Hi All,
Last year I found a Phoenix Amiga 1000 on the street, which is now restored but could really do with some more RAM and acceleration.
Unfortunately the Phoenix is a bit picky and doesn't get on well with most of the devices I've tried. The TF534 is really close to being the solution except for a couple of problems..
With the TF534 installed the machine will crash instantly when the onboard SCSI driver (pbscsi.device) is loaded. If the onboard SCSI boot ROM is enabled then it will crash at power-on. If SCSI is disabled then the system will boot and the TF534 runs like a champ. If the updated polled-pbscsi driver is loaded later from floppy then the system will crash immediately (black screen). I've tried different cables, HDDs, termination etc. but no change. TF534 + pbscsi.device = crash.
Unfortunately I can't use the TF534 IDE either because it's not reliable in this system. I've tried very short cables, CF cards, SD adapters, HDDs etc. but I have not found any reliable combo. It usually won't format an 80MB partition without errors. I have a buffered interface on the way and will try that soon but TBH the SCSI was a big feature of the Phoenix and it just wouldn't feel right to leave it disabled.
So currently the TF534 CPU and RAM are great in the Phoenix but neither the SCSI nor IDE can be used. SO close yet so far..
Is there a way to disable the TF534 IDE or any other ideas to try? Thanks for your time.
TF354 + Phoenix A1000 pbscsi.device conflict?
Moderators: terriblefire, Terriblefire Moderator
Re: TF354 + Phoenix A1000 pbscsi.device conflict?
Hi Again,
An update on the above, the problems with the TF534 IDE appear to have been solved by adding a buffered interface board and it's now working well with a HDD and optical drive on the single channel.
It would be great to solve the conflict with the Phoenix SCSI though...
An update on the above, the problems with the TF534 IDE appear to have been solved by adding a buffered interface board and it's now working well with a HDD and optical drive on the single channel.
It would be great to solve the conflict with the Phoenix SCSI though...
Re: TF354 + Phoenix A1000 pbscsi.device conflict?
Good to hear you got it going!
A couple of random thoughts - the pbscsi seems to have a "Pseudo DMA" feature so my first guess would be that is overlapping the TF534 RAM address space. If you can, try with the RAM CPLD erased and see what happens. If it all works then you could modify the TF534 verilog to move the 4MB RAM to the upper address space and use it with addmem. Anyway that's where I'd start looking - good luck!
A couple of random thoughts - the pbscsi seems to have a "Pseudo DMA" feature so my first guess would be that is overlapping the TF534 RAM address space. If you can, try with the RAM CPLD erased and see what happens. If it all works then you could modify the TF534 verilog to move the 4MB RAM to the upper address space and use it with addmem. Anyway that's where I'd start looking - good luck!
Re: TF354 + Phoenix A1000 pbscsi.device conflict?
That's unlikely because the 534 places the ram in the Z3 space. What is more likely is that the driver is trying to DMA into Z3 which obviously won't work.dalek wrote: Mon Sep 26, 2022 3:16 am Good to hear you got it going!
A couple of random thoughts - the pbscsi seems to have a "Pseudo DMA" feature so my first guess would be that is overlapping the TF534 RAM address space
So one solution would be to find a way to disable SCSI DMA or to force it to do DMA into Z2 ram only. You may also want to try the TF536 as that one is generally a _lot_ more compatible with expansions than the 534.