Latest firmware stops PAK/30 from booting.

[stephen_usher]

I'm not sure what changed in the October release of the TF536 firmware but it stops the PAK/30 ROM from booting correctly.

It gets through the RAM tests and then hangs. Earlier versions allowed it to boot to the dekstop.

[exxos]

I had that the other day.. But it wasn't firmware related.. I think it was a bad connection or bus pull ups issue.

[stephen_usher]

I had that the other day.. But it wasn't firmware related.. I think it was a bad connection or bus pull ups issue.

I actually noticed this way back when I updated the firmware. Directly before the update PAK/30 booted, 20 minutes later after the upgrade it didn't, with no other changes, so it would have to be the firmware.

[exxos]

Backtrack to a previous one then ? All the ST firmware is experimental and not had much testing by anyone yet I think.

I do hope to do my own port of the firmware, but I have to learn how to use the xlinix tools first, which I don't have time to do as I'm in the middle of moving servers for starters.

[stephen_usher]

OK, that's interesting.

Although, as I said, between the OS booting and not booting was merely a change in the firmware I did discover this morning that disconnecting the UltraSATAN did allow it to boot. So it probably *IS* a bus problem.

I know that with the TF enabled and the UltraSATAN plugged in and turned on, accessing the floppy is completely unreliable. With the UltraSATAN turned off floppy access is broken for both the 68000 and TF. This is why in the past I changed the bus pull-ups first to 2.2K all around (which broke floppy and ACSI access completely) and then up to 4.7K which allowed things to generally work OK.

I think that I'll replace all the pull-ups and put them back to stock 10K as that did seem more reliable.

[stephen_usher]

I just tried as an experiment changing the pull-up I'd put on the 1772, which was 10K, to 4.7K as the UltraSATAN pulls the data lines to 2.54V, which is not an acceptable level for either TTL or CMOS. It made mo difference. Something in the UltraSATAN is *REALLY* clamping the voltage to that level very strongly.

I may try a 2.2K resistor network this evening.

[exxos]

Which ultrasatan have you got exactly ?

[stephen_usher]

Lotherek version 1.2

[stephen_usher]

With 2.2K ohm pull-ups on the DMA data bus I'm now geting the barely acceptable 2.7V. At least the floppy is reliable with the UltraSATAN powered, though under PAK/3 it still won't boot from the U-S drive, though it does work if I run HDDriver later, having left the SD card out during the reset.

The floppy still doesn't work with the U-S powered off though. There must be another line causing this behaviour.

I've noticed that I'm now getting large bursts of noise from the power feed to the pull-up.

P.S. Holding a something with WiFi next to it generates lots of noise. It's acting as a WiFi antenna!
P.P.S. Forget that! It would help if I connected the ground lead of the probe wouldn't it?!

[frank.lukas]

You mean with PAK/3 or PAK/030 the TOS 3.06 patched for a ST with 020 or 030 CPU, right?

With what OS you boot the TF536 and how boot the patched TOS 3.06?