Hi there,
I'm sorry for the slow reply. Thank you, the 9905 schematic did help a bit, luckily the part of the circuit that was causing some problems was pretty much identical to the 9905.
Long story short, the problem is 99.999% solved. The remaining .001 is a nagging doubt. I've been soak-testing this thing for many hours and I've not seen the problem again since I removed, socketed and replaced a pair of ICs.
Firstly the 74LS04 hex inverter that was driving the troublesome lines was removed and tested on a breadboard and found to be sort-of working (the output levels looked a bit iffy but the response time and noise levels were in spec).
So I binned it and put in a fresh one. The problem immediately improved by 75%.
Secondly a 74LS75 bistable latch was swapped out in the same way and now the problem seems to be solved.
My home-brew data logger custom made for this instrument works nicely now that the BCD signals are not mysteriously sticky any more.
If anyone's interested, it's an Atmel ATTiny84 with an MCP23S17 SPI bus expander (which is super useful because it has built-in configurable inverters on it's inputs).
Data is output over a simple 5V, 9600 baud "8N1" serial connection that my PC can read easily. Power for the logger is taken from the 5V output on the Racal's output port.
In the photo here it's the small black box on top of the Racal. The serial port data output is shown on the right side.
" alt="" class="bbc_img" />