40m QIL Cryo_Lab CTN SUS_Lab TCS_Lab OMC_Lab CRIME_Lab FEA ENG_Labs OptContFac Mariner WBEEShop
  PSL  Not logged in ELOG logo
Message ID: 1705     Entry time: Sat Aug 13 20:41:34 2016
Author: awade 
Type: DailyProgress 
Category: Other 
Subject: South cavity error signal troubleshooting and cavity alignment 

The south path cavity was realigned and I fixed the control box error signal monitor issues.

The south path alignment had drifted over time and in the course of aligning the north cavity I had touched the south periscope knobs in error.  After realignment the reflection signal dip on resonance was V_min,V_max = (304.0,80.0) mV on reflection => Visibility of 58%. This was about the same as before*. 

In the past we found that that there was no error signal coming out of the mixer monitor BNC on the front panel box.  In the course of checking the power supply configuration, before moving it to the rack on the north side of the table, I noticed that the common banana plug pins on the +/-17 V and +/-24 V were very wobbly.  In fact they were barely making contact.  Banana plugs are terrible. This can be a big problem in general if the 24V lines disconnect as HV amps can often fiz without their lower voltage supplies connected. 

Also, none of the power supplies were grounded to a common ground, which might open us up to ground loop noise, depending on the details of the grounding arraignments inside the FSS boxes themselves.  We also don't want them floating all over the place at low frequencies as this may have unintended consequences for noise. I have no experience of this being a problem as I have always just grounded things properly, better to do it and be safe than sorry.

On another note, one of the HV power supplies (the +ve one) was grounded with what turned out to be a piece of solder linking the front panel pins: bad.

HV amp grounded with piece of solder

I have now removed a lot of the thin wires grounding the common pins on each power supply replaced them with thick wire (better to go with super thick wire) and grounded the front panels together using brand new solid banana plug patch cables (good to many amps).  I snipped off the old green banana plugs a soldered on new ones and added two diameters of heat shrink to provide some strain relieve for the joint.  Still don't think this is good enough, but it will do for now.

After doing all this and powering back up I saw a decent error signal at the error signal monitor port. There might have been an issue with one of the low voltage supplies not having been referenced properly to a common ground. We now see an error signal of order 744 mV peak to peak.  

Also, I got a delayer box and checked the phase delay of the LO relative to the reflection PD RF signal to see if we had the best possible error signal. In the south path the demodulation quadrature was correct and no cable changes were needed.  I didn't get to locking or measuring a transfer function, but this should be done to characterize performance and give us an idea of what to optimize next. All of this is leading to a better quality beat note and is good to have on record for future design iterations.

- awade Sat Aug 13 20:41:28 2016

*Note I have used a different methadology FringVis = (V_max-Vmin)/(V_max+V_min), this should give the actual contrast rather than ratio of peak to total power.

ELOG V3.1.3-