[Halld-pid] TOF PMT gain and rates

[Elton Smith]

Hi all,

Just a couple of observations, perhaps redundant with previous comments: 
The average current draw is readily computed from the rate (16MHz), 
pulse height (I take 0.5*minimum ionizing assuming interaction half-way 
into the scintillator and a pulse of 1 V/minimum ionizing) of 0.5 V and 
estimated triangular pulse with base of 50 ns. Then, the charge per 
pulse = 0.5 * (0.5 V/ 50 Ohms)*50x10^-9 s = 2.5x10^-10 C/pulse. Taking 
Beni's number of 16 MHz, we get 4 mA average current (at 10^8 g/s). The 
rate at low luminosity (10^7 g/s), we scale to get 0.4 mA.

The maximum ratings for average current  the XP2020 is 0.2 mA and for 
the R105343 tubes the maximum rating is 0.1 mA. The definition by 
Hamamatsu is that running at this maximum rate for 1000 hours will 
decrease the amplitude of the tube by about 50% [Note that there is 
considerable variation among tubes, so these values are to be taken as 
typical]. Now for 50% accelerator/detector operation efficiency, a 35 
week per year run corresponds to about 3000 hours. Thus operating at 
this average current will result in having to change HV values to 
maintain gain several times during the course of one year. At some 
point, the PMT itself will have to be replaced.

Ways to address this problem
1. Reduce gain of tube (either by going to pmt with less stages or 
adjustments in voltage divider). This can be compensated either by using 
an amplifier and/or low thresholds on the discriminator. It seems like 
the latest electronic developments tend to allow to go to low thresholds 
by controlling noise, so this may be an option.
2. Change size of counters to reduce actual rate per counter. I don't 
know if this is worth considering at this point, e.g. make all 
half-width also half-height and vice versa.
3. Reduce the pulse size for a typical minimum-ionizing particle (it 
does not need to be 1 V, what about 100 mV?)
4. Some combination of the above.

Now some comments:
1. Caution is required even at 10^7 g/s.
2. With no changes, even at 10^8 g/s it will be challenging to operate 
since the pmt gain will likely be changing during beam periods.

These notes intended for further discussion and feedback. Cheers, Elton.

On 9/19/11 3:32 PM, Paul Eugenio wrote:
> Hi Beni,
>
> Recall originally the plan was to use XP2020UR tubes.  The XP2020UR is a 12 stage tube with gains of 5E7.  The Hamamatsu R10533 which had an average gain of 1E6 (now 4E6) is the 10 stage version of their ultra-fast 8-stage R9779 tube.
>
> We did comparative test between the R105343 and the XP2020UR.  The timing, based on time difference were comparable.  We even showed that the  R105343 show slightly better resolution.
>
> Since the gain of  R105343 was 1.5 orders of magnitude less that the XP2020's,  we measured the counter efficiency as a function of hit location along the scintillator (0 - 100 inches).   The efficiencies of the XP2020 tubes were fine, but we found that one of the  R105343 tubes (the lower gain tube) was less efficient(97% for hits 95" away) at detecting the hits which emanated at the opposite end of the scintillator.  At the time, we had two  R105343 tubes, one with gain 1.1E6 and the other with gain 1.5E6.    These tubes were one the first of this type made by Hamamatsu.  Also at the time, they promised an average gain of 2E6 with 1E6 minimum.  Now, their average gains are 4.2E6 of which we have a requirement that the minimum gain for selection be 2E6.    This is the so called "FSU push for high gain PMTs".
>
> The average anode current for the R105343 listed on the spec sheet is 0.1 mA.  You say that we can not operate the PMT at a gain that results in currents that the close to 0.1mA.   Are you sure that you are not off by a factor of ten?
>
>
>
>
> --
> Prof. Paul Eugenio
> Florida State University
> Department of Physics
> Tallahassee, Florida,  USA 32306
>
> (850) 325-0314
> eugenio at fsu.edu
>
>
>
>
> On Sep 19, 2011, at 9:11 AM, Beni Zihlmann wrote:
>
>> Hi All,
>> last Friday the Hamamatsu representatives were here at the lab and
>> we had a short conversation about the TOF PMTs. From this short
>> meeting I conclude that we need an extensive discussion about the
>> PMTs.
>> One important issue is that here at Jlab we do not understand why
>> you at FSU push so hard for high gain PMTs. Even at the low intensity
>> beam of 10^7 photons the inner most paddles will see rates of about
>> 1 MHz and consequently the current in the PMT will be substantial if
>> the gain is high. We can not operate the PMT at a gain that results in
>> currents that the close to 0.1mA. Such currents will diminish the intrinsic
>> gain of the PMT substantially in a short time (50% in 1000 hours operation).
>>
>> Can you estimate how many photo electrons you get for a MIP that passes
>> through the center of the paddle and deposits the mean energy of a landau
>> distribution? Having signals of 1000mV for such events will results in too
>> high currents.
>>
>> I would like to put this issue on the agenda for next meeting.
>>
>> cheers,
>> Beni
>> _______________________________________________
>> Halld-pid mailing list
>> Halld-pid at jlab.org
>> https://mailman.jlab.org/mailman/listinfo/halld-pid
>
> _______________________________________________
> Halld-pid mailing list
> Halld-pid at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-pid