Some of my HxC Goteks are not writting ADFs

[Jason]

Just connect the vss/gnd and the TX pin the usb adapter RX pin. ignore all the other stuff.
please use a ttl adapter, not an rs232 version.

No pins in mine to connect to. I assume it will void the warranty if I install some. Any new updates?

[Jeff]

Just connect the vss/gnd and the TX pin the usb adapter RX pin. ignore all the other stuff.
please use a ttl adapter, not an rs232 version.

No pins in mine to connect to. I assume it will void the warranty if I install some. Any new updates?

No, no update. I have a PiStorm but i still don't have the KS 3.2 ROM dump so i still can't reproduce the issue here.

[Jeff]

New major update : v5.0.0.4
Artery cpu running at 144MHz (was 72MHz) plus a lot of changes in the low level layers. this may help for this issue.

EDIT : UPD file removed, the fix is in this post :
viewtopic.php?p=24887#p24887

[PascalDe73]

Hi Jeff, Is this file is a patched version of the file included in the 5.0.0.4 ZIP file? or is it the same because I have already made the update from the zip file and here the siez isn't the same... Another revision of v5.0.0.4 ?

[PascalDe73]

New major update : v5.0.0.4
Artery cpu running at 144MHz (was 72MHz) plus a lot of changes in the low level layers. this may help for this issue.
HxCUSBFE.UPD

Hi Jeff, Is this file is a patched version of the file included in the 5.0.0.4 ZIP file? or is it the same because I have already made the update from the zip file and here the siez isn't the same... Another revision of v5.0.0.4 ?

[Jeff]

No this is not the same. Have you any issue related to this thread ?

[PascalDe73]

No this is not the same. Have you any issue related to this thread ?

Not currently. My different Gotek are all STM32 models. I didn't notice anything as a problem.
But it must be said that I use them in reading mode for IPF games converted to HFEv3.

[lowlevel]

New major update : v5.0.0.4
Artery cpu running at 144MHz (was 72MHz) plus a lot of changes in the low level layers. this may help for this issue.
HxCUSBFE.UPD

It does seem to help, however it still does not work correctly for me.
It goes through the motions now, allowing me to copy files to a adf file, and they seem to stay there... however I'm still seeing data corruption/bad data when I try to read it back. (I'm assuming is write is where it goes bad, as reading downloaded adf files seems to work ok.)

Hopefully you can eventually figure out a working solution to this issue, but I'm going to take a break from it for now.

[Jeff]

It does seem to help, however it still does not work correctly for me.
It goes through the motions now, allowing me to copy files to a adf file, and they seem to stay there... however I'm still seeing data corruption/bad data when I try to read it back. (I'm assuming is write is where it goes bad, as reading downloaded adf files seems to work ok.)

Hopefully you can eventually figure out a working solution to this issue, but I'm going to take a break from it for now.

Yes i agree that this is going nowhere this way. i need to have this KS 3.2 ROM dump to reproduce the issue and fix it. Can someone send me it by email ?
Btw does a WB 2.X on a KS 3.2 machine reproduce the issue ?

[Jeff]

Some "good" news : i managed to setup an KS 3.2 + WB 3.2 test machine (A500 + pistorm). And i manage to reproduce the issue ! :

Some notes :

- All is working fine with the FW v5.0.0.4 and the new AT32F435 Gotek.

- With the FW v5.0.0.4 and the QFN32 AT32F415 this is a different story : The track change interruption seems too slow when running with the pistorm accelerator : some steps are missed. And i can see the exact missing sector when writing.

[Jeff]

Some "good" news : i managed to setup an KS 3.2 + WB 3.2 test machine (A500 + pistorm). And i manage to reproduce the issue ! :

Some notes :

- All is working fine with the FW v5.0.0.4 and the new AT32F435 Gotek.

- With the FW v5.0.0.4 and the QFN32 AT32F415 this is a different story : The track change interruption seems too slow when running with the pistorm accelerator : some steps are missed. And i can see the exact missing sector when writing.

First issue fixed : The interrupts priorities was not set correctly...

Second issue : Work in progress :

Screenshot at 2022-08-17 17-14-29.png (42.75 KiB) Viewed 16541 times

[Jeff]

And finally here is the fix :

HxCFEUSB_V5_0_0_6a.UPD

(88.5 KiB) Downloaded 401 times

The problem with the KS/WB 3.2 was that there is almost no gap at the end of the track. The very last bytes are part of the last sector.

(Note : This fix is only for Amiga users with AT32F415 Goteks.
Please use the normal firmware if you have an STM32 or AT32F435 Gotek and/or if you are not an Amiga user).

[lowlevel]

And finally here is the fix :

HxCFEUSB_V5_0_0_6a.UPD

The problem with the KS/WB 3.2 was that there is almost no gap at the end of the track. The very last bytes are part of the last sector.

(Note : This fix is only for Amiga users with AT32F415 Goteks.
Please use the normal firmware if you have an STM32 or AT32F435 Gotek and/or if you are not an Amiga user).

Seems to work now... I successfully formatted an ADF image from the Amiga, and successfully copied a disk full of files to that image without corruption. I'm having trouble accepting it, but I think you've finally got it licked

For anyone finding this later, it's an AT32F415 based gotek in an A2000 w/wb3.2 and tf536.

[AlexOughton]

I just found that this firmware is necessary for me on Kickstart 1.3.2 as well. Only with this firmware version do I pass the write test on Amiga Test Kit, and only on this version can I properly duplicate disks from physical to image.