Some results from CSC Preamp Prototype tests
The setup
Eight prototype preamp channels are connected to the station 2 prototype
which is in a cosmic-ray test stand which provides a trigger when
two scintillators (one above the prototype chamber and one below the
prototype chamber) fire in coincidence. This trigger is sent to the
test board which holds the prototype preamplifier and N samples prior
to the trigger and M samples after the trigger are digitized and sent
to our cosmic-ray test stand readout system.
Some Events (Gee Whiz)
Shown in the plots
below are ADC values versus sample number (samples are taken once
every 100 nsec) for a set of contiguous strips in a couple of
sample events. In the first file
you can see 2 strips with non-saturated but also non-pedestal values
followed by a strip which reaches saturation followed by another
non-zero/non-saturated strip, followed by a strip which basically
has a pedestal value. The second plot shows a smaller
number of ADC samples, and three strips with the first two showing
charge and the third one showing a pedestal value. (Pedestal values
are ~channel 3500 and saturation is ~channel 0.)
25 ADC samples from five cathode strips in an event
Another event with fewer samples
Pedestal (noise) measurements
The first plot below shows a pedstal and ADC spectrum
for channel 2 from the preamplifier. As can be seen, the pedestal
is somewhat wide (~12 channels sigma). Note, again, the pedestal
is at the high end of the spectrum and large cathode charge is at
the bottom end of the spectrum
Pedestal and ADC spectrum from Channel 2
Since we know that common-mode noise is a large portion of the
noise that is currently seen in the preamplifier, I try below
to separate the common-mode noise from the "white" noise. The
first plot shows the pedestal of channel 2 versus the
pedestal of channel 5 and you can see a strong correlation which
indicates that a large fraction of the noise is, in fact, common-
mode noise. (Just to confuse you, I have plotted 4095-ADC here
so that the pedestals are now at the bottom of the spectrum
rather than the top.) The next plot shows the difference between
these two pedestals, which should give you the non-common-mode
noise for two channels in quadrature. The resulting pedestal
appears to be ~6.2/sqrt(2)=4.4 channels. Converted to electrons,
this is ~(4.4 channels)*(1mV/channel)/(7mV/fC)/(1.6E-4fC/e-) =
3900 electrons.
Pedestal of channel 2 versus Pedestal fo channel 5
Pedestal of channel 2 minus Pedestal of channel 5
Pulse-height distributions
The plot below shows histograms of the charge on the peak
cathode strip for all events where a cathode cluster was found, for
chamber voltages of 1550 and 1600 volts. Pedestals have been subtracted,
which were of the order of 100-200 ADC channels so the top of the ADC
spectrum is ~channel 3900-4000 in these histograms. Break-down voltage
for this chamber gas is ~1950 V, so these spectra were collected at
well below the break-down voltage.
Peak strip distributions for 1550 and 1600 V.
Calibration data
The eight preamp channels were calibrated by putting a calibration card
with built-in capacitors between the detector output and the preamplifier
and sending step voltages into the capacitor card. The voltages were
not especially precise--set with a scope and a standard pulser. Around
150 events were taken at each voltage step, covering the range of
the ADC. The plots below show the output ADC value versus the input
pulser voltage for all eight channels and the full and lower portion of
the calibration data from channel 2. All data look very linear over the
range of interest (at least within the errors of the calibration method).
A couple of channels show a small flattening at the
lowest portion of the spectrum which is probably the onset of pedestal
values. The final plot shows 3 channels, lower portion. You can see
that the differences between the straight line fit and the data points
are similar which either says that the non-linearity of the preamp is
uniform or points out the error in the input voltage which was fed
simultaneously to all channels.
Calibration data from all 8 preamp channels.
Post-script
Calibration data from all preamp channel 2.
Post-script
Lower portion of channel 2 calibration with full straight-line fit.
Post-script
Post-script
Lower portion of three calibrations.
Calibration data with DAC
A sample set of calibration data taken for channel 3 of the preamplifier.
A sample set of calibration data taken for channel 3 of the preamplifier,
showing just the lower portion of the spectrum.
The residuals of a straight-line fit to the calibration data. The residuals
are in units of ADC channels where full scale=4095, and the residuals are
plotted agains ADC value. Just the
portion above DAC=15 was used for the fit. The two plots show three separate
measurements taken for channel 3 (top plot) and channel 5 (bottom plot).
Note that to get a calibration measurement to better than 1% of the most
probably pulse (which should fall in ~ channel 400), the ADC gain must
be calibrated to better than ~4 channels.
Last update 11/6/98 by
Melynda Brooks