Last modified: 7-OCT-1999
Relative channel gains are estimated by integrating the net signal from pulsing the ground wires or injecting charge into an FEE channel. Both techniques suffer from variation of the ratio of test pulse coupling to real signal coupling between channels (sector edge effects, variations in coupling capacitance, variations in time response).
Absolute channel gain, including gas gain, can be determined with cosmic ray signals at known TPC gas pressure.
It has been generally assumed that the gain for a channel is independent of time bucket and time.
Note that the electronics people (e.g., Chinh Vu) usually refer to gain in terms of the peak signal, while others of us refer to the integrated signal; the proportionality of these two measures can vary between channels and with the shape of the input signal.
FEE gains are trimmed, somewhat, so that the measured gain distribution has an rms of about 7%. However, the tails contain some interesting systematics. Preliminary results with a subset of the production found a small systematic variation of gain according to chip on the FEE boards, and an even smaller systematic variation according to chip channel.
Typically, gain measurements have used 64 pedestal-subtracted events; this is probably excessive. One would like to determine the number of events actually needed.
SAS shaping times vary from channel to channel, so that there is not a fixed ratio between the gains based on peak pulseheight by Chinh Vu and those based on integrated signal that will be used for analysis. This variation and its systematics should be determined.
For 20 pre-production cards, measured gains were stable within about 1%; Fabrice Retiere may have values for the production cards.
Some study done for pre-production cards, but not studied for current cards.
At least three measurements of relative gain are available, and one measurement of absolute gain. It turns out that, for the Summer 1999 data, termination conditions at the unpowered MWPC's allowed variations in the anode-wire grounding conditions between rows, and, hence, in the ground-wire pulser gains
FEE charge injection is not guaranteed to be more accurate than 20% (because of variations in the integrated-circuit coupling capacitor), but it is believed to be more accurate than this, and we should measure this.
The energy-loss straggling distribution should be measured and compared to model predictions (Geant=PAI, cluster model).
Gains, as estimated by ground-wire pulsing, show strong systematics on the inner subsector; we need to understand the systematics for both the ground-wire pulsing and for the usual signals.
Gains, as estimated by ground-wire pulsing, show some systematics on the outer subsector; we need to understand the systematics for both the ground-wire pulsing and for the usual signals.