68030 /BERR

[techknight]

is it possible the Sonora is bad? woudl take a donor board to figure out though. 

Is it possible to invoke the 68K trace and see where its hanging up? which opcodes, etc. 

Im going to go with my gut and say its some sort of initialization fault. 

 

[mmx01]

Slow progress due parts and covid shipping delays... got myself two boards in the mean time so have now access to some parts. This project is now driving few other projects and got out of proportions

> Still waiting for more LA probes to be able to handle 64+ channels and do address/data decoding

> Got inverse assembler software for TLA714 and started building adapter it requires to run TMS203

> Got addons for my hot air station to heat only IC pins, not etire chips to avoid heat damage. Gives more confidence with touching these oldies - so far successfully

> As expected RAMDAC is just a stupid chip, with it removed good board chimes as normal. Chip from bad board produces image in the working one

> I am about to do same trick to SCSI controller

Next is to get TMS205 working and decode what the board is doing. With CPU, buffers and RAM + VRAM replaced not much options I see but still curious to see if it fails at RAM test as part of rom boot or elswhere. Provided it fails with RAM this would mean indeed Sonora doing mem mgmt. is bad.

 

[mmx01]

Interesting, with SCSI chip removed (U8) good board chimes good as well. So neither RAMDAC nor SCSI are really queried for chime of death. Removal of on board VRAM or RAM results in chime being distorted.

Soldered both RAMDAC and SCSI from bad board into a good one and both are okay.

In the mean time fixed old 10Base2 Ethernet card from Asante which had -9v inverter faulty resulting in loopback transmission test failure and no connectivity. Newer ones have DC-109 which is obvious but of course mine was custom padded with no markings. Not repairable but with minor power supply wiring tweak a modern one with full isolation works like a charm: RO1209S

 

[mmx01]

Finally some progress! Lifted, cleaned up and re-soldered QFP 208 fine pitch Sonora chip. Not funny and impossible without a microscope as shorts are tiny even when soldering with soldering paste.

Chime of death is gone and I was able to boot MacOS, keyboard and mouse are fully functional. Another wire was needed for SCSI controller next to the corroded battery holder.

Still some work needed as colours are distorted but otherwise the board is pretty functional. Seems like on an LCIII board chime of death is nothing related to RAMDAC, SCSI or PRAM. When chime is distorted there are issues on the bus related to VRAM/RAM contacts/presence. Low level ROM boot test appears to be limited to BUS/RAM/MMU as lack of other ICs don't change behavior.