QIE Tests at the ADMEM Teststand

QIE calibrations

Read out QIE calibrations with Frank Chlebana's code. Check that results meet specs. (Hao Wei, Frank Chlebana) *DONE*
Read out QIE calibrations repeatedly over long period (days). Check that results meet specs. ( Hao Wei, Frank Chlebana) *DONE*
Modify Rod Klein's code to include QIE calibration in simple text-driven program which can be easily modified to add new tests. (Hao Wei, Rod Klein) *DONE* (Frank points out that both of the above should share code. We should work toward this goal.)
Modify delays in QIE calibration to make it as fast as possible. (Hao Wei, Rod Klein) *DONE* Understand how calibration fails when delays are too small. Ideally, delays should be in code and set by timer routines, not loops, so that routines are independent of processor they are run on.
Presumably, trigger outputs also vary with QIE calibration. But this also presumes logical values are in trigger output lookup tables, also (does code exist to download these?). Attach trigger outputs to digital register (or better yet, DRAW card?) and check linearity of trigger signals vs. QIE signals. (Hao Wei, Steve Hahn, Peter Wilson?)

Charge (i.e. PMT signal) calibrations

Make sure dark current from PMT does not effect QIE pedestals. (Steve Hahn, Hao Wei) *DONE*
Trigger LED with HALT frame from TESTCLK emulator clock (TSIE), and read corresponding PMT signal in L2 buffer with single ADMEM channel. (Hao Wei, Steve Hahn, Mike Lindgren) *DONE* Do this repeatedly over long time scale (24 hours).
Vary grounding scheme (how to make this more or less like detector ground?) and note effects on pedestals and LED signals. Source monitor channels, too, for that matter.
Move LED signal relative to 132 ns gate and note signal loss and RMS. (Hao Wei, Steve Hahn) *DONE* Vary PMT signal width (by varying LED signal width) and repeat. Match PMT LED signal to actual detector signals as much as possible.
Use filter wheel to vary LED signal amplitude and test linearity of QIE response for PMT signals. (Not clear how to do this: use high resolution, high speed VME-based ADC and read out both ADC and ADMEM? So far, Mike has not found a good candidate.) (Steve Hahn, Hao Wei, Mike Lindgren, Howard Budd?) *Long-term study*
Read four adjacent L1As so as to sum four 132 ns gates. Can this method be used to determine what fraction of energy is lost in tail of PMT signal? (Steve Hahn, Hao Wei)
Read four adjacent L1As. Compare ADMEM readout vs. results of reading same gates out with integrating digital scope. This is to look for possible charge loss in QIE operation. (Steve Hahn, Hao Wei) *DONE*

Check trigger outputs vs. QIE signals for PMT input signals for linearity. (Steve Hahn, Hao Wei, Peter Wilson?) *Long-term study*

Connect real CHA and PHA phototubes to ADMEM inputs with coaxes of varying length, readjusting the delay to put the signal in the correct location in the gate, and readout the signal in each of the four L1 buffers to see the effect of cable length above and beyond the 24' reflection point. (Steve Hahn, Hao Wei, Mike Lindgren)

Connect real CHA and PHA phototubes to ADMEM inputs with coaxes of approximately the right length and measure the noise at the pmt dynodes (they must be terminated in 50 ohms). This noise will determine the minimum threshold for the CHA and PHA TDCs. (Steve Hahn, Mike Lindgren)

Current calibrations

Read out 1000X current readout for dark current (1 nA of dark current should be 1 nA * 1000 / 50 nA = 20 counts). Artificially raise dark current (LED with DC current varying current; could also use filter wheel), and look at larger signals. (Steve Hahn, Hao Wei) *DONE*
Need some method of calibrating current readout. Limiting factor will be noise level of current source. Should not let this problem prevent us from trying current source varying current over ranges corresponding to source values in the central and the plug (0-500 nA?) as soon as possible, however. Do we need to tweak gain of current readout to maixmize resolution? Did first crude measurement reading current from PMT anode with picoammeter. (Steve Hahn, Hao Wei, Howard Budd?) *DONE* *Long-term study*
Also, need to worry about settling time of current readout. If time is too fast, cosmic ray spikes ruin source calibrations; if time is too slow, delay becomes evident in source peaks. (In Run 1, settling time was 100 ms with source moving at 0.67 cm/sec and readout rate at 5 Hz in central; Charlie has used the same 100 ms for the Café cards. All of these need to be tuned together eventually, for both the central and the plug.) (Steve Hahn, Hao Wei, Mike Lindgren?) *Long-term study; needs detectors*

Scaling calibrations

Do all the above for a single channel on ADMEM *DONE*
Intermediate tests: do all the above for maximum channels usable with existing power supplies (6 channels?) *DONE* some
Do all the above for entire ADMEM (20 channels = 20 Café cards) *DONE* some (even 2 full ADMEMs in a crate)
Do all the above for maximum number of ADMEMs in one crate (currently, 6 ADMEMs in central crate)

Noise tests

Check RMS values in QIE calibration with different power supplies. (Hao Wei, Steve Hahn, Lee Scott) Investigate this as function of power supply load. Also, look at noise spectrum with spectrum analyzer (Charlie Nelson) *Long-term study*

Unipower switcher modular supply, with and without extensive filtering *DONE*
Linear power supplies (Condor Electronics)
Wiener switcher (linear regulator?) modular power supply
Rittal (Lambda) swither/linear regulator modular power supply
Others?

Use artificial noise source (the noise-blaster?) and QIE calibrations to check ADMEM immunity to noise at various frequencies? Can we somehow quantify this? (Hao Wei? Steve Hahn? Gary Drake?) *Long-term study*

Cross-talk tests

Café card cross-talk test: do QIE calibration on one channel and read out every other Café card on same ADMEM (Hao Wei, Steve Hahn) *DONE*
ADMEM cross-talk test: do QIE calibration on every channel on one ADMEM and read out every channel on adjacent ADMEM cards (Steve Hahn, Hao Wei) *DONE*

Preparation for data-taking

Check that normalized readout into (TR)YBOS raw data banks works correctly. (Frank Chlebana? Kurt Biery?) *DONE*
Incorporate QIE calibration into RUN_CONTROL scheme including new calibration banks. (Kurt Biery? Jim Patrick?) *DONE*

Miscellaneous

"Hiccup mode" - check that this works as advertised by reading PMT signal data and checking that raw data event corresponds to normalized event as per lookup table in FRAM.
Measure actual currents for each power supply (+5 digital, +5 analog, -5 analog, +15 analog, -15 analog) both in quiescent and active modes *DONE*
Check that pipeline and L2 buffers work correctly at every pipeline depth?

QIE tests
modified May 14, 1998