dml attempts a Raven

[agranlund]

Since it's a quick thing to do I would take out the ram simm in the off chance it's tainting the address or data bus -- you won't need any ram to get the first sign of life out the serial port.

Perhaps also try to verify that the address bus shows access to rom in the 0x04xxxxxx range.
And also the OE pin for the rom simm (which comes from pin51 on Nessi, RB3RD)

Nessi overlays rom at $0 for the first few cycles, and after that it should be running at $04000100 onward assuming it was able to read the PC from the rom.

Might be worth checking the IPL pins too. Interrupts should be disabled at this point during boot, but an unexcepted IRQ7 might cause issues at this stage, perhaps.

[dml]

Since it's a quick thing to do I would take out the ram simm in the off chance it's tainting the address or data bus -- you won't need any ram to get the first sign of life out the serial port.

Thanks - that is very useful to know.

Perhaps also try to verify that the address bus shows access to rom in the 0x04xxxxxx range.
And also the OE pin for the rom simm (which comes from pin51 on Nessi, RB3RD)

:thumbsup: will try....

Nessi overlays rom at $0 for the first few cycles, and after that it should be running at $04000100 onward assuming it was able to read the PC from the rom.

Nessi is the only thing that seems to be doing much of anything just now - although it's difficult to access the CPU from the top so I'll flip the board over for more tests soon.

Might be worth checking the IPL pins too. Interrupts should be disabled at this point during boot, but an unexcepted IRQ7 might cause issues at this stage, perhaps.

Ok thanks. Will update when I have been able to gather a bit more info.

[agranlund]

Was going to ask if you are testing with a non-EC cpu but you mentioned LC earlier so it won't be this but I'll mention it anyway:
Looking at the boot code I realize it will most likely silent crash before the first boot message when using an EC cpu.
This can, and should, be changed at some point to detect and print some message to serial before halting :)

[dml]

IPL1 drops LOW on reset and resumes HIGH.

IPL0/2 don't appear to ever change from HIGH.

Testing with RAM stick removed, no change in behaviour. Will leave it out for the time being, until I see something change.

Will have a look at the rom next.

[dml]

Was going to ask if you are testing with a non-EC cpu but you mentioned LC earlier so it won't be this but I'll mention it anyway:

I'm testing with an LC060RC50 - it's a little more expendable than the other one :P

[dml]

And also the OE pin for the rom simm (which comes from pin51 on Nessi, RB3RD)

Pin 51 sits high and there's a burst of convincing looking low activity shortly after a reset.

Pin 35 (A25 - ROM address high bit?) is high on RESET, drops low and stays there.

[agranlund]

IPL1 drops LOW on reset and resumes HIGH.
IPL0/2 don't appear to ever change from HIGH.

IPL behavior sounds correct.
Pin status is correct during reset when we're configuring CPU mode of operation, and then no interrupts happening after the reset period.

[dml]

Noticed something else now.

UART1:TXB from the 68752 is emitting clean bursts of data on reset and it is seen on UART1:TXB (pin 13?) on U305. I don't think that was happening before I messed with the RAM card.

I don't see anything on the listening terminal but it looks like something is being sent to serial, or trying.

There is also amplified data on pin 10 of U305, COM1:TXD

Huh.

And there are pulses on the COM1 physical connector...

I need to find loose DB connectors and make a serial cable. I don't trust any of this crap I have here farther than I can throw it.

[agranlund]

And also the OE pin for the rom simm (which comes from pin51 on Nessi, RB3RD)

Pin 51 sits high and there's a burst of convincing looking low activity shortly after a reset.

Pin 35 (A25 - ROM address high bit?) is high on RESET, drops low and stays there.

Sorry I was mistaken, rom slot lives at $40xxxxxx (A30 high)
A25 indicates it's accessing the last ram slot (the one closest to the cpu since we reversed the order of them recently) which does not make much sense to me at this point. That's not even the location which is initially overlayed by the rom.

Not seeing OE on the ROM at all would certainly be a reason for it to run garbage and crash itself. It's going to be interesting to know the root cause when it's all figured out and up and running (you may want to double check that OE for rom is indeed 51 on Nessi, perhaps I'm wrong about that too :) )

If it doesn't read the rom, for some reason, the initial PC will not be correct and perhaps that's why it's now in $02xxxxxx trying and failing to run floating-databus-data.

[SteveBagley]

I need to find loose DB connectors and make a serial cable. I don't trust any of this crap I have here farther than I can throw it.

If you can't find any and have a raspberry pi handy, you could use the 3.3v serial port on it's GPIO headers to see what is being produced.

Steve