Hall A Helicity Circuit

Contents

Overview
Operational notes
Safety
Responsible Personnel
Old documentation

Hall A has several instruments that require knowledge of the helicity state of the beam pulse. Here we focus on the use of the helicity pulses in the ``standard'' spectrometer CODA DAQ, as opposed to for the Moller detector or for the specialized HAPPEX parity detectors.

The helicity circuitry was set up initially for the FPP experiment E89-033, and used subsequently for the FPP G_E^p experiment93-027. Modifications were subsequently made to put helicity signals in the electron arm, for the independent spectrometer DAQ in E94-010. Also, the polarized ³He target experiments required much better false asymmetry knowledge, ~10^-5 to 10^-4, as compared to the FPP experiment requirements of ~10^-3 to 10^-2, necessitating a reexamination of the system.

Input to the system is one correlated-in-time h+ helicity signal from MCC. Since fast scaler readout has not been done in the standard system, the helicity-correlated pulse provides the simplest monitoring. Either pseudorandom or toggle modes may be used for the pulse sequence.

The h+ pulse, and its complement, are sent downstairs to the two spectrometer detector stacks. Circuitry in exah arm then generates a shortened pulse, chopping off the leading edge to allow the helicity to stabilize. The output pulse is also ``run'' gated.

For each arm, and both the h+ and h- helicity pulses, the shortened, gated, output pulse is fanned out to the following:

two ADC channels, to provide redundant information on an event-by-event basis of the beam helicity for that particular event,
a scaler gate, so that helicity-gates trigger and beam current information is obtained
a scaler channel, so that the number of helicity pulses may be counted to crudely check system functionality.

A postscript drawing of the hadron-arm helicty circuit is available. The point of the system is to use one channel of a gate and delay module delay the incoming helicity pulse, and then with the delayed output to trigger and latch the signal on a second channel of a gate and delay module. The latched signal is cleared by the start of the helicity pulse of the other phase.

The e-arm circuit is functionally identical, but uses different modules. The LeCroy fan it / fan out is replaced by a Phillips 740, and the Phillips 794 quad gate & delay is replaced by a LeCroy 2323A CAMAC dual gate and delay, to generate the delays, and a LeCroy 222 NIM dual gate and delay, to provide the latching. Note that the Phillips and LeCroy units are labelled differently:

function LeCroy label Phillips label

start gating start trigger

stop latched signal stop reset

prevent output signal blank inhibit

set ``time scale'' to hold level latch FF

One further difference is that, on the e-arm, we use the nim-bar output of the gate and delay directly as input to the CAEN scaler veto - the CAEN scaler is vetoed to not cont, rather than being gated. On the hadron arm, the new 32-channel scaler installed March 1999 uses a twisted pair veto input, provided by taking the fanned-out gate signal, and inverting the twisted-pair wire leads to the scaler unit.

function	LeCroy label	Phillips label
start gating	start	trigger
stop latched signal	stop	reset
prevent output signal	blank	inhibit
set ``time scale'' to hold level	latch	FF

The run gate is put into the circuit so that the scalers may be stopped and cleared at the start of a run, and stopped and read out at the end of a run. We have confirmed, for both the Phillips and LeCroy gate & delay modules, that the output gate goes false at the end of the run. (In both cases, run-bar goes true into the inhibit/blank, clearing the output gate of the gate & delay.) Note that the trigger supervisor run output is actually left true when runs are not in progress, so that scalers can also be monitored then. It is only momentarily set false at the start and end of runs.

Operational notes

The helicity circuitry is set up so that no adjustments are needed when the pulse rate is changed. Nevertheless, it is desirable to confirm its operation at the start of an experiment and at regular intervals. Previous problems have included MCC no longer sending us a helicity signal, and MCC sending us the clock signal, when in pseudorandom mode, rather than the helicity signal. (Both incidents occurred in the middle of runs, after operation had been checked.) Since the spectrometers and the Moller detector use the same signal, a few hour Moller measurement can confirm the status of the system, in case the spectrometer physics requires significantly longer times.

In toggle mode, helicity pulse lengths are fixed and the helicity alternates. There is some confusion about how to label the pulse frequency, since 1 s long alternating pulses have been called 1 Hz, whereas 33 ms long alternating pulses have been called 30 Hz. (These are the two pulse lengths most often run.)

In pseudorandom mode, one may get two successive pulses of the same helicity, generating a single double length pulse. The net effect of this is to reduce the frequency by 25 %. That is, with 1 s long pulses in toggle mode, one will measure 0.5 Hz of h+ and 0.5 Hz of h-. In pseudorandom mode, the scalers measure 0.375 Hz of h+ and 0.375 Hz of h-.

Because of the phase of turning on / off the run relative to the helicity pulses, one can measure up to 2 pulses different between h+ and h-, leading typically to ~10^-3 false asymmetries in pulse numbers and integrated charge. Any conclusion about charge false asymmetry requiures at least a time correction from the pulser scaler. A 0-offset correction is also desirable if beam has been off for significant times.

Changes to the circuitry may result in the helicity pulse getting to the scaler before, or too soon after, the scaler gate. Circuit modifications should be checked.

Previous versions of the circuit used electronics that sometimes generated spurious signals to be counted, or missed generating signals to be counted. These problems were typically at the ~10^-3 level. The ADC signals and the scaler gating were not affected, and there should be no problem with the data.

It was generally believed that a 200-microsecond blank-off time at the start of the pulse was sufficient to allow the helicity to stabilize. Early 1999 information from HAPPEX indicates that the helicity does not really stabilize for 400 - 500 microseconds. (On Mar 24 1999, the blank-off time was increased to nominal 500 microseconds.) This may lead to a dilution of the beam polarization, at a level no larger than 200 microseconds / T, where T is the average pulse length. For 1 s long, pesudorandom pulses, T = 4/3 s, and the dilution is no larger than 1.5*10^-4. It should be the same for both helicities and should not lead to a false asymmetry.

Because of the blank-off time, some fraction of the events should not be either good h+ or good h-. For example, with 1 s pulses in toggle mode, the fraction without good helicity signal is the blank-off time / 1 s. In pseudorandom mode, it is 75% of this. For upcoming April 99 commissioning time, with 33.3 ms pseudorandom pulses, the fraction is 75% * 0.5 ms / 33.3 ms = 1.125%

There are no safety issues concerning the helicity monitoring circuitry.

Responsible Personnel

The people most familiar with the Hall A helicity circuitry are: