.......... So to get 2Mhz pin 2 would need to be at about 3.25V. Is it?..........
Bryce.

Bryce , this above  is valid under condition that PLL loop is closed: to pin 30 RDD is applied stable Data Read signal, or if control line "SYNC pin24" is set to "0", then to PLL is applied reference 250kHz from pin 9 of 293 counter. In this case PLL is synchronised and stable. IMHO it is for faster synchronisation in case switching PLL to RDD signal.

[JonB]

Just remember we are grounding the CLR on the left hand 221 before measuring frequency. I think this is intended to lock it down (so it won't depend on the disk signals). Have I got that right?

[Bryce]

Bryce , this above  is valid under condition that PLL loop is closed: to pin 30 RDD is applied stable Data Read signal, or if control line "SYNC pin24" is set to "0", then to PLL is applied reference 250kHz from pin 9 of 293 counter. In this case PLL is synchronised and stable. IMHO it is for faster synchronisation in case switching PLL to RDD signal.

According to the service manual, this is also valid if he grounds pin 3 of the 221. So without any disk spinning or data reading he should be able to do all these tests.

Bryce.

[kevin.c]

I see Bryce... It looks that clearing this pin 3, the phase comparator is reset to "0".
JonB check following please :
1. ground pin 3 of left  221 as before
2. check if pins 5 and 9 of LS112 are both in "0"state
3. measure voltage on pin 6 (output) of 741, turn gently current source regulator R4712 and observe voltage change on 741 output.
For any case mark regulator 4712 factory position...

[Bryce]

Good check, but I've a feeling the voltage won't change much, if at all. My bet is on a failed Zener at the moment.
Check the Zener output voltages while you are at it.

Bryce.

[JonB]

Grounding Pin 3 of 221




Pin 5 0 state with pulses about every 500ms
Pin 9 0 state with pulses about every 500ms


Does not make sense but anyway...


Pin 6 of the 741 is at -300 mV and adjusting the pot makes about 30mV difference either way.


That said I am having difficulty holding the wire onto 222/P3 and trying to make measurements. I will need to solder a strap to GND in place, because I don't have any of those little clip things.


Please see attached image. Might it be this (highlighted) bit that's not working? Feels like I'm fumbling in the dark at the moment.

[kevin.c]

Check the voltage on zener diode as Bryce said.

[JonB]

Good check, but I've a feeling the voltage won't change much, if at all. My bet is on a failed Zener at the moment.
Check the Zener output voltages while you are at it.

Bryce.

Don't get you. The voltages before or after the zeners? You mean 7505 / 7503?


Zener: -> points to black band (output?), measurements at both sides:


7503 0v -> 5.5v
7505 -12.15v -> -12.15v

[kevin.c]

To simplify electronics jargon... 
Voltage measured on Zener diode 7505,  as shown on schematic, should be always 5,6V  (tolerance 0,1-0,2 V)

[JonB]

http://www.nxp.com/documents/data_sheet/BZX79.pdf


That's a voltage regulator. The input and output voltages are the same (measuring wrt GND) so without actually measuring across it as you show in your annotation, I'd say it will measure at 0v. But I'll check it when I get home tonight.

[kevin.c]

....The input and output voltages are the same ....

This means you found the fault! The Zener diode (=voltage regulator) is broken down. (short circuit). I hope there is no more  faulty elements because of short circuit. As you can see on schematic,  -12V came directly to resistor 3314.
By the way... Bryce bet on good horse. 

[Bryce]

  Looks like my suspicions were right. Unfortunately, this might also mean that further parts have been damaged (looking at you BC549 - 7302 not 7303).

Bryce.

[JonB]

OK, I should have 2 C5v6 Zener diodes and 2 BC549s by the weekend. I'll swap out the first Zener and recheck.


Is there a way to test the ctransistor or should I just swap it out anyway? I'm reluctant to do it because both transistors are bound tightly together by a rubber tube (why?) so will be tricky to take out.

[Bryce]

OK, I should have 2 C5v6 Zener diodes and 2 BC549s by the weekend. I'll swap out the first Zener and recheck.


Is there a way to test the ctransistor or should I just swap it out anyway? I'm reluctant to do it because both transistors are bound tightly together by a rubber tube (why?) so will be tricky to take out.

Binding the two resistors is to ensure that they both experience the same thermal drift (gain change due to temperature). This is often done on voltage reference circuits when one Transistor is going to have a higher load than the other. By binding them together they share the same temperature and hence the same amount of drift.

Bryce.

[JonB]

Hmm thought as much. So I will need to bind them somehow after replacement. Hopefully I can slide the sleeve off without damaging it.

[kevin.c]

For the begining replace the faulty diode only. First measurement of 741 output voltage will answer if transistors are ok.

[JonB]

I was going to ask about that. Turns out Maplins had no transistors in stock so I only have the Zener. Do you think it is likely the replacement diode will blow if the transistor is dead?


Anyway, I'm going to do it. First bit of repair work on the P2000C - it's been diagnostic up till now.

[TFM]

So the P2000C has two Z80 CPU's right? At which speed to they run? Is it 2.5 MHz?


It's hard to get docs about that nice piece of hardware

[JonB]

Supposed to be one of the best CP/M machines there is.


Twin CPUs, both 4Mhz. One for I/O, one for general CPU tasks. 64K RAM, has DMA for disk transfers. 3 second cold boot, if YouTube is to be believed. Also has built in interfaces: SASI HDD , serial, parallel and external floppy plus external monitor connector in case you don't fancy squinting at the dinky 9" screen. Supports high rez graphics (monochrome). Twin 640k 5.25 floppies (Teac FD55); modern, low profile keyboard. Has MSDOS capability if optional 8088 board is fitted (not present in mine, unfortunately). There are other boards that could be fitted to the one slot, but they are rarer than hen's teeth.


Oh and it's "portable" too, ha ha!


Yet another computer museum - P2000C


All docs & disk images at the above link, including (crucially) the service manual. Mine has all user documentation plus some extra software on disk.

Super fast boot video :

Now you can see why I want to repair it so much. It's pretty, compact, capable and totally kicks CP/M ass.

[JonB]

Drat, foiled again. Maplin had the transistors but not the Zener diodes (their web site lied). So now I have to wait "3-5 working days" before I can progress.