Update Sep 3
Experts on the integration have been in IHEP to improve the general status of the setup for the Summer Integration week that has been held from Aug 14 to Aug 25. Since Aug 27, L1 and L2 are separated waiting for the arrival of new L1.
Here you can find a summary of the activities divided by the topic:
1) Hardware:
One GEMROC has been replaced. Cause: missing reset cables. This has solved some 4 TIGER's missing TP. FEBs were not properly resetting with the other.
In the cabling operations, we have noticed that some of the cables were not properly closed, but just a few pins were attached. When properly connected, the signals in the data became more uniform.
4 FEBs will be replaced or checked during the following weeks
2) Communication errors:
New firmware version has largely improved the stability of the 8b/10b errors. A file to keep track of the improvements is being prepared.
Turning off/on the lights or using the vacuum door in the clear room induces a large noise in the clean room ground. Several solutions have been tested. At present, the lights MUST be always on.
3) Grounding: In order to understand the source of the interference in clean room we have worked on the ground of the system. As a general improvement, we have shortened the length of all the ground cables from GEMROC to DLVPC and from DLVPC to CAEN. Several configuration have been tested: - all together to CAEN port (standard) - detector ground to a "special ground": inside the clean room there is a copper braid that shall be connected to a pole in the ground. However we have verified that it is also connected to the electrical ground of the room. - detector ground to a very large metal table outside the clean room.
In all the different configurations we have always collected some cosmic data and counted the number of strips below 4 fC and below 5 fC. The best condition resulted to be the standard one, in which 84% of the strip is already below 5 fC and more than 70% is already below 4 fC noise. We have continued to acquire data in standard configuration.
4) Total charge:
After we replaced the gas bottle the cluster charge has been readout to standard values comparable with test beam data. The first hypothesis of an overpressure problem has been tested and excluded. The educated guess is that the quality of the gas was not enough. We will run new test with the separated layers.
In parallel, we have worked to the implement the ToT readout. A large effort was needed, since a new calibration dedicated has been implemented in the code: the new calibration consider the distance between the threhold and the baseline. The signal length has a very large dependence on the final charge output.
The signal duration had been placed under investigation after Torino people have simulated of the chip response using the simulated signals from Liangliang and Lia. Longer signals with respect to the design one have large impact on the total charge (due to ballistic effect). To verify the effect, an integration time scan has been performed. This parameter controls the time after which the charge is sampled in S&H mode. While the design value is 5, the data seems to favour to a larger value (namely 6). The difference is very small (~3%) for all the possible range of values (from 4 to 10): this could show that the signals have very different signal length. From a very brief investigation, this seems not to affect the time resolution performances.
5) Uniformity
Both on strips and clusters with low charge a set of recurrent holes have been observed in data. These are due to separation between two microsectors. This effect is well known in literature. Once the charge has been higher, the holes have disappeared.
The floating microsectors are well visible in data in terms of charge on the strips. We have observed a reduced amount of charge on those strips which face a floating microsectors: charge is not anymore distinguishible from the noise.
