Reports until 16:01, Tuesday 29 August 2017
H1 SUS (CAL, DetChar, ISC, SUS, SYS, VE)
jeffrey.kissel@LIGO.ORG - posted 16:01, Tuesday 29 August 2017 - last comment - 17:31, Tuesday 29 August 2017(38430)
First Cycle of ETMX TMDS Complete
J. Kissel, N. Mavalvala, K. Ryan, D. Sigg, J. Worden

Executive / Expert Summary:
We've completed our first cycle of the Test Mass Discharge System. 

We ran into a little confusion when we first started with the TMDS valved into the chamber (around 13:30 PDT) because our clean air flow rate was too low (only ~30 [L/min] or "slm"). With that input rate, our electrometer readings of the ion flow were only ~+/-3 [V], and the pressure inside the TMDS (as measured by the Baratron) only read ~18 [Torr]. However, our patience drove us to investigate when the BSC chamber pressure only was increasing at a rate of [0.1 Torr/ 10 min]. 
After increasing the clean are input flow rate to ~70 [L/min], we were able to reliably put in +/- 7.5 [V] (we were not able to get as high as the design value of +/- 13 [V]), and pressure in the TMDS came up to near the desired value, around 95 [Torr] (desired value is 120 [Torr]). 
However, even with ~70 [L/min] of ionized gas going into the chamber -- it took us twice as long (about an hour) to reach the desired BSC chamber pressure of 30 [Torr] -- because we have two BSCs worth of volume (BSC9 and BSC5) (obvious in retrospect!). 

After completing the first cycle (by 14:30 PDT -- note, twice as long, LLO only took ~1/2 hour to get that high), we're now pumping on the two-chamber volume with the QDP80, but it'll likely also take twice as long. We're going to continue to monitor the pressure into the evening, with the hope to get another cycle in tonight. If not -- two cycles tomorrow, one starting first thing in the morning the other mid-afternoon.

Right now we're suggesting that gate valves will be open first thing Thursday (but we may get there faster).

Details
The documentation that proved to be most helpful:
T1500057: -v2 of Rai's procedure for operating the TMDS, and associated documentation uploaded from S. McCormick and R. DeRosa's using of the TMDS at LLO.
E1500252: Video of Calum and Ben running the TMDS system (valved out of chamber)
T1400713: "Final" Design Documentation, Figure 1.

Changes in the system since the above documentation:
- The electrometer and ion generator are all powered from the TMDS interface, D1500152, and connected to the electrometer and ion generator as shown in LHO aLOG 38358
- Not included in the "final" design Figure 1, but included in Rai's hand-drawn diagram in T1500057 is a temperature controlled heater that's re-heating the clean input air to room temperature after being cooled and filtered by the cold trap. With the input clean air as low as we started out, at 30 [L/min], we did not need to use the heater. However, once we increased the flow rate to 70 [L/min], 

Future User Guidance (as gleened from documentation and use of the system by Daniel and Nergis):
- Step 12 of T1500057 is the most exciting part, and the paragraph is pretty loaded. One of the more important sentences is "Reduce the amplitude of the square wave [fed to the electrometer] to 2 Vpp, and leave this value for most of the filling. Occasionally raise the square wave voltage to 25 Vpp and note the +/- ion current."

    What Rai's elluding to: The electrometer is measuring the ion current (converted into voltage by a resistor). However, the higher the amplitude of square wave, the more ions are being sucked into the electrometer to be measured. That means less ions are getting into the chamber. So, one should
    - Before the TMDS is valved into the chamber, constantly "measure" the ion current (once at the desired clean air input flow rate) with a high square wave amplitude (say the max, 25 Vpp), while turning up the HV VARIAC knob, watching the HV readback of the ion generator to ensure you're just at the border line of major distortion of the HV waveform (some small amount of distortion is OK -- we want the most amount of ions going into the chamber as possible). The ion current (voltage) from the electrometer should show *roughly* equal +/- voltage. Both Ryan's results (see LLO aLOG 35636), and our experience shows that one cannot get a perfect balance; a deviation from perfect of ~0.5-1 [V] / ~10 [V] = 5-10% is OK.
    - Just before valving into the chamber, turn down the square wave amplitude to something low -- 2 +/- 0.5 [V] is fine, no need to be more precise. Once you've turned down the square wave amplitude, the measurement on the electrometer degrades, so you'll see a reduction of ion current (voltage), and likely the imbalance of +/- will increase or change. This is OK, you're just trying to back the electrometer out of the way, and let all the charged ions go into the chamber.
    - Now Rai's second sentence, and Ryan's data makes more sense. Every once and a while (every 10-15 minutes, or at your favorite increment of pressure), bring the amplitude of the square wave back up, and measure the ion current (voltage). This is just a double check, to make sure that your ion current is still balanced, and you're still sending in (roughly) equal number of positive and negative ions

- Before doing anything, Kyle had turned off all high-voltage power supplies to anything in-and-around the chamber. Good. However, as a triple layer of precaution, we should also make sure that we're not requesting any digital high-voltage. As such, in the future, operators should 
    - Bring the QUAD's ESD Bias Voltage to Zero
    - Reduced the requested Ring Heater power to Zero
    - Turn off any suspension alignment offsets (though it's nice to leave the suspensions damped -- don't forget about the TransMon / TMTS)
    - Turn off / disable the in-vac pico-motor control
    - Make sure that the transmon pico motors are disabled.
    - 4 Piezo controllers are on the ISCT EX / EY for the green WFS centering, and centering the green input onto the TransMon QPDs. Since these are external to the chamber, you don't *have* to turn them off, but we don't need them, so might as well.
Comments related to this report
jeffrey.kissel@LIGO.ORG - 16:18, Tuesday 29 August 2017 (38432)
Attached are the raw notes / timeline from today's work -- attempting to capture both vacuum information (i.e. like Scott's log) and the electrometer information (i.e. like Ryan's log) -- but because of the confusion with pump speed and electrometer "on/off" during the measurement, we only got ~2 readings of the ion "current" (voltage), so they're not really that useful.

We'll take more useful / sensible measurements tomorrow now that we understand the system better.
Non-image files attached to this comment
kyle.ryan@LIGO.ORG - 17:25, Tuesday 29 August 2017 (38436)
The ESD and Ring Heater power supplies were de-energized and the Picomotor drives were disabled prior to the admission of gas.  This was done during my preparations for this partial "vent" but was not conveyed to the TMDS folks.  Also attached is a picture of the spring table support used to temporarily mount the Surface Discharge Ionizer to BSC9.  This "contraption", along with the Ionizer. gets removed after the discharge exercise.  
Non-image files attached to this comment
jeffrey.kissel@LIGO.ORG - 17:31, Tuesday 29 August 2017 (38437)
It looks like it won't be until ~19:20p local time (PDT) before we hit the desired BSC chamber pressure of ~0.1 [Torr], so we've deciding to punt on a second cycle for the evening. 
(For the record, that's 17:30 - 14:30 PDT ~= 5 hour pumpdown time from ~30 [Torr] to 0.1 [Torr].)

The game plan is now to perform a second cycle starting at 07:00a local time (PDT) tomorrow, and another after mid day. Still hope to open EX gate valve to the arms and begin IFO recovery Thursday morning.
Images attached to this comment