On 3/4/2018 7:16 AM, Michał Pleban wrote: > Hello! > > Jim Brain wrote: > >> My unit also fails on the last 3 tests (direct 1 succeeds). > It might be interesting to see what values you are getting in these tests? You are looking for $0f, and you get back $ff > > That is strange and rather unexpected. Can you post test results in this > case too? Actually, I had it backwards. $f works, but $e causes the ram test to fail. > These tests are meant to test whether the read back returns only lowest > 4 bits, i.e. if you write $FF, the real 6509 will read back $0F. You > might want to check your Verilog for that behaviour. I changed the code to send back $0 for the high nybble, with no change in results. I tested various options: 1) set high nybble of $0/$1 as $f on read 2) set high nybble of $0/$1 as $0 on read 3) send unused bank on $0/$1 read, with high nybble of $0/$1 as $f 4) send unused bank on $0/$1 read, with high nybble of $0/$1 as $0 With 1 and 2, ram test works, and I get 3 failures on your last test With 3/4 and unused bank as $f, I get failures on 2 of the 4 read tests (direct and indirect 1 pass, the others say $0f expected, $ff read. With 3/4 and unused bank as $e, ram test fails Interestingly, I switched to a 65C02, and the ram tests under all configurations, so I am going to work on that when the new PCB comes. Jim Message was sent through the cbm-hackers mailing listReceived on 2018-03-05 01:00:03
Archive generated by hypermail 2.2.0.