Log entry time 09:56:27 on March 10, 2010

Entry number 311268

This entry is a followup to: 311250

I thought it might be useful to copy this follow-up here.  The conclusion
was to use a GDG to set a fixed 200usec MPS pulse for the DAQs that need
it (Moller, maybe Compton?).


> From: Kent Paschke 
> Date: March 10, 2010 9:29:28 AM EST
> To: "Oleksandr Glamazdin" 
> Cc: brads@jlab.orrg, rom@jlab.org, camsonne@jlab.org,
> romanip@jlab.org, suleiman@jlab.org
> Subject: Re: Settle time duration question (MPS signal)
> 
> Hi Sasha,
> 
> To be clear: you are saying that when beam goes from OFF to ON (in
> the "User" pulse mode) it may wiggle around a bit in position,
> intensity, or energy for 200 microseconds.  (I've heard that
> statement also, but can't independently verify it.)
> 
> And as you point out, Brad's log entry (and my reply) were about an
> entirely different issue: the polarization transition, which is
> much faster.
> 
> So, I'm not sure if anyone needs further discussion of the
> polarization transition or helicity control signals, but I'm
> certainly willing to help with that if needed.
> 
> Cheers,
> Kent
> 
> 
> On Mar 10, 2010, at 9:06 AM, Oleksandr Glamazdin wrote:
> 
>> Hi,
>> 
>> I think there is some misunderstanding. 200 microseconds is the
>> accelerator stabilization time. This time is important for the Moller
>> measurements in pulse mode.
>> 
>> Sasha
>> 
>> 
>>> Hi,
>>> 
>>> I guess that the question depends on what we mean by settle time, but the
>>> Pockels cell transition (and thus the polarization change) is complicate
>>> in less than 10 microseconds and the optics have settled (even by
"parity"
>>> standards) in less than 60 microseconds.  This is confirmed by bench
tests
>>> with the laser and also by measuring the electron beam characteristics at
>>> a range of T_settle settings.  There may still be variations in the
degree
>>> of polarization at the level of a few 10^-3 at the beginning of the
>>> helicity window, but we've tended to ignore these effects... and they are
>>> greatly reduced or eliminated at the higher flip rates. (Pockels cells
>>> hate to hold one voltage state for a long time, so the 30 Hz flipping
>>> generated a host of strange time-dependent problems.)
>>> 
>>> I wouldn't object to exending the settle time to 200 microseconds (at the
>>> cost of potentially ~2.5% of our running time) for DAQ reasons.  On the
>>> other hand, if more time in needed between helicity windows, a DAQ can
>>> time out early on each window,  discarding the end of each helicity
window
>>> in order to use that time for readout. (I think the Compton DAQ will do
>>> this, at the level of 100-200 usec additional deadtime per window).
>>> 
>>> By the way, that would be better than discarding the front of the
helicity
>>> gate.  If there were a time-dependent change in polarization, one would
>>> expect to have more variation at the beginning of the window than the
end.
>>> 
>>> Cheers,
>>> Kent
>>> 
>>> On Mar 9, 2010, at 5:58 PM, Hall A Electronic Logbook wrote:
>>> 
>>>> User name brads Log entry time 17:58:39 on March 09, 2010 Entry
>>>> number 311250 keyword=Settle time duration question (MPS signal)
>>>> 
>>>> Right now the T_settle signal that we receive in the counting house
>>>> is typically 100 usecs wide. I see this is a parameter that can be
>>>> set in the Helicity Control board GUI (Figure 1).
>>>> I understand through Sasha (who spoke to Hari) that the physical
>>>> settle time is at least 200 usec. That is, it takes the injector a
>>>> minimum of 200 usecs to reliably transition from one stable
>>>> helicity
>>>> state to another.
>>>> 
>>>> If that is true, then I would strongly suggest that the minimum
>>>> T_settle time in the GUI be set to 200 (or 250?) usecs. Otherwise,
>>>> we're just begging for problems/confusion down the line. If that is
>>>> not the case, what is the physical settle time (and how has that
>>>> been measured)?
>>>> 
>>>> Comments?