MEGA ST2 - "random" characters on the screen

[DominoTree]

During boot, I get a bunch of unexpected characters on the screen as text is displayed. I also see similar behavior with running YAART from the cartridge port, although YAART itself runs through multiple cycles with no errors.

That sounds like pullup problems, id check all the address and databus on CPU to make sure you have 2.2K on them all...

Thanks, went ahead and socketed the CPU first, all data and address lines tested at the socket have 2.2k to VCC at the socket so that all seems good.

Also went ahead and swapped in another known good M68k while I was at it and it didn't change the behavior.

Are there any particular signals that I should look at with a scope or logic analyzer that might be culprits here?

I should have a keyboard/mouse rigged up today as well and I was thinking running GEMBench might provide some interesting data as well.

[exxos]

I would check the CPU pins which go to the MMU & GLUE , MEGA ST's seem to suffer from loads of broken vias all over the place.

[DominoTree]

Plugging in a keyboard seems to have resolved it. Is that normal? Floating ground somewhere?

[DominoTree]

Ran through GEMBench just fine too - numbers compared to my 1040STf - curious why VDI is slower but maybe this is just because it's a MEGA

st.jpg (422.3 KiB) Viewed 2885 times

[exxos]

Plugging in a keyboard seems to have resolved it. Is that normal? Floating ground somewhere?

So it works with the keyboard plugged in, so if you un-plug it now, does it go nuts again ?

Those scores look odd.. looks like the blitter is running slow for some reason

EDIT:

If not done already.. try changing RP1 to 2.2K as well....

[DominoTree]

If not done already.. try changing RP1 to 2.2K as well....

So after a few more boots, it still does it about 50% of the time with the keyboard plugged in, and 100% of the time without.

I’ll give swapping RP1 a try and go over all of the related keyboard connections one more time and make sure all of the solder joints are good

[exxos]

Maybe slap 4700uf on the 5v rail near the MMU see if it helps or not..

[DominoTree]

Ha, so one mystery figured out - I was running GEMBench with SpeedoGDOS loaded - everything is at 100% now.

Tried swapping RP1 with 2.2k but didn't see any different behavior when the keyboard is unplugged.

You may be on to something with the capacitor idea (I'll have to dig one up) - I've noticed on some boots when it acts like the keyboard isn't there even though it's plugged in (and I hear the "stuck key" sound and get random characters) simply touching the contacts of what I've rigged up is enough to get it to settle it down and behave and everything seems stable afterwards.

[DominoTree]

Of course, that *could* be a very touchy short as well, but it still doesn't explain the random characters thing

[exxos]

Sounds like a bad connection... But even so, they keyboard shouldnt cause screen corruption.. That doesn't make any sense ... Might be worth tracing the IRQ lines from the ACIA chips to the mfp etc... Maybe even E clock from the CPU... If the acia is selected by some random means, it could corrupt the bus.. But I wouldn't think the thing would even try to boot if that was the case..

I guess you could lift the enable pin ( e clock ) from the acia , can't remember if its high or low to disable it, then see if the corruption goes away or not...