[Halld-pid] PMT test results question

[Alexander Ostrovidov]

> Hi Paul,
> 
> sorry that I missed the meeting. Mark explained to me the situation and 
> I have some
> remarks and questions because I think something is completely wrong with 
> the ADC
> spectrum using the Am source.

Hi Beni,

Thanks for taking a look into this test. Let me start by saying that things
are better today than yesterday when we were getting a factor of 4
inconsistency in anode sensitivity between tubes. What apparently has
happened is we overloaded NIM crate power supply with too many modules.
For whatever reason, this led to variations of about 20-30mV in baseline (offset) 
of the signal which I didn't notice initially. This, obviously, didn't go well with 
a 25mV discriminator which is used to form ADC gate. We pulled out a few 
unused modules and the offset went back to zero. Now all tubes measured
so far show relative consistency within 30% when their readings are normalized
to their listed anode sensitivity. Check the last column here (standard gluex
user/pass): http://hadron.physics.fsu.edu/TOF/database/PMT_am.html

I absolutely agree with you that one would expect a relatively narrow gaussian
from alpha source. Unfortunately, that's not what I see  with the scope. Signals
(straight from PMT into the scope with 50 Ohm terminator, and I put the scope
into 5sec accumulation mode) are spread from about 50mV to 300mV with
maximum closer to the lower values. Their FWHH visually varies from 5ns to 15ns. 
It's no wonder with such variations that ADC spectrum is nowhere near to being 
monochromatic. 

I couldn't find anything wrong with ADC setup. ADC gate is about 160ns.
Signal comes about 80ns after the start of the gate and nicely fits inside it.
Signal visual spread on the scope roughly corresponds to the spread of
the ADC spectrum. The most probable signal corresponds approximately
to about 10 initial photo electrons at the cathode.

At this point, I'm  ready to suspect that something is wrong with Americium
source itself. I can see a small silver metallic button on the other side of
the scintillator but I'm not sure if there is something that stops (or slowdowns) 
alpha particles before they enter the scintillator. May be, what I see is such
slowed down alphas. Or a conversion of 60keV gammas instead of alphas 
which are also emitted by Am241.  Or betas from its decay products. 
At least, what I do know is:

a) The rate, from whatever it is, does go up to 2000Hz (with 30mV threshold)
comparing with 10-30Hz rate from PMT with nothing inserted, and 20-40Hz
rate when a comparable size scintillator (with no source) is attached to PMT.
Just from the probability, it's very unlikely that we trigger on noise or cosmics.

b) ADC spectra do look differently in these 3 cases Take a look here
http://hadron.physics.fsu.edu/TOF/pmt_am/ZQ2263/
File noscint* is with nothing attached to PMT, file nosource* is for
a passive scintillator, and am10x* is for a scintillator with Am source.
>From the difference of these spectra as well, it also doesn' t look like we trigger 
on cosmics or PMT noise that much.

We'll take a closer look at the Am source with Paul on Monday. Meanwhile,
even if this is not a pure Americium source, I guess, I can still treat it
just as a consistent, albeit puzzling, source of light. Therefore, relative
comparison of PMT anode/cathode sensitivity can still proceed as long
as light illumination is stable from tube to tube. At least, current 30% is
much better and almost acceptable comparing with factor of 4 (or even 
factor of 10 from   UV LED rates) I had before.

Sasha


> 
> 1) the Am 241 generates 5.6 MeV alphas. This alphas are mono-energetic 
> and should
>       produce a lot of light in the scintillator and therefore a very 
> large signal in the PMT
>       if you put this little scintillator directly in front of the photo 
> cathode.
>       -> so have a look on the scope how the signals look like. They 
> should be huge if you
>       use 1750V
> 
> 2) Because the alphas are mono-energetic the resulting energy spectrum 
> in the ADC
>       should be a peak. Or maybe two peaks depending of the pureness of 
> the Am241
>       source and the resolution of the PMT/scintillator system. This 
> peak should be rather
>       sharp, high above the pedestal and Gaussian. If any tail this tail 
> should go to lower
>       energy not higher energy. These are the main reasons why I think 
> the ADC spectra
>       you see are completely off and something in the setup must be off. 
> My first thinking
>       is that you look a noise and the actual signal is at much much 
> higher ADC values.
>       -> again that is why looking on the scope at the signals and in 
> particular large signals
>       should be done.
> 
> 3) Because I say that the Am241/scintillator source produces so much 
> light it might be that
>      1750 is far to high Voltage, and for sure you do not need 
> amplification.
>      Note: 5.6 MeV from the alpha particle is deposited in the first 1mm 
> of the scintillator and
>      the scintillator in total is not that think so there is hardly any 
> attenuation and all the light
>      is seen by the photo cathode. => lots and lots of light => huge signal!
> 
> 4) you trigger on the same signal as you are looking at. That means you 
> should never see
>      a pedestal. However the data show a pedestal => that does not make 
> sense.
> 
> 5) the peak in the ADC is at a similar location above pedestal as the 
> single photo electron peak
>      is at the same HV and amplification. Again that does not make sense 
> because from the
>      alpha source you should have lots and lots of light.
> 
> 
> I hope this gives you some ideas of what you can look at in your setup 
> to see
> what might be going wrong.
> 
> 
> cheers,
> Beni
> 
> 
> _______________________________________________
> Halld-pid mailing list
> Halld-pid at jlab.org
> https://mailman.jlab.org/mailman/listinfo/halld-pid
> 
>