During this week-end, I checked the whole acquisition system. and used some of the data of Friday. Now everything seems good from the programming point of view. A screen shot is given just below where you can see :
- a TOF between MCP (start)-Anode (Stop). Here, the typical temporal resolution is around 450 ps FWHM. I wrote a "self calibration program" which determines the center of the TOF (with gaussian model) for each channel. From this one could expect to achieve 300-200 ps in real experiment.
- the distribution of the hit on the DS16. The purple dots are for the regular lines and the red for the mirror image lines: A->B, C->D, E->F,G->H. The strong irregularities are mostly due to a to high discriminator threshold or too low gain on the MCP.
- the hit distribution in pixels underline this assumption. Since I cannot go higher than 2200V through the plate without noise 100Hz, it is hard so far to do anything better...
- the resulting image
As we can see, one main problem is that many DS16 channels are never answering. Thus this morning I checked the termniation of the DS16 (50 Ohm). Most of the channels are around 51.4+/- 0.3 Ohm, some at 49 and have been marked with a yellow tape. Unfortunatelly, the card A present a lot of channels with 80 Ohm. After a final test with the pulser, this card stoped to work !! Damned....
During the afternoon, I did some double checked the data of friday with little different voltage configuration:
- front MCP - 2100 V
- rear MCP +200 V
- anode +250->700 V
The background pressure is around 10^-7 mbar. When the gage is off the noise is slowly fluctuating between 20 Hz up to 300 Hz. I plot the efficiency of localization as a function of the voltage on the anode with the dark count.
| Anode | 250 | 300 | 350 | 400 | 450 | 500 | 550 |
|---|
| % | 2 | 5 | 9 | 11 | 14 | 17 | 20 |
| number pixels for 1e | 2 | 2 | 2 | 2 | 2 | 2 | 2 |
| number pixels for 2e | 0 | 2 | 0 | 2 | 2 | 2 | 2 |
I cannot use the gage on since it just localized all the hits in the center of the detector, and does not give an average effeciency. At 550V a second peak (wings) with 5 pixels start to appear, the time of flight is pretty symmetric a small step at short time of flight. At 600 V, the time of flight is perfectly symmetric, but the MCPs behaved very strange, up to 10 kHz noise, then went down to a few Hz...mmm… Back to 550V and I increased rear MCP.
| rear MCP(V)-anode (V) | 200-550 | 250-550 | 300-600 | 350-650 | 400-700 |
|---|
| micro A | 55 | 56 | 58 | 60 | 61 |
| % | 21 | 25 | 26 | 28 | 32 |
| number pixels for 1e | 2 | 2 | 2-3(smaller) | 2-3(equal) | 2-3(equal) |
| number pixels for 2e | 4 | 4 | 4 | 4 | 4 |
| number pixels for 2e | 0 | 6 | 0 | 0 | 0 |
At higher MCP voltage the picture got more homogeneous … For a strange reason the count rate is stable around 30 Hz, but the acquisition program is unstable ? memory leak … While I was checking the code, the signal of the oscilloscope has started to behave strangely much too large, and many mutli-hits. However, the count rate was stable somewhat a little lower: 10-20Hz. Since, I suspected something really abnormal, I checked everything and found that the current went up to 80-90 micro amp, again ! Why do these bloody MCPs not work normally, I cannot reach more than 2.25 kV through the plate ?
