REV 3 - REV 5 - The beginning (ST536)

[exxos]

Well, the good news, is it ran all day again without issue. The bad news is, it wasn't supposed to :lol: :roll:

I had a known buggy setup. Left it over the weekend. Made firmware updates.. Tested fine.. Went back to the buggy setup and its magically not buggy now :stars: I even bashes about everything, bent the 536 about and heated the whole thing up. Still kept chugging away just fine..

I'm going to leave it running GB6 overnight and see what happens. I'll likely flash the current firmware in the other boards tomorrow and leave those running..

There was a signal which was hardly reaching a logic high on the CPU about a week ago. Doesn't look like I took any notes or images. According to AI I never mentioned it. I remember putting a pullup on a few signals but made no odds. But can I figure out what I tested.. I got it in my head it was AS30 for some reason. But scoped it out and doesn't seem to be that..

[exxos]

GB6 was running all night again. Annoying when things don't fail on demand.

20260313_092651.jpg

20260313_092658.jpg

[exxos]

Decided to give @coonsgm board another look at.

Booted up into EMUTOS 8mhz fixed to see if it gave any more clues...

IMG_4790.JPG

IMG_4791.JPG

IMG_4792.JPG

Cannot say it did :lol: I will have to continuity check the address bus for starters to see if that is FUBAR somewhere.

[exxos]

That board is an absolute puzzle :pullhair:

Mostly is just some sort of exotic crash screen generator at the moment. Sometimes EMUTOS will display the splash screen and sometimes get to desktop.

I've continuity checked the 030 back to the 58K headers and PLD and all seems good there.

IMG_4793.JPG

IMG_4794.JPG

IMG_4796.JPG

IMG_4797.JPG

It's running at stock 8mhz. While it is running from the recreated clock, it won't even boot if I run it from the ST's master 8mhz clock :shrug:

Answers on a postcard....

[PhilC]

Rob it of parts and put it in the dustbin?

[exxos]

Rob it of parts and put it in the dustbin?

Yeah I am tempted just to bin it.. The problem is in my mind, what if it's a hardware or firmware fault which is just masked on the later boards..

It could be just GND bounce on the earlier boards causing issues.. It would make it practically impossible to diagnose the problem though. But you would still think that even 8MHz it would be stable. I mean it did run fine at one point. I've changed the PLD, CPU socket, SDRAM already. Nothing seems to be wrong with the PCB.. So why is it broken :roll:

[PhilC]

Rob it of parts and put it in the dustbin?

Yeah I am tempted just to bin it.. The problem is in my mind, what if it's a hardware or firmware fault which is just masked on the later boards..

It could be just GND bounce on the earlier boards causing issues.. It would make it practically impossible to diagnose the problem though. But you would still think that even 8MHz it would be stable. I mean it did run fine at one point. I've changed the PLD, CPU socket, SDRAM already. Nothing seems to be wrong with the PCB.. So why is it broken :roll:

The hardest bit with R&D is knowing when to quit with a design and move on to the next one. Otherwise you end up going down the rabbit hole to oblivion.

[exxos]

The hardest bit with R&D is knowing when to quit with a design and move on to the next one. Otherwise you end up going down the rabbit hole to oblivion.

True. I could move onto the next design and getting it manufactured but then it just more expense and more boards in the mix.

I also have a later 5.50 board which died for no reason as well.

A second 5.50 board gets to ROM verify OK then locks up.

Overall that's the problem, to many factors going on to even get a idea where to start.

The kicker as well, is your TF536 goes nuts when MAPROM loads. Its why I started to question the H5.. I changed to a H4 at one point. Same issues.

The series resistor board which adds 100R in every signal doesn't seem to help either.

So currently I have basically 4 boards which all behave differently.

I just tried my second rev 5.62 board, it copied ROM to RAM, then reset and locked up. So 5 boards now then. My other 5.62 board has runn running for days almost without issues.

I mean , if it wasn't just running at 50mhz, at least that would be something to go on. But even stock speed causes headaches. Even if I strip the firmware right back to bare bones which I've done in the past, it doesn't seem to help either.

I will try and see if EMUTOS works better or not on anyboards. I mean its possible my patched 206 has a bug somewhere. So need to rule the OS out next. But beyond that.. i'm running out out of ideas. Possible I would have to start with full bus isolate between the 536 and H5.. but thats another project to do based on a guess..

[exxos]

My second rev 5.62 board boots with EMUTOS and then fails on YAARTTT. It passes own address test, 0000 test and FFFF test , then fails on the same 0101 test , its odd.

IMG_4798.JPG

Seems like D8 is broken ?

1111111111111110 $FFFE
1111111011111110 $ FEFE

Not sure if that's really D8 on the 030 or D24 though ?

Either Way, it seems like the SDRAM must have died somehow, or some other fault somewhere. But always seems to fail at the same point. But if the ram was faltered and why would 0000 and FFFF pass in the first place :crazy:

[Badwolf]

Seems like D8 is broken ?

1111111111111110 $FFFE
1111111011111110 $ FEFE

Addresses are all even words, which would make it D24, I think.

BW