Last summer(1995) the positive ion tails were measured for the P0 test chamber using a 30 nsec peaking 4 pole semigaussian amplifier. The results are described in detail in the proceedings of the Lake Tahoe CMS week Sept 1995. Twenty pulses were collected using a Techtronix Digital scope. It was found that the data fit the expected t0/(t+t0) distribution poorly, but the positive ion drift seem to be well described by the expression:
f(t) = (1-xa-xb)*exp(-l1*t)+xa*exp(-l2*t)+xb*exp(-l3*t)where xa=.210, xb=.054, l1=.046,l2=.013,and l3=.0036.
On April 16 new data was collected on the T1 prototype. The gas in the chamber was 50% (C02) + 30% (Ar) + 20% (CF4). The data were taken at a voltage of 4.2 KV using the 50um wire. 4-pole semigaussian amplifiers with a peaking time of 30,50,100,300 nsec were used. The amplitude of the signals from the 30 nsec amplifier were so small as to be useless. The 300 nsec amplifier was behaving poorly with large low frequency noise components. The analysis described here uses only 50 nsec and 100 nsec amplifiers.
The data was fit both to a t0/(t+t0) distribution and the same three component exponential listed above. This was convoluted with the delta function response of the amplifier, once again measured using a Techtronics digital scope. To account for the electron drift arrival times the charge was assumed to arrive in 5 descrete bunches. The arrival time for these bunches was included in parameters for the fit.
In contrast to the P0 prototype data the T1 data showed good agreement with a t0/(t+t0) distribution below drifts of 1000 nsec. For greater than than 1000 nsec the data fell off faster than a t0/(t+t0) distribution. The t0/(t0+t) fit pulse distributions are shown in the following ps file: 1/t tailfits . Files 41-83 correspond to the 100 nsec peaking amplifier while 84-120 correspond to 50 nsec peaking amplifier.
The data has been fit to a universal 3 exponential curve. These fit pulse distributions are show in the following ps file: exponential tailfits . Files 41-83 correspond to the 100 nsec peaking amplifier while 84-120 correspond to 50 nsec peaking amplifier. This curve is seen to fit almost all of the data very well. The parameters are:
f(t) = (1-xa-xb)*exp(-l1*t)+xa*exp(-l2*t)+xb*exp(-l3*t)where xa=.15, xb=.035, l1=.1,l2=.015,and l3=.002.
Shown below are the universal curves extracted for the P0 and T1
chambers. For comparison a 2.5/(t+2.5) distributions is overlayed
on these universal curves. It is seen that T1 agrees well with the
2.5/(t+2.5) distribution for ion drifts below about 1000 nsec.
The P0 data is substantially different. This difference is not
presently understood.
