Just to flesh out this log entry (we were interrupted by the party):
The "test run" 13215 was taken in a configuration suitable for linearity tests. The short version is that a 25 kHz triangle was is mixed with a DC signal having known asymmetry from the TRIUMF fake asymmetry box.
The parity ROCs are still mostly bare following the tests and triage of the VQWK channels over the last couple weeks. For these tests I have re-powered ROC1 and ROC2 and re-cabled the gate and clock signals to all modules in ROC1. In order to get a helicity signal, I had to pull the delayed helicity signal from the fiber transceiver in the trigger supervisor, duplicate it using a level translator, and send copies back to the trigger supervisor and to the fake asymmetry box on the SPIN ST IN input. (The delayed helicity NIM output on the fiber transceiver doesn't seem to be able to drive two 50 Ω loads in parallel.) The fake helicity box is using the same ~1 kHz gate and 20 MHz clock signals as the VQWK ADCs.
The manual for the fake asymmetry box claims that, in order for the spin state to change promptly with the gate signal, the SPIN input should be present at least 0.150 μs before the GATE input. Unfortunately we have the delayed helicity transitions about 1 μs after the gate begins, and I wasn't successful at finding a delay module. The modulation seems to work anyway, but in a way that may depend on the load present at the output: watching the modulated voltage output on the scope, transitions seemed to lag the gate by as much as 400 μs; but, watching the modulated current and/or voltage driving the few kΩ load presented by the preamps, the transitions seemed to be relatively prompt. I don't understand this, but possibly the "programmed" asymmetry will only appear in the last subblocks.
The fake asymmetry box is set to generate 2.5 V and 2.5 μA outputs, each with an asymmetry of 100 ppm.
There are three VQWK inputs. One is a linear fan-in-fan-out copy of the 25 kHz triangle wave, amplitude 0.100 V (give or a take a factor of two), and one is the modulated current, preamplified (gain 2 MΩ). The two are mixed using a voltage divider with a gain 1/100 arm and a gain 1/1000 arm. The preamp's voltage gain is -200, so run 13215 should have a negative 5 V "signal" (with a 0.5 mV asymmetry) modulated by a free-running 20 mV "raster." That's consistent with observations on the scope.
John Leacock has a better handle than I do on mapfiles and the analyzer and has offered to help me look at this data.
Figure 1
