1040 STE - 2 bombs on boot - strange ACSI timing

[lismarc]

@retro16:

My ACSI2STM works perfectly with my 1040 STF but I have two bombs on booting with STE. I did some mods to be sure the STE is in good shape: recapped PSU, upgraded TOS to 2.06, changed ACSI port buffers ICs 72LS445, 74LS745. In ACSI2STM: removed R9, R11. Unfortunately, until now - no success.

Then I found section Troubleshooting - how to check my Atari ACSI port is operative, like mentioned here: https://github.com/retro16/acsi2stm/blob/5.0/doc/troubleshooting.md So with digital analyser I did checking for timing on both STF, STE ports.

1) No ACSI2STEM connected:

• STF: shows exact 8 CS pulses with perfect timing 250ns/50ms,
• STE: 8 CS pulses with timing 250ns/120ms. The rest of signals like A1, IRQ, DRQ, ACK corresponds perfectly to CS as required on both.

2) Connected ACSI2STM

• STF: here both devices start talking and it works,
• STE: I can find 6 CS pulses, one "coated" with low A1 then IRQ goes to low and bombs appear.

The question the ACSI2STM firmware is ready to have 120ms spacing on init? Yes instruction says 50 ms more/less, but spacing is more than two times more!!!

Or maybe I have to find a reason why 120ms spacing in STE? Can be problem in degradation of Atari ICs or some caps.

Here 8 CS/A1 pulses from STE (no ACSI2STM connected):

[lismarc]

btw. I can share more captured signals from STF/STE. Also I'm open for whatever ideas what I can check more to obtain and solve the problem.

Additionally, I tired to recompile firmware form sources to dig into the code - debug and play a little with another timings. However, every time I compile the firmware by own it does not work. My firmware is a "bit" bigger then officially shared e.g. official release tag 5.0a: acsi2stm-5.0a.ino.bin → 57,8 kB, where my binary 59,3 kB (default setup in directives). I think the problem can be in upgraded libraries. If so, the project needs to be refreshed.

I'm using ST-Link for programming (official binaries works after programming ST-Link).