HD Floppy drive in A1200

[pipper]

I have a 1.4D A1200 that came with the ESCOM floppy modification. I performed the ESCOM floppy fix (to get proper RDY signal) and replaced the Panasonic drive with a Chinon FB-375A.

The issue, though, is that HD floppies won't be detected. It boils down to the OS reading out the wrong 'Drive ID' and thus assuming that it can't handle the inserted disk. The tool 'driveid' reports that the drive reports 0x55555555, but looking at my scope it produces 0x55555555 on the physical /RDY line, thus it should have been reported as 0xAAAAAAAA logical.

The same drive works and is reported as HD drive on an A3000, drive id reports 0xAAAAAAAA there.

The only difference in hardware I can see between the A3000 and the A1200 is that the /RDY line of the internal drive is connected straight to the CIA in the A3000, while it goes through a 68Ohm resistor on the A1200.

Is there a bug in the kickstart software that makes the timing between activating /SEL and reading back /RDY very sensitive?

Any hints are welcome, thanks!

[DanyPPC]

I know that if you use a standard disk drive Amiga you don't have to modify nothing on the motherboard.

[AmigaHope]

Originally Posted by DanyPPC

I know that if you use a standard disk drive Amiga you don't have to modify nothing on the motherboard.

To clarify, he's talking about using a high density drive to read high density floppies, which on the Amiga requires a special drive that spins hd disks at half-speed because Paula can't handle the higher bitrate if they spin at normal speed.

He has one of those special drives, taken from an A3000 (which shipped with it), which works fine in the A3000, but fails to detect HD floppies when connected to his Escom A1200.

[pipper]

Originally Posted by AmigaHope

To clarify, he's talking about using a high density drive to read high density floppies, which on the Amiga requires a special drive that spins hd disks at half-speed because Paula can't handle the higher bitrate if they spin at normal speed.

He has one of those special drives, taken from an A3000 (which shipped with it), which works fine in the A3000, but fails to detect HD floppies when connected to his Escom A1200.

AmigaHope summed it up perfectly

[cpiac64]

1200 have 3,1 kick?

[pipper]

Yes, it does have Kick 3.1. Kick 3.1.4 has the same issue.

[cpiac64]

next week i can test some drive hd and report the result

[solarmon]

What does Amiga Test Kit report the drive ID as, on the A1200 and on the A3000?

[pipper]

I did some more research on the scope. It seems the way the drive ID mechanism is initiated on an A500/A3000 is different from how it is done on the A1200.

See the attached pictures.
(I inverted the signals as they are all active low, so the drive ID values I note down are already in logical domain)

On the A500, when the /MTR0 signal deactivates and the /SEL0 signal activates, the /RDY line responds with a 1.
The rest of the sequence thus becomes 0xAAAAAAAA which identifies the drive as HD 150rpm drive.

On the A1200 the very first time /SEL0 activates after /MTR deactivates, the drive responds with a 0. The rest of the sequence then becomes 0x55555555 which makes the OS reject the disk.

If you look at the last image, I overlaid both sequences. The timing is different on both machine. And I wonder if there's a difference in the Kickstart code that initiates the sequence?!

[Bruce Abbott]

Originally Posted by AmigaHope

To clarify, he's talking about using a high density drive to read high density floppies, which on the Amiga requires a special drive that spins hd disks at half-speed because Paula can't handle the higher bitrate if they spin at normal speed.

To clarify, he's saying that if you are using a standard (or HD) Amiga disk drive then you should undo any modifications that have been done to the motherboard for use with PC floppy drives.

Fixing the AT A1200 FDD connector

[Bruce Abbott]

Originally Posted by pipper

I did some more research on the scope. It seems the way the drive ID mechanism is initiated on an A500/A3000 is different from how it is done on the A1200.

