Operating Instructions
ComTec GmbH
5-2
All of the many possible contributors to less than optimum performance cannot be listed here.
The purpose of this section is to note the usual causes of loss of resolution, and to suggest
curative steps.
Do not expect to diagnose all problems with the detector, a preamplifier, a main amplifier, and a
multichannel analyzer. The spectroscopy system records results, it does not necessarily lead to
the identification of causes. A good, modern oscilloscope will be needed. Also a high quality tail
pulse generator will be extremely useful.
The simplest test is, of course, to connect your detector, apply bias, present a source, and
accumulate a spectrum. Be sure a pulser is not feeding the preamplifier while the spectrum is
accumulting, or resolution loss may result. If the results obtained are far different from what is
expected, it than becomes necessary to troubleshoot the system.
First observe the amplifier output on an oscilloscope at various time base and amplitude settings.
Is the amplifier properly pole/zero cancelled (do the output pulses cause undershoot that persist
for longer than two or more main pulse widths)? Set the main amplifier pole/zero cancellation
(without DC resolution) to obtain the most rapid, complete baseline recovery.
The next step is to remove all sources and, with the detector still connected and bias applied,
present a test pulse to the detector input of the preamplifier using the test input described in
Section 5-2. Make sure the pulser polarity is correct. Set the amplitude of the pulser so that ist
peak accurs near the region of the peak of the source previously used.
Observe the output of the amplifier without DC restoration. Note that the amplifier is not properly
pole/zero cancelled for the pulser feeding the preamp (due to the extra time constant of the
pulser). This is of no consequence for a pure pulser input. Are the baseline fluctuations of 50/60
or 100/120 Hz frequency? A ground loop is indicated. Insert all system line plugs into the same
output. Or are the baseline fluctuations of ramdom frequency between 10 Hz and 15,000 Hz?
The area may be too noisy, causing microphonic problems.
If high frequency noise is observed, is it random or periodic? Periodic noise is a sign of
electronics failure; isolate the cause by observing the preamplifier output. Is the same pattern
observed, or is the problem in the main amplifier? Random high frequency noise may be detector
load resistor or input capacitor breakdown.
Next, accumulate a pulser peak on the analyzer. Calculate its resolution. Repeat with the detector
removed and the input connector of the preamplifier shielded. (Wait five minutes to remove the
preamplifier from the detector after removing the detector bias.) You now have three resolution
figures available for essentially equal energy peaks:
R
S
: source
R
D
: pulser and biased detector connected
R
E
: puser without detector connected
If R
E
is not less than 2.0 keV for two microseconds unipolar Gaussian shaping time constant, the
problem is in the electronics and probably in the preamplifier.
If R
E
is acceptable, but R
D
is greater than 2.0 keV plus 10 eV/pF detector and connection
capacitance, then the problem is either in the detector (microphonics, excess leakage current
noise, breakdown due to moisture or grime on the detector output connectors), or in the
preamplifier (leaky input capacitor, dirty or moist detector load resistor, dirty or moist detector
input connector).
If R
E
and R
D
are acceptable, but the live spectrum (R
S
) is not as good as expected, the problem is
probably in the detector (bad detector, poor charge collection, insufficient bias) or in the
electronics following the preamplifier (count rate too high, improper amplifier pole/zero
cancellation, wrong main amplifier time constant, wrong amplifier shaping-bipolar vs. Unipolar,
improper amplifier shaping for the ADC being used, ADC cannot take the count rate, amplifier or
ADC drift). These many alternatives are not easy to check. Substituting, one-by-one other
detectors, preamplifiers, amplifiers, and multichannel analyzers may help pinpoint the problem.
Checking the above common problems will aid in spotting the source of trouble.
