[d2n-analysis-talk] Quick scaler question (fwd)

[Diana Parno]

I thought several of you might be interested in an email from Vince 
regarding the structure of the scalers in the BigBite and LHRS stacks:

---------- Forwarded message ----------
Date: Sat, 24 Jul 2010 07:57:19 -0400 (EDT)
From: vasulk at jlab.org
To: Diana Parno <dseymour at andrew.cmu.edu>
Subject: Re: Quick scaler question


Hi Diana,

Again you caught me when I was most busiest.  But then that always seems
to be the case these days.  At this time, I am planning to go to China for
a workshop.  My trip begins tomorrow morning bright and early.

Yes the bbite scalers are read out asynchronously and inserted into the
data stream as event type 140 every few seconds.  The evbbite scalers are
synchronously readout in the data stream every 100 events.  So the evbbite
scalers have more entries than the 140's.

The scalers continuously count, so there is no averaging.  You just see
finer structure with the evbbite because the readout is more frequent.
Hence you can see the beam trips more clearly

There is actually a subset of scalers that are readout each event.  These
are similar to the evbbite scalers.  Just more frequent readouts.

I hope that helps.  But you have the general idea already.

Best Regards,
Vince

> Hi again Vince,
>
> OK, I think I've worked this out -- please let me know if I'm wrong.
>
> The bbite variables are read out from the scalers asynchronously, in
> special events with the code 140. The evbbite variables are copies of the
> same variable (?) read out synchronously, in the event data stream. Thus
> an evbbite value might change from event to event.
>
> I'm thinking that perhaps the rate distributions look different because of
> the different readout times. The bbite values are essentially averaged
> over a longer period of time, so you end up with tighter clustering around
> the central value -- the outliers seen in the evbbite values cancel each
> other out.
>
> Does that sound about right?
>
> Diana
>
> On Fri, 23 Jul 2010, Diana Parno wrote:
>
>> Hi Vince,
>>
>> You said a while back that you felt neglected since I wasn't emailing
>> you
>> lots of questions. Here's one!
>>
>> I'm confused by the scaler options in the BigBite data stream. On the
>> one
>> hand, I have the scaler group "bbite." On the other, I have "evbbite". I
>> have
>> done some searching to try to figure out the difference between them,
>> but all
>> I've found is that evbbite is ROC 12 data, which does not quite clarify
>> things for me. Yet they do seem to be different things -- you can see in
>> the
>> plot I've attached that, for example, the rate for the d10r variable for
>> the
>> beam-current scaler has quite different distributions in the two blocks.
>> (This was plotted using the bbite_d10r and evbbite_d10r variables in the
>> main
>> tree T.) When I define trees for each scaler group in my .odef file, I
>> get
>> radically different numbers of entries in each one -- for example, 5218
>> entries in EVBBITE vs 19 entries in BBITE for the same period of time.
>>
>> So -- there's clearly a difference between BBITE and EVBBITE. What is
>> it?
>> Which do I trust for, say, the beam-current distribution?
>>
>> Thanks for your help. Hope things are going well. Stay cool -- I hear
>> it's
>> pretty awful outside in VA!
>>
>> Diana
>