Before taking the photometry equipment to the telescope some preliminary calibration must be done. Normally these need be done just once. While it is suggested to start with a high voltage of -950 VDC, once the signal-to-noise calibration is set, you can experiment with other settings. If you decide on a different high voltage setting, you may wish to repeat the signal-to-noise calibration, however, unless the high voltage has been changed by more than 10 volts, it may not be necessary.
Signal -to-Noise Adjustment
There are two variables to consider with a digital photon-counting system. One is the high voltage on the photomultiplier tube and the other is the threshold voltage in the pulse conditioning circuit. The optimum settings for these voltages are when the best Signal-to-Noise or S/N ratio is obtained. This is a point where the net number of counts per second (e.g., with and without a fixed amount of light) has the largest difference, i.e., the ratio of light counts to dark counts. If you do not wish to do this calibration, for a 1P21 or equivalent photomultiplier tube, a good threshold voltage is -9.5 VDC for the MVL100 circuit and 20 millivolts DC for the new pulse conditioner circuit, with the high voltage set at -950 VDC or the voltage determined from the next section. Let the tube sit in the dark with the voltage applied for about 30 minutes to let it stabilize. The ambient temperature should also be stablized. If the temperature varies by more than a few degrees during the calibration, it may be best to wait and do the calibration once the temperature has stablized. A cooler temperature is preferred to a warmer temperature. Refer to Figure 32 for a test set-up to do this calibration.
 |
| Figure 32 Signal-to-Noise Calibration Setup |
- With the Vhv set at -900 VDC and the Vth set at -10.0 VDC (MVL 100) or 10 millivolts DC (for new circuit), adjust the light intensity for about 5000 counts per second.
- Set Vth to -6.0 VDC (for MVL100) or 5.0 millivolts DC (for new circuit).
- Set Vhv to -600 VDC.
- Record the counts per second with (CL) and without (CD) the light source.
- Increment the Vhv in 50 volt increments, repeat step 4 after each increment, until -1050 VDC is reached.
- Next, increment Vth in 0.5 VDC or 5 millivolts DC increments while repeating steps 3, 4 and 5 at each increment.
- Make a graph of the results. The vertical axis should be the Log10 of the counts. The horizontal axis should be the Vth settings.
- The resulting plots should be a set of curves (CL and CD) for each High Voltage setting.
- The best S/N is with Vhv and Vth settings that produce the largest difference between the CL and CD curves.
- Note these voltages and adjust the system for them. This will give the best performance.
What High Voltage Setting to Use
The higher the high voltage, up to about -1,000 VDC, applied to a photomultiplier tube, the more sensitive the tube is. Higher voltages will also create more dark current (noise or dark counts). A lot depends on the tube being used. If more sensitivity is required, then a higher voltage may be considered. For most photon counting use, -950 VDC is a good value. If the signal-to-noise calibration has been done, use the high voltage determined by that.
Note: The system zero points and dead time will change with a high voltage change.
The following is optional and discussed for those wishing to understand their photometry system better.
If you wish to determine what the optimum high voltage is you for your photomultiplier tube you will need to perform the following experiment. You will need a light-tight box with a small wheat type bulb or LED to act as a constant light source. See Figure 29 for a suggested arrangement. The photomultiplier tube with socket and voltage divider you plan to use in the final equipment along with a magnetic shield should be used. With no light on in the box, set the tube's high voltage for -1,000 VDC and let the tube "cook" for about 30 minutes. This will stabilize the noise counts.
Note: It is important that the ambient temperature be constant as it will affect the sensitivity and noise counts. Slight temperature variations (a few degrees) will not be much of a problem, but large swings (10's of degrees) may have significant impact and should be avoided.
Next, adjust the high voltage to your starting voltage, usually the lowest you want to check. Now set the brightness of the lamp or LED to get counts in the tens of thousands. The counts will vary slightly due to system noise, but should average close to each other for a constant high voltage and a 10 second integration. Make a table of high voltage vs. counts. You can also prepare a table that varies the high voltage with no light. This will show the system noise vs. high voltage. From the tables compare sensitivity vs. noise and decide what high voltage you think is best.
Once the desired voltage has been determined the power supply should be regulated ±1.0 VDC or better. A digital readout of the high voltage is helpful for keeping an eye on it. Figure 33 shows a typical plot of high voltage (-900 VDC to -1,000 VDC) vs. sensitivity. As can be seen, a change of the 100 VDC can produce nearly a 0.3 magnitude change.
 |
| Figure 33 High Voltage vs. Photomultiplier Tube Sensitivity |