After 2 months of beamtime, it seems that the MCPS are working. It is not so good, but at least at high bias current (round 2300 V) there is a rather homogeneous answer. It seems however, that one side is answer much less (DS C &D). The same diagnostic was made at the lab, with the dark count only. Of course, during the experiment other effect ought to be considered...

Anyway, after every baking out, the current has increased quite a bit. The new table is given here:

V	1000	1100	1200	1300	1400	1500	1600	1700	1800	1900	2000	2100	2200	2300	2400	2500	2600
micro A	32	35	38	41	44	48	51	54	57.5	60.5	63.5	67	70	73	77	80	84

The day of yesterday has been very good until, I decided to switched off the system. High dark count rate occured just then. Today, the background pressure is arround 4.10-e7 mbar (in 3 days), slightly better then yesterday. The polarisation scheme is as followed:

• front MCP : -1000 V -> -14000 V
• rear MCP : +1000 V
• anode : +1350 V

V	100	200	300	400	500	600	700	800	900	1000	1100	1200	1300	1400	1500	1600	1700	1800	1900	2000	2100	2200	2300
micro A	2	4	7	9.5	11.5	14	16	18	20.5	23	25.5	28	30	32	34.5	37	39	41.5	44	46.5

With respect to yesterday, I can notice that there is 1-2 micro A more above 500 V. Courant leak ?… A first pulse with a repetition rate of 1Hz is detected with 1600 V and the anode at 1300V. The pulse width is about 10 ns and the amplitude around 15-35 mV and is independant of the gage status (on / off). It repetition rate seems to be less with the votlage on the anode around 1100 V. A second pulse with a width <5 ns and and amplitude of 5-10 mV is present for a voltage of 1900 V. At 2000V the darkcount is around 30 Hz while yesterday it around a few Hz !

The system had a strange comportement during the morning and finally the reference channel (D16) died during lunch ! In the afternoon, something had changed, the pictures where much better and we could start to increase the DS16 threshold. The detector was still noisy around 500Hz (+/-300Hz) , and we let it run the whole night at 2100V.

The diagnostic from August (the 8th) have shown that there was very poor electrical contact between the feed-through and the anode. During the week (27th of August-> 1rst of september), we replaced the gold-wires glued with sivler epoxy (Epotek 21D) with regular SnPb weldings. After three iterations (including the super-sonic bath) the system was tested with RF successfully, and set up under vacuum. The HV test showed again clear limit with the feedthrough above 1100 V:

• The front side of the MCP can stand up to 2.6 kV; micro discharge is seen around 1.3 kV at 1Hz. The origin and location is unknow.
• The rear side of the MCP can stand up to 1.1 kV; above this there is a high load of the oscilloscope
• The anode can stand up to 1.4 kV; above this there is a regular "clac-clac" noise in the experimental room

The first attempt to use the MCPs like they were, was obviously not a good idea: high current , dicharge, and almost no nice signal. During, the cleaning of the ceramic holders 4 odd things were pointed out:

• the Cu ring where very much oxyde
• the Cu ring for the grid where at a floating potential,
• the electric holder for HV where very close to the MCPs, and potentially could induce short circuit
• the front MCP was "burned" on the side probably due to a discharge...

I swap the front and read MCP, remove the dusts with Helium, and try to remove the burned part on the MCP. Then wait 72 hours !