Re: kicker

[Brendan Kiburg <kiburg@npl.uiuc.edu>]

Dear Peter,
  Sorry I wasn't clear about my thoughts with the muSC discriminator.  I
was trying to imply that we could use the muSC disc in the area as our
only signal, rather than using the linear fanout first.  This would reduce
the delay by the time it takes to go through this one module.  Of course,
we would lose the muSC-lo threshold as well as the ability to set the
threshold from the barrack.  Thus it is not a reasonable option.

  I agree there really isn't much room for improvement wrt the kicking
delay.

Best Regards,
Brendan

On Fri, 12 May 2006, Peter Kammel wrote:

> Dear Brendan,
> 
> The kicker muSC discriminator IS in the area to avoid switching delays.
> I think that's all what matters. So no way to improve simply, I
> believe.
> 
> The change to undating is a bit tricky electonically.
> 
> As this is relevant for the whole collaboration, we should get this
> discussion into the elog. Please post your calculations so that everybody
> can check it and I will follow tomorrow.
> 
> Actually I don't get 20% PP losses but rather 8% in my estimate. Moreover
> I still believe the red and black curves in the kicker right/up online
> display differ only by 10% not by 20% as Fred mentioned in last
> teleconf.
> 
> There is a lot of good stuff on this kicker page. The incident rate
> can be derived from a fit of the lower/right panel. The overall PU
> losses can be calculated by integrating the kicker protected region
> in the upper left panel. That should be done to comment on the
> 8-20% discrepancy.
> 
> Thus my impression is that at current rate conditions we lose <10%
> by PP, so not much can be gained by improvements. But that disagrees
> with Fred's trend plot and needs to be scutinized.
> 
> Best regards
> 
> Peter  
> 
> On Sat, 13 May 2006, Brendan Kiburg wrote:
> 
> > Dear Peter,
> >   I agree that there are improvements that could be made to the hardware 
> > pile-up protection.  Specifically, we could use an updating pile-up 
> > scheme, as you mentioned, or we could add in all signals which 
> > ultimately contribute to our software pileup protection (muSC-lo, muSCA) 
> > into our kicking scheme. 
> >   From my brief studies, it seems the best way to improve the maximum 
> > allowable rate is to minimize the time between the kicking muon and the 
> > effective kick (now ~600 ns).  It is not clear how to do this other than 
> > to move our main muSC discriminator to the area, or use the one there.  
> > This is undesirable because we can't tune the discriminator thresholds 
> > remotely unless we move to some CAMAC system in the area. 
> >    Since adding additional signals to the hardware pileup protection 
> > would potentially extend the kicking delay, we should think carefully 
> > about minimizing this delay.  Fred and I discussed the updating 
> > discriminator scheme briefly today, and it seems to be the most 
> > promising method of ensuring that muons aren't vetoed "from the left" by 
> > a sneaker.  They would simply need to survive their own fill without a 
> > sneaker.  Perhaps a brief Monte Carlo is in order to understand the 
> > effects on the background from variable length kicker gates.
> > 
> > Until then, we will try to maintain quality data taking and maximize our 
> > statistics in the current configuration.
> > 
> > Best Regards,
> > Brendan
> > 
> > 
> > Peter Kammel wrote:
> > 
> > > Dear Brendan,
> > >
> > > Thanks for your independent thinking and for reminding me of the
> > > real situation. I made similar studies in
> > > http://kaon.physics.berkeley.edu:8080/analysis-run10/1
> > >
> > > One can be more tricky than we are at present.
> > >
> > > Why wait to the end of the gate when we know that there was a pile-up?
> > > Why start a new kicker with a pile-up muon?
> > >
> > > The solution to both problems is switching the gate generator
> > > in http://kaon.physics.berkeley.edu:8080/analysis-run10/1
> > > to updating. I believe that, but would have to think much more carefully
> > > before doing this.
> > >
> > > The result would be that the kicker start rate is cleaned from bad 
> > > starts,
> > > somewhat reducing the deadtime.
> > >
> > > However, matters are complicated enough, including the beam tuning.
> > > With muon on request beam tuning is too indirect.
> > > Probably we should wait until my return in 10 days before trying
> > > that...
> > >
> > > That are my "remote" feelings at least.
> > >
> > > Best regards
> > >
> > > Peter
> > >
> > >
> > >
> > >
> > > On Fri, 12 May 2006, Brendan Kiburg wrote:
> > >
> > >> Dear Peter,
> > >>  I was considering the statement that you made that we could obtain 38
> > >> kHz effective rate in an ideal world.  If our kicker were perfectly
> > >> effective and instantly deflected the beam whenever a muon passed 
> > >> through
> > >> it, then this would be true.  But, as you have shown, it takes time for
> > >> the signal to reach the kicker and it doesn't kick perfectly, so we 
> > >> lose some events due to pileup.
> > >>
> > >> I have calculated the "ideal" rate we would observe in the case that no
> > >> sneakers exist, as well as the case that we actually have, where the 
> > >> full
> > >> beam is unsuppressed for ~550 ns and then we have some extinction factor
> > >> of 100.
> > >>
> > >> I considered the possibilities that a muon is vetoed because a sneaker
> > >> comes in on the right (later times), or because the muon came in too 
> > >> close
> > >> to a sneaker during the previous fill.
> > >>
> > >> I have not taken into account the increase in pileup veto due to the
> > >> muSCA/muSClo signals.
> > >>
> > >> In the attachment, I have the ideal NPP rate in red, and the actual PP
> > >> observed rate in blue, as a function of incident unkicked rate.  For our
> > >> current running conditions we havea a kick fraction of 0.63 and a 
> > >> rate of
> > >> 25 kHz, implying our beam is about 70 kHz incident.  This comes out to
> > >> some low 20s of kHz which is close to the 0.8 PP efficiency times the 25
> > >> kHz scaler 9.
> > >>
> > >> Ultimately we still have a lot of room to imprve if we could increase 
> > >> the
> > >> beam rate by a factor of 2, but after that the returns are 
> > >> diminishing for
> > >> our physical setup, and the maximum PP rate is only about 25 kHz.  
> > >> Mostly
> > >> I was just interested in seeing the maximum theoretical rate with our
> > >> setup.
> > >>
> > >> Best Regards,
> > >> Brendan
> > >>
> > >>
> > 
>