I've used the attached "script", which is also in userapps/lsc/h1/scripts/DARM_transitions, to transition DARM control from the ETMX ESD to ITMX, leaving the upper stage control on EX.  

The attached screenshot shows the DARM OLG in the normal configuration (EX bias 130V, drivealign gain 184.65) and in the current configuration (IX bias -428V drivealign gain -108.4V). 

Robert did a ground injection while sweeping the EX bias.  

I then switched the ESD bias back to -125V, and started to switch the DARM control back to EX starting with the settings in 66751. This matched the DARM gain pretty well, with a discrepancy in the OLG measured at 200 Hz of .15%.   I didn't have time to do a broadband pcal measurement before we lost lock, but this should have kept the calibration OK. Interestingly, the increase in DARM noise seen in 66751 and 66814, as shown in the attached screenshot.  

We were in this configuration starting at around 22:45:30 UTC August 9th, until a lockloss at 22:49:24 UTC.   I'm not sure why we lost lock, we were sitting at -125V for 5 minutes for some quiet data when it happened.