View Full Version : Weird serial problem on OSI

February 23rd, 2014, 03:27 AM
Hi all,
I have a Ohio Scientific 4P that was working great about 3 months ago, but now I have trouble sending serial data from it. I need to get this working so I can continue to trouble shoot my disk drive.

Let me explain a little about the machine. It has an OSI backplane with 4 boards plugged in:

- 505b (CPU + Serial I/O + Disk I/O) (the RS232 serial voltages are non standard at 0V and 5V, but that was never an issue before. I have tested these voltages and they still look ok)
- 2 x memory boards
- A video board
* I also have a spare 505b CPU board, and a couple of spare ram boards too.

- The last two bits on the serial out data are nearly always stuck at ’00’. I have tested this at the ACIA TxDATA line and the serial out port (DB25) with the same results.
e.g. The expected output character ‘3’ or 00110011, but will always be sent as a ‘0’ or 00110000

- The only time the last 2 bits aren’t stuck is the moment I reset the machine using the ‘BREAK’ key. The very last character sent before the reset is always correct.

- The inputs to the ACIA (D0 - D7) seem to be correct as tested with a logic analyser.

- I have swapped the ACIA for another and the problem persists.

- I have even swapped the 505b board with my spare 505b board and the problem still exists.

- Otherwise the machine functions perfectly and receives serial data with no problem.

So this is weird because there is no other serial circuitry on any of the other boards (Video or RAM), and I have exchanged the entire 505b board with another with no change.

Just wondering if anyone has any wild guesses as to what this could be.



March 7th, 2014, 08:57 AM
Baud rate. Those cheap computers never had a crystal controlled serial port, one of the little tricks was to ALWAYS use two stop bits so the ACIAs could resync on every character. I am assuming you have a good solid earth between the two devices, that will cause problems you can only dream about.