LO became unlocked at different Az positions during r1610. Randall inspected the twister cable and found one had a kink. It also made a grinding noise if flexed. The problem was just below a hoop in the twister. That length of cable was suspected to be Heliax. It was the 100MHz cable to the LO. That length has been temporarily replaced with a piece of LMR400 that was also in the loom. The loss of LO lock problem seems to have gone away and the R1610YG experiment has continued. We should watch to see if LMR400 produces any more phase noise compared to the Heliax.
Randall confirms replacement cable is of same type as the original : LMR400. Photo of damaged cable attached.