Hangs on reading pilot checksum about openrabbit HOT 12 CLOSED

Can you capture serial communications and compare it to serial communications of RFU sending the cold loader and pilot?

from openrabbit.

I just checked that the hardware is okay: The same setup (same cable, same module, no prototyping board, lab power supply at 5V) works fine with the RFU from Dynamic C 9.62A on Windows 10.

from openrabbit.

I'll look into capturing the communication later: I'll be away from most of my equipment including the lab power supplies and the Windows machine for a week from today.

from openrabbit.

I wonder if this is due to some difference between the coldbload.bin and pilot.bin that OpenRabbit was used with at the time of the 0.1.1 release vs. the current ones.

AFAIK the author of Openrabbit 0.1.1 used it with pilot.bin and coldload.bin from Dynamic C 8.51. Are pilot.bin and coldload.bin from Dynamic C 8.51 available somewhere? Do we know anything about the differences to the current versions?

from openrabbit.

Testing with coldload.bin and pilot.bin from Dynamic C 8.61 I was able to proceed beyond this problem.
So this is indeed a version incompability.

from openrabbit.

I have documented the bootstrapping process from Dynamic C 9.62 in commit digidotcom/DCRabbit_9@4e527a9.

from openrabbit.

Turn out I was able to get beyond this exactly once so far. Now I run into #5 even when powering off the RCM2200 overnight.

from openrabbit.

With the timing fixes, it works reliably using the Dynamic C 8.61 binaries.
It also works with our coldload.bin (from Dynamic C 9.62), but not with the pilot.bin from Dynamic C 9.62.

from openrabbit.

My guess is that PILOT.BIN from DC 8.61 used a different protocol than the PILOT.BIN from DC 9.62.

from openrabbit.

Looks like it. Just using the pilot.bin from Dynamic C 9.62A fails due to different file format of pilot.bin.
But even when I change OpenRabbit to use the DC 9 pilot.bin file format, it still fails:

reset rabbit
loaded 831 bytes from ../coldboot/coldload.bin
flushing data for baudrate set
set baudrate to 2400
sending 828 initial loader triplets
flushing data for baudrate set
set baudrate to 57600
loaded 5517 bytes from /tmp/pilot.bin
sending 5517 pilot
csumR 0xd469 != csumU 0x0000

from openrabbit.

It works now. With pilot.bin from Dynamic C 9 (which is included in OpenRabbit).
After all it was just a file format issue.

from openrabbit.

OK, so now it sounds like the coldboot/pilot from Dynamic C 8.61 and 9.62 are interchangeable with the protocol used in OpenRabbit's RFU replacement? That's great. And you have a way to assemble coldboot.s with SDCC, and convert it to triplets for a replacement coldboot.bin. Next would be to update pilot.c and its dependencies to compile with SDCC (perhaps a bit more difficult, given Dynamic C's library format).

from openrabbit.