5:30: Gustavo deploys the arm, we have pbl starting AdSecSx: it periodically faild while loading the DSP program. We try many times to power cicle. LB finally can set the shell (how?).

We close the loop on LGSW only. Hexapod centered:

SX HEXA

3.57

0.66

-0.82

147

-202

Test of telescopeOffset

We try to send a telescopeOffset command as explained in Lorenzo's email of 11th May:

argos.arbitrator.execute("""self._runner._foreman._pointingCtrl.offsetTelescope(dx, dy)""")

The first try was on 14:06:59 UT. The command is properly sent from AARB to AOS and PCS GUI displays wired values:

We attempt to send a new command at 14:33:21. The output of the test was explained in the following email from Jose to Luca and TCS folks. A issuetrack was created with the number #5978.

Today we have exercised the AOS offsetTelescope command from Argos and
in principle seems to work fine. However, I would like to share with you
some issues we had.

After restarting AOS or sending this command without a valid preset, the
AOS hung and PCS got invalid numbers.

in AOS.cpp, in repoint():

 749     double px=(double)aos.side[_side].cmds.ao.presetao.rocoordx;
 750     double py=(double)aos.side[_side].cmds.ao.presetao.rocoordy;

px and py are NAN. The AOS calculates the offsets (x and y) and it is
sending this to PCS, PCS gets bad numbers, and AOS is not responding any
more.

So, wouldn't be a good idea that AOS returns an error if the above
variables were never set ? (i.e. no preset or AOS restarted).

Additionally, we also noticed that PCS accepted numbers that were not valid.

CL test of LGSW + PYR BIN4

We check that the March RECs are still valid: we close on LGSW only, no disturb. We can rise the AO gains up to 0.5; 0.5; 0.4 before the system start to oscillate.

TN	AO Gain	CL	TT Res. WF [nm rms]	HO Res. WF [nm rms]	AO rec	DIMM ["]	JCL	PCL	Notes
20160512_140034	0.2,0.1,0.05	T	231.136	30.8062	20151212_220700	-1	F	F
20160512_145503	0.5,0.3,0.15	T	152.432	30.3344	20151212_220700	-1	F	F
20160512_145524	0.5,0.3,0.15	T	148.742	30.1465	20151212_220700	-1	F	F
20160512_145557	0.5,0.5,0.15	T	147.74	37.1656	20151212_220700	-1	F	F
20160512_145609	0.5,0.5,0.15	T	150.795	36.9825	20151212_220700	-1	F	F
20160512_145652	0.5,0.5,0.4	T	146.864	41.0891	20151212_220700	-1	F	F
20160512_145703	0.5,0.5,0.4	T	138.407	39.9885	20151212_220700	-1	F	F

Then we load 1" disturb and we check again the best gains:

TN	AO Gain	CL	TT Res. WF [nm rms]	HO Res. WF [nm rms]	AO rec	DIMM ["]	JCL	PCL	Notes
20160512_150028	0.5,0.5,0.4	T	147.846	62.139	20151212_220700	-1	F	F
20160512_150040	0.5,0.5,0.4	T	148.124	61.7845	20151212_220700	-1	F	F
20160512_150149	0.5,0.5,0.4	T	154.029	63.5662	20151212_220700	-1	F	F
20160512_150231	0.5,0.5,0.5	T	161.024	68.4223	20151212_220700	-1	F	F
20160512_150245	0.5,0.5,0.5	T	156.868	65.7741	20151212_220700	-1	F	F
20160512_150258	0.5,0.5,0.5	T	171.148	73.594	20151212_220700	-1	F	F
20160512_150449	0.5,0.4,0.3	T	144.228	68.2542	20151212_220700	-1	F	F
20160512_150502	0.5,0.4,0.3	T	152.604	68.4732	20151212_220700	-1	F	F
20160512_150512	0.5,0.4,0.3	T	149.682	68.3466	20151212_220700	-1	F	F

With High Order gain= 0.5 the loop start to oscillate. HOgain=0.3 is too low.

TN	AO HO Gain	HO residual [nm rms]
20160512_150149	0.4	41
20160512_150245	0.5	43
20160512_150449	0.3	50

Then we switched to LGSW + PYR, mag 10, no disturb.

TN	AO Gain	CL	TT Res. WF [nm rms]	HO Res. WF [nm rms]	AO rec	DIMM ["]	JCL	PCL	Notes
20160512_151338	0.5,0.5,0.4	T	87.0356	42.0655	20151212_233400	-1	T	F
20160512_151419	0.5,0.5,0.4	T	79.6252	41.9633	20151212_233400	-1	T	F
20160512_151432	0.5,0.5,0.4	T	83.6337	42.3457	20151212_233400	-1	T	F
20160512_151521	1,0.5,0.4	T	71.2101	44.3984	20151212_233400	-1	T	F
20160512_151537	1,0.5,0.4	T	72.8999	51.9782	20151212_233400	-1	T	F
20160512_151616	1.2,0.5,0.4	T	259.378	281.778	20151212_233400	-1	T	F
20160512_224722	0.2,0.1,0.05	T	55.6389	28.4444	20151212_233600	0	T	F

We also give a try to the autogain procedure. We swap different # of iterations and # of repetitions. The best result in DT we get it with low # of rep and higher # of iterations (4 and 3000 respectively, wrt 10 and 100 default).

9:00: we start to shutdown the HW.

9:45: Power off of the telescope.

14:00: cooling is back, people starts to switch on again HW.

15:30: we have back the telescope, we deploy the cal unit swing arm.

We get back to LGSW + PYR BIN4 tests we set mag 15, PyrWfs set to BIn4, 200Hz. The PyrWfs is plenty of background light.

BcuNumber2 installation on LgswUnitSx

15:50: SX BCU dies after 2min of CL operation, we reset and try again. It dies 2 more times, few minutes after the loop is closed. We decide to swap the SX BCU with the spare one.

The spare BCU has a known pbl with the second HVC that fails to configure. We give a first try: after more than 5 consecutive failure in configuring it we decide to swap the HVC from BcuNumber2 (spare BCU) with HVC of BcuNumber1 (old SX BCU).

• We need to replace the HVC with a working one: contact Mario and ask for time and cost of a new HVC module.
• We must contact MicroGate also to schedule the BCUs power supply upgrade. All 3 BCU will be shipped back to EU at the end of the run.

17:00: Jose modified the oscilloscope SW to retireve the peak-to-peak and average voltages of the 4 channels. This was necessary to monitor the BCU power supply voltage read by channel 4.

• The property crawler must be updated and entries in Zabbix must be created

17:25: we give a try to the spare BCU w/ HVC we enable the PyrWfs slopes even w/o light. We also enable the BCU sloping to AdSec. The system seem to work fine for about 20min then the BCU stucks.

• JZ will trim the BcuNumber1 and BcuNumber2 voltages to 3.5V (or even more?) as fisrt task tomorrow.

-- TommasoMazzoni - 12 May 2016