Limitation of Detectors

[gerhard]

Dear forum,
I am new in Gammaspectroscopy an have a lot of questions.
i am in trouble to understand Merek Dolleisers manual to PRA 15.
He wrote on the first pages, that there is a limitation for around 1000 pulses per second to the limitation of 48000 samples.
But, newer ADC or soundcards can sample more then 192000 samples.
So, ist the limitation of 48000 samples due to the limits of the Crystal or the PMT (Maybe the deadtime ?).
If it is so, ADC adapters with more than 48000 samples are a waste of resources ?
Second, Marek Dolleiser recommends a low pass filter in the left channel to the ADC converter.
So, how low ? 3000 khz are right ?
Or, is in GS a built in low pass filter in left channel allready included, so i do not worry about. ?

Than You, for answer

[Sesselmann]

Hi flexit12,

Your assumptions are quite right, the limitation of 1000 cps applies to 48 kHz sound card sampling only, and it comes about because PRA uses 16 samples to shape a pulse regardless of how fast the sample rate is. With 192 kHz sampling you can count around 4000 cps. lets do the math..

1 sec / 48000 samples * 16 samples = 0.000333 seconds or 333 µs

So if you divide 1 second by 0.000333 you get 3000 but realistically with random pulses you can't sample at that rate and the best you are going to achieve is around 1000 counts before you get pulse pile up (PPU). PRA filters out badly formed pulses caused by PPU, so at high count rates you will have high counting loss.

The actual pulse from the crystal when amplified by the PMT is in the order of a µS but this is too fast for a slow sound card, and the chances are most of the pulses will come and go between samples and if you were lucky enough to catch one, how would you know that you actually caught the peak?

So the GS-1100 amplifies the pulse a little and passes it through a low pass filter so it takes around 100 µs from the first rise until it falls below the baseline again, tis pulse length is also a factor of the sound card impedance, so if you look at the pulse on a scope and use a 50 Ohm terminator you will see a much shorter pulse.

Steven

PS: Hi flexit12 sounds a bit silly which is why we prefer real names on this forum, you can update the usename in the UCP.

[gerhard]

Thank You, very much, Mr.Sesselmann for Your explanation.Now i see clear.
Sorry for my user name,but when trying registration,my name was not accepted and i have a password and username generator and he did this job.So i try to change it again,perhaps it works.
Another thing ist, that i found,that crystals are different in resolution and sensitivity.
So, i found,that smaller detectors have a better resolution an less sensitivity and bigger detectors have less resolution and better sensitivity.
It ist not much, but it seems to me. I never read about this, so maybe i am wrong.It belongs to NaI (Ti) sensors.
What do You think about ? Perhaps the best world of both is a 2 inch sensor ?.

[Neil.]

Gerhard,

It is good to find you participating on this forum and though your questions were directed to Steven I though I would add what I can.

I have tried various sized NaI:Th crystals and have not found any association between size and resolution, in fact my smallest crystal tends toward lower resolution though this could be due to the age of the crystal which was produced in the 1950s. One of my larger crystals has a built in absorber for energies less than about 50keV which further confuses this issue with my equipment.

Indeed, larger crystals are more sensitive but be careful with what is meant by sensitivity. If interpreted as sensing more events for a given source then this is correct but as is the case in many types of instrumentation this type of increased sensitivity also results in increased noise. The noise I reference in this case is background radiation and good use of shielding is a way of reducing this.

It is generally accepted that larger crystals will respond to higher energy gammas due to the increased travel distance through the crystal which increases the probability of an interaction but I have not found this important for most of my measurements which are below 1.5 MeV. Even the smaller crystal will give some peaks in the 2 to 3 MeV region.

I am using home built probes and have found a big factor affecting resolution is the choice of photomultiplier with the best around 6.5% and the worst around 15% (a used tube with brown discoloration indicating probable age or abuse). Most new tubes I have tried give 6.5 to 8%, The best resolution seems to happen about 5 to 10 minutes into the accumulation and then gets worse with longer time perhaps because something is drifting with time or temperature.

All that being said, I am getting good results using a 40 X 40 mm crystal with a 1.5 inch PMT which is more to the point of your question.

Neil

[gerhard]

Dear Neal,
Thank You very much for helping me.
I also come to conclusion, that a very good interaction of all parts is necessesary to get a good resolution, especiallly with NaI:Ti crystals.
Theremino describes on its Software manual this problems and what to do,to get FWHM better.
The problem ist,that i dont understand most of this electronical advices, but i am learning.
A perfect Crystal and a perfect PMT ist a good start.
Maybe, the preampifer should built near or on top of the Resistor-divider of the PMT and not together with the high voltage source.
A resolution of 6.5% with 137 in Your homebuilt probes ist really more then perfect. More than most factory built probes can realise.
i have to apologize for my bad english, in school i had to learn latin and greece.... (not usefull, as i see now).

[luuk]

Hello Gerhard,

A small or big crystal/detector has nothing to do with resolution=quality of the crystal/detector.
It all depends on the right crystal material, pmt, treatment of the crystal, and final wrapping and fitting in the housing
If you look here in the forum in the in the spectrum section under: Test of a 4x4x16”detector, there it is obvious that even a very large NaI(Tl) detector can give a great resolution.
You can improve for a little the resolution from a detector in a positive way by choosing a larger pmt than the detector e.g. 1”detector on a 1.5” or 2” pmt will give overall a better resolution than on a 1” pmt, the main reason for this effect is that the Cathode of a pmt is not homogenous special around the edges, the centre is mostly much better that is why the resolution is better with the small crystal on the centre of a larger pmt.

Regards,
Luuk

[gerhard]

hello luuk
Thank You for this information. A larger PMT, this seems obvious to me.Whenever building a detecor, ich will consider this.

[ChrisRadium2023]

Hello,
Could you please tell me if I got it right : so in the end if I bought a 192 kHz sound card instead of the 48kHz I currently have , will it help to improve the resolution or just allow a higher cps rate ? or both? as the number of bits/pulse is always 16 ....?



CHris.

[iRad]

Chris - Now that you're getting decent spectrums, I don't think you will see that much of an additional improvement with a super high-end sound card. I did not see much difference myself between the higher end cards when used with Theremino or PRA, although it's been years since I have even messed with them. Every PC or laptop I've had since then has had a good quality soundcard, or I've used the GS-PRO with its own quality USB sound card driver and interface built-in. The advantage is using only a single cable for both power and I/O. I used to modify and sell some of those inexpensive Chinese cards to reduