Raven060 (luciodra build).

[luciodra]

If it were a normal PC motherboard, in my experience, I would immediately think about the RAM, then the PSU and finally the CPU. With PSU and CPU I tried with others in my possession, for the RAM... could there be incorrect soldering? Yes. Could there be incorrect welding on all three? Yes but difficult.
I have the 68150 which should arrive soon and I will draw the first conclusions there.
Or is there something very stupid that I'm missing, like isa_slot.prg... should I put it in AUTO? I tried this too but nothing.

NB No certainty: today after a bit of testing Castaway also crashed and when it crashed Soundplay gave me active debugging on Mac...

[agranlund]

ram/rom would have been my initial guess too, but that wouldn't likely cause NMI to not respond.

Getting bad data would most likely result in some kind of crash, or even if it ended up running some kind of infinite loop looking just like a hang it would still be running in the sense that NMI would work.
NMI not working is more likely to be some kind of hardware bus hang; a cycle that never ends making the cpu more or less just wait for a signal that never comes so I would personally go down that route first.

I don't suspect the CPU as such. You've tried two of them, even non-overclocked. And that LC is very likely going to be happy even at 96mhz.
24mhz oscillator is an unknown to me. I never tried that speed so cannot rule out if there are timing issues in the bus logic perhaps causing issues at that speed in combination with the hardware you have. Or just the hardware you have at _any_ speed. That said, I've tried mine with 48 and 25 (and a few others too but so little that I wouldn't take those as proof of anything)

What I would do:
Run until you get a hang where the NMI button does not work.

- Oscilloscope on a low address line to verify if the cpu is indeed stuck waiting, or cpu status pins. Whatever is more convenient. But even without necessary stuff to do so I would skip on to the next step anyway.
- Assuming it's stuck, measure all the peripheral /CS signals to see which, if any, is being constantly asserted so we know what device the cpu is attempting to communicate with and probably waiting for.

Then take it from there checking some more signals to narrow down how far in the process it is and why a potential "I'm done" is not reaching the cpu properly.

Will be interesting to hear the result when you get the other 68150 too.

[agranlund]

NB No certainty: today after a bit of testing Castaway also crashed and when it crashed Soundplay gave me active debugging on Mac...

This is interesting but since NMI button worked I'd file it as potentially unrelated to the more pressing issue of those hard-lockups.
Still, the register printout when you pressed the button would be interesting.

[luciodra]

NB No certainty: today after a bit of testing Castaway also crashed and when it crashed Soundplay gave me active debugging on Mac...

This is interesting but since NMI button worked I'd file it as potentially unrelated to the more pressing issue of those hard-lockups.
Still, the register printout when you pressed the button would be interesting.

This is the log printout after pressing the NMI button, after Raven crashed while playing the song with Soundplay:

Code: Select all

 d0: 0000004c  d2: 0000004c  d4: 00000020  d6: 00000020
 d1: 0000012e  d3: d1960007  d5: 00000000  d7: 0000000f
 a0: 0002680c  a2: 00e01d30  a4: 00e072c7  a6: 0000232c
 a1: 00e078d0  a3: 00e078c1  a5: 00e07218  a7: 0000225c
 pc: 0117cd64  sr: 2700                   usp: 01187a2c
vbr: 00581500  tc: 00008210 srp: 00581200 urp: 00581200
dt0: 00000000 dt1: 00000000 it0: 00000000 it1: 00000000
pcr: 04310403 bcr: 00000000 ccr: a080a000

[luciodra]

Could it depend on the fact that I used SST39LF802C-55-4C-EKE for the ROMs?

[Oldskool]

In my build I had some unstable performce due to a 68150 33 from aliexpress.
Since then I have bought 3 more which are all stable.
One 40 mhz from aliexpress and 2 from ebay (33 and 40 little diode shop).
I now have installed the 40 mhz from little diode permantly. The 40 one from aliexpress is a bit loose and clearly used (but stable).
I can send you a stable one if you don’t manage or still doubt things if your new 68150 arrives.

I personlly would not fully trust the aliexpress stuff as they will can be used, broken or fake.
As I have a faulty 68150 and the isa slots are also a bit non standard (although they look new). (all from ali)
The other parts on my Raven are all new non aliexpress (or exxos or known good sources) and I don’t regret that.
I regret the ali stuff..

[luciodra]

In my build I had some unstable performce due to a 68150 33 from aliexpress.
Since then I have bought 3 more which are all stable.
One 40 mhz from aliexpress and 2 from ebay (33 and 40 little diode shop).
I now have installed the 40 mhz from little diode permantly. The 40 one from aliexpress is a bit loose and clearly used (but stable).
I can send you a stable one if you don’t manage or still doubt things if your new 68150 arrives.

I personlly would not fully trust the aliexpress stuff as they will can be used, broken or fake.
As I have a faulty 68150 and the isa slots are also a bit non standard (although they look new). (all from ali)
The other parts on my Raven are all new non aliexpress (or exxos or known good sources) and I don’t regret that.
I regret the ali stuff..

Thank you, you are very kind! I agree with everything you say, and if the 68150 I purchased give me problems I will be happy to accept your offer.
I'm confident that the problem may stem from there.

IMG_1666.jpg

Last thing: how nice it is to work remotely (Term 4.8) on Amiga 4000Tx and Raven ;))

[luciodra]

@agranlund

Why does the system seem more stable with copyback in TT-ram cache?
I would like to install Freemint but I can't figure out which archive is correct...

[Oldskool]

https://tho-otto.de/snapshots/freemint/ ... clones.zip

[luciodra]

https://tho-otto.de/snapshots/freemint/ ... clones.zip

Thank you !