Instruction for Si Sensor Long-term Current Testing
Starting Test
1) Start with sensors bonded into special gelpacks. Be sure gelpacks are carefully marked with sensor ID. Sensors should be matched by operational voltage.
2) Make sure that 5V and bias supplies are both OFF.
3) Plug gelpack phone cords into adapter box.
4) Move lemo cables (starting with 1) to outputs corresponding to phone plug inputs.
5) Turn on 5V supply.
6) Turn up bias supply smoothly and slowly to operational voltage. Use the largest value among the sensors being used. Make sure that this doesn’t exceed the highest plateau voltage of any of the sensors.
7) Using multimeter, check voltages on all sensor outputs by unplugging lemo cable from lemo-ribbon adapter box. Make sure they are not too high (>100mV for guard, >500mV for bias). Make sure bias voltages aren’t too small (<100mV).
8) Check that all cables are plugged in correctly.
9) Double click on the file "Pfadc" (not Pfacd12 or Pfadc15) and edit it to correspond to sensors and signals being tested. Save it and exit Notepad program.
10) Double click on the file "start" (the MS-DOS Batch File one).
11) Click on button labeled "Run" on top menu bar, pull down and click on "Start" entry.
12) Use about 300 seconds for time delay (time between measurements).
13) Enter number of channels being read out (total including both guard and bias).
14) Enter bias voltage.
15) Enter your name and comments.
16) You will see a listing of program parameters. The number of samples is hard-coded to equal 10 inside program. Output file is always output.dat. Verify that all other information is correct.
17) Press Enter key to start program. First measurement done right away, then it waits the delay time you entered. Stay long enough to verify it reads more than once. The values on the screen display are in nanoamps so 1000 corresponds to 1 microamp.
18) Check results on cleanroom computer or web occasionally. There is a program running on the Phobos-srv which reads output.dat every few hours and writes in the database so it make take a while to appear on the web. The web link is Detector > Si Detectors > Si Database Access > Measurements > Click on "details" link for desired batch> Click on desired sensor number > Long term. Both bias (active area) and guard current will be shown when you click on any long-term test link. If current starts to shoot up to hundreds of microamps, you should stop the test.
19) You can minimize the program window but not close it. If you hit the wrong key by accident, you will get a warning and will be asked for a confirmation before the program is actually killed.
Ending Test
1) If not done recently, check currents on all wafers to find any bad ones (Link is: Detector > Si Detectors > Si Database Access > Measurements > Click on "details" link for desired batch> Click on desired sensor number > Long term. Both bias (active area) and guard current will be shown when you click on the test link.)
2) Stop program by killing window. You will be asked to confirm that this is what you really want to do.
3) Turn down bias voltage (bottom supply). Turn down slowly and smoothly.
4) Turn down 5 V supply.
5) Disconnect gelpack phone cords from adapter box.
6) Mark boxes for sensors to be unbonded and returned to their original gelpack.
Bonding and Gelpack Details
There are 9 special gelpacks with a phone cord connected to 3 pieces of metal inside. The sensor sits on top of the copper strip which makes the back plane contact. The two other pieces are used to connect the bond and guard voltages. The sensor must be positioned so that the readout pads are next to one of these metal pieces and the other is along the side of the sensor. The side piece bonds to the guard ring running along the side of the sensor and the other piece bonds to one of the bond voltage input pads (labeled "B" on the sensor).
NOTE: Frequently, a sensor will still have its guard and bond lines tied together which is done for an earlier Si test step. Usually, but not always, this is indicated by a sticker on the gelpack for the sensor which says "B-G" or some such. It is critical to remove this bond before running the long-term test! You should remove or cross out the BG note on the sensor’s gelpack once this is done.
Other Miscellaneous Details
There are 2 power supplies hooked up to the long-term test system. The old one is the 5V supply for the temperature probe, the new HP supplies the silicon back-plane bias. This voltage should be set according to the full depletion voltage tests of the sensors on the long-term test station. On the web table displaying the properties of sensors available for mounting you can find the operational voltage for each sensor. In case that more than one sensor is mounted on the long-term test-station you have to take the highest of the operational voltages. However, the sensors have an operational range (called plateau in the properties web page) which the voltage setting of the long-term test station should not exceed. In practice this means that you select the sensors for the long-term test to have roughly matching operational voltages.
There is no VI for the long-term test. The reason is that the long-term test has to run for a long time. Only one LabView instance can be running per machine. Since sensor testing on the probe station is going on, a VI for the long-term test would have to run within a LabView program which is controlling the probe station. This was not practical to implement since the long-term VI would be killed every time you switch program. A QBASIC program does the trick. It can run in parallel with any other program you start on the PC and it is nice and small.
On the cleanroom machine you'll find the directory: C:\Users\Sitester\Long-term\
It contains everything that is needed to run the long-term program.
Start.bat should start QBASIC. you end up in the editor, which displays the program, you go to run and run the program
Pfadc.cfg the configuration file containing the sensors under test, what is tested on them (guard and/or bias) and to which port they are hooked up.
Output.log the logfile so you can see what happened afterwards.
Output.dat The output of the program goes here. Don't ever touch this file! This is the file that gets read by the program on the server which inserts the data into the database. This happens about every 3 hours.
The telephone plug connectors of the test gelpaks have to be plugged into the box next to the power supplies. The LEMO cables connect the ADC to this box. When more than 5 sensors are tested at one time, we only read out the guard ring current. Make sure that everything is OK after you hooked up a sensor and switched on the bias voltage. Measure the voltages on the LEMO outputs on the adapter box and see if they make sense (i.e. a few uA for the active area and 100nA to ~100uA for guard). If not, then probably a bond has popped off.