'2011-06-21
Modify board:
Change R12 and R13 to 1k; ground bottom end of C8
(this makes two 20kHz RC low-pass filters in series)
Cut trace from C8 to J3
Add .01uF from J3 to HV output
See scope traces
Channel 2 (top) is pico output (Ac-coupled) Channel A (mid) is top of C7 (first stage filter) Channel 3 (bottom) is top of C8 (second stage filter)
Changed C7, C8 from 0.01 to 0.1uF. Filter cutoff freq is theoretically now 2kHz. See scope traces Note change of scale on bottom trace.
Ripple on 2nd stage (C8) decreased from ~1V p-p to ~200mV p-p.
'2011-06-15 (hazen, damask)
Test conditions for ripple test:
See sketch
Jumper JP4 installed, JP6 out (feedback taken from Pico input)
Output loaded with 1Meg, output voltage ~ 650V.
Circuit modified slightly.
R8 top end (HV divider) now connected to SHV output (after filters).
1k inserted between JP4/JP6 and U2-2 (add feedback resistor)
Input to Pico: 4.76V Output at TP1: 3.23V (647V output) Load: 1MEG
See scope traces
Channel 1: PICO output through 47pF series, 100k to GND to scope probe Channel 2: PMT output BNC Channel 3: PICO input at pin 1
PICO spec is for 2% ripple. 2% of 650V is 13V, which is what we see! Remove JP4, install JP6. Reset output voltage to ~650V (its different with PICO out of the feedback loop due to non-zero startup voltage of PICO).
See scope traces They are essentially identical. So, we can conclude that the ripple is not being caused by our op-amp circuit.
'2011-01-20, hazen, linden
Continue testing. Install U3 and measure HV output, TP1 and input (on JP4) at several voltages. See Google Spreadsheet
Remove feedback wire from JP2-JP7 and install JP6 to complete feedback loop from HV divider. It works! Stop for now to ponder next steps.
'2011-01-19, hazen, damask, linden
Initial testing. Fix a couple of issues on the board:
- Change R11 to 100k
- Change C11 to 1uF
First test is open-loop. Use wire-wrap to short JP2 and connect to U2-2 via JP7 pin. This closes the feedback loop at U2. Confirm that HV output is present and at a plausible level.
