Had another session wwith the FRK Rubidium. Unfortunatly it does not quite match either of the two available manuals. Still a close manual is better than none.
Thinking about it while studying the manuals I relaised the lamp voltage was completely out of range at 2.5V. Now the "lamp voltage" is not the voltage across the lamp, its derived from the DC component of the photo diode output (without the 127Hz lock-in lamp modulation). So as the lamp is lit I thought I'd start with that as it's fairly simple, just a ua739 dual op-amp. The lamp voltage output is resistor (R7) isolated from pin one (output) of the 739. The boards have thick conformal coating so even with sharp probes it takes a bit of pressure to get contact. So checked all supplies, all OK. Went to pin 1 of 739 and it's at 2.5V as expected but poor contact so I press harder and the voltage rises, but harder and it's 10V and the unit locks
So the board is hard to work on the rear of because the are wires soldered to all four sides and some were routed under components before being soldered to turret tags and conformally coated.
So I disconnect 3 sides lift the board up and sart doing continuity checks. There is something odd between the photodiode connetion tag and pin 5 of the 739. I can get between 1 ohm and 200 ohms depending on pressure applied. All that is between the two points is a 2.2uH choke in a minature moulded case with axial leads. All joints look OK. Solder a wire to turrent and pin 5 connect to DMM on ohmes and start poking about. Turns out to be the turret tag where it's swaged and soldered to the PCB.
Still looks ok under the coating. Clean the coting off with some acetone and the reitance is now 1 ohm! Under high magnification and the right light you can just see the crack.
Soldered up, put board back down and connected and the lamp voltage is 10V and the unit locks up in 8 minutes even without insulation or cover
I've even found some suitable insulation. Med will approve, it's blue.