See the attached pictures.
(I inverted the signals as they are all active low, so the drive ID values I note down are already in logical domain)

It should follow the sequence described here, but I am having difficulty analyzing it due to the inverted signals. Can you show us the waveforms for the entire sequence (32 serial bits plus anything immediately before and after) with the actual signals (not inverted)?

[pipper]

Attached you'll see the sequence at various zoom levels, non-inverted.
I focused on the start of the sequence as this is where the problem starts. The rest of the sequence is just repeating.

My guess is that the drive is deactivating /MTR to reset the drive_id sequence and would start reporting the alternating 0s and 1s clocked by the /SEL.
But the deactivation of /MTR and the first activation of /SEL virtually happen at the same time and thus the reset may not have a chance to fully reset the circuit. In the last closeup you can see that there's a small dip in /RDY where it looks like it wants to report a low, but for some reason can't pull down the line.

If there was a small pause between deactivating /MTR and the activation of /SEL, the drive could have more time to do its thing. I don't know why it works on A3000 and A500, but not on A1200... But as I said in the first post, there are hardware differences in how /MTR0 and /SEL0 are being generated. On A500 and A3000 its a piece of discreet logic, while Gayle is doing it on the A1200.

[cpiac64]

arrived in my lab a dead fz-357a, need time for repair and test on a1200

[cpiac64]

Sorry, my drive don't work

[Bruce Abbott]

Originally Posted by pipper

Attached you'll see the sequence at various zoom levels, non-inverted.

This confirms what I thought - the ID bits are not inverted. The first bit should come out during the first /SEL pulse after the pulse that turns the motor off. On the A500 this is a 1, whereas on the A1200 it is a 0. Another reason we know the bits aren't inverted is that the ID for 'no disk' or no drive is 0xffffffff, which equates to /RDY being continuously high (held there by the CIA port pullup resistor).

I cannot explain why your drive appears to send an ID code of 0x55555555 on the A1200. I don't have an HD Amiga drive, but I found a circuit for a modified HD drive which includes an ID generator. I will try building this and see what waveforms and drive detection I get from an A500 and A1200.

If there was a small pause between deactivating /MTR and the activation of /SEL, the drive could have more time to do its thing. I don't know why it works on A3000 and A500, but not on A1200... But as I said in the first post, there are hardware differences in how /MTR0 and /SEL0 are being generated. On A500 and A3000 its a piece of discreet logic, while Gayle is doing it on the A1200.

The waveforms look close enough to identical to me that timing shouldn't be a factor (it would be very poor circuit if it was sensitive to such small changes). However to test my ID circuit I intend to extract the code from the ROM and modify it to produce different timings, so if the timing is critical it should be discernible.

Now to find that 23 pin connector I had stashed away...

[pipper]

Thanks for jumping on this mystery!

I suggest using the internal drive connector (that's what I'm doing). /SEL0 and /MTR0 seem to be specially generated. I have not attempted to see what happens when the drive is attached as external drive.
Besides the FB-357A (the mystery drive), I also have an FZ-357A which currently doesn't work (the drive ID mechanism always just returns "DD floppy drive").
I am hoping to fix this soon, but still waiting on a part. Lets see how the FZ will perform versus the FB.

[jbenam]

Hi pipper,

I have a working Chinon FB-357A (from my A4000) I can test on my A1200. I will need to dig it out, but no problems

Where do I find the driveID tool?

[pipper]

driveid is part of this archive:
https://aminet.net/package/docs/hard/PCFloppy2Amiga

It also comes with a drivespeed tool that measures the rotational speed (similar to how AmigaTestKit does it).

[jbenam]

Originally Posted by pipper

driveid is part of this archive:
https://aminet.net/package/docs/hard/PCFloppy2Amiga

It also comes with a drivespeed tool that measures the rotational speed (similar to how AmigaTestKit does it).

There we go - drive ID 0, just like on my A4000D.

RPM tests good too, ~300RPM just like my A4000D. It also reads a floppy without any issue.

[pipper]

You seemed to have inserted a DD disk. Please try with an HD floppy disk.