The problem:
Some users report that their CSP modules are often destroyed by their detectors. Some detectors are prone to 'sparking', which send very large and damaging pulses to the CSP input. Gas based detectors such as proportional chambers can be problematic as they are particularly prone to sparking. High voltage APDs are also notorious sparkers.
When a detector sparks, the detector is momentarily 'shorted out'. This causes the blocking capacitor (0.01 μF on the CR-150-R6 eval board) to discharge through the preamplifier input. If the detector bias voltage was set to 1000 V, for example, this represents a charge of 107 picocoulombs very quickly draining into the preamplifier input. To provide a sense of scale, consider that a 2 picocoulomb signal saturates the CR-110 CSP.
The absolute maximum size of the input pulse:
After testing many CSP modules with sparking detectors, some practical limitations have been observed. These tests were made using the CR-110-R2.1 and CR-111-R2.1 circuits and these two CSP designs failed under similar conditions. These limits are:
- largest positive input pulse: 3.0x10-6 coulombs
- largest negative input pulse: 1.5x10-6 coulombs
Assuming AC coupling is used (such as with the CR-150-R6 board) and the applied detector bias is negative, the CSP receives a positive input pulse. Alternatively, if a positive bias is applied in an AC coupled circuit then CSP then receives a negative input pulse.
Assuming the blocking capacitor in the AC coupled circuit is 0.01 μF (as is the case in the CR-150-R6 eval board), a CR-110/111 CSP should be relatively safe from sparking detectors if the detector bias is limited to the range -70 V to +150 V. If higher detector bias is required please read the next section.
CSPs that are destroyed in this way have an output voltage at one of the rails: approximately 3V (or -3V)
Tests were also made on the older 'Revision 2' versions of these CSPs and the same limitations were observed.
What to do if your detector is destroying preamplifiers:
Externally applied diodes cannot protect the input under these conditions. Other types of devices such as fuses have not been shown effective in protecting the input either. That said, there are a couple things we can do to mitigate the problem. The most effective method of protection is to limit the current flowing into the preamplifier input by adding a series resistor to the input. The resistor value should be in the range of approximately 220 ohms. If the value is too small then the current is not sufficiently limited and the preamplifier may still be damaged. If the resistor value is too large then the preamplifier output rise time can be adversely slowed. Also there is a small noise penalty with the addition of this resistor, but often this is an acceptable trade off.
If you are using the CR-150-R6 CSP evaluation board, a convenient location for the resistor is R11. The CR-150-R6 is provided with the position shorted. To add the protection, remove the short and install the 220 ohm resistor which is provided with the CR-150-R6 board.
The other thing we can do to mitigate the problem is to reduce the size of the blocking capacitor from 0.01 μF. In general, we would like the value of this capacitor to be large compared with the detector capacitance. If, however, the detector capacitance is under 100 pF then consider reducing the blocking cap to 1000 pF. This will reduce the magnitude of the sparked charge in our previous example from 107 picocoulombs to 106 picocoulombs (10-6 coulombs) which may help significantly in protecting the preamplifier from a sparking detector.