Hi Guys
I thought I’d share my findings of the CD32 Error Rate Counter boards
used for testing Amiga CD32 units at West Chester production, and maybe update as I go.
CD Error Rate Counter boards:
I purchased the two boards individually from eBay toward the beginning of the year.
I’ve only powered them to check nothing warmed up, and all was ok.
They have a 7805 regulator, and were probably powered from a 12 Volt bench supply.
From there, investigation of the 26 pin header connector suggested strongly
that it’s actually a 25 pin parallel port connected to the test station computer.
When a cable was made to connect a 26 way IDC header connector to a ribbon
25 way D connector with pin 1 aligned with the connector at each end,
all pins aligned for a parallel port with all grounds connected and common to each other,
data bus D0-D8 all connected, and also control pins for BUSY and Select are connected.
The 26th pin of the header connector is not connected to anything on the PCB,
and is simply cut off as it doesn’t fit in a 25 way D ribbon connector.
Commodore probably just pulled the 26th wire off the ribbon cable completely.
25 pin D Parallel cable:
25 pin PC/Amiga Parallel Port (for reference) :
https://upload.wikimedia.org/wikiped...inouts.svg.png
Over to the input header connector, it looks to connect the error status monitor
output of the Sony CXD2500BQ controller chip on the CD drive controller board.
The MNTR test points can be seen toward the top right of the CD controller board.
The CD drive unit is Chinon brand, but uses several Sony chips.
Chinon CD Drive Controller Board (Amiga CD32) :
Investigation of the data sheet for Sony CXD2500BQ suggests manufacturers
connect some equipment to this 4 bit port to monitor error status output.
The text also seems to suggest that the value of all 1s (decimal 7) on the
4 bit port is the only uncorrectable error, as a failed C1 correction is passed
on to C2 error correction, and if that fails, the error is uncorrectable.
Sony CXD2500BQ (CD Drive Controller) Monitor output Error Correction Table:
It occurs to me quite possible that the error counter boards logical AND all
four bits of the port together, only looking for the value b1111, and count that.
Though I haven’t looked too hard at the logic on the board yet, it only has three
pulse extender hardware chips, which seems consistent with use of this port,
as the bit MNTR3 is always at least twice as long as the other three bits.
The RFCLK line referenced on the timing chart may not be on the same test port,
but is available on a Sony CXD2500BQ chip if it is needed for the port timing.
It could be the lone test point off to the right a little from the MNTR port.
Pin 5 of the 5 pin input port to the Error Counter board is Ground,
so I don’ think it is used at all, lending to the theory that the 4 bits are logical ANDed.
Sony CXD2500BQ Monitor Bus Timing:
Now I think the counter data is encoded for the 8 bit parallel port output,
and the BUSY and Select pins used perhaps for latching and resetting the
counters on the PCBs. Probably the case that’s it’s simply a frequency counter,
but yet to be determined.
It seems fitting to write the software for an Amiga, but to get it done faster,
sticking with what I know, it would probably be smoother for me to get it done
with a dedicated microcontroller board, and maybe over to an Amiga from there.
Cheers, Art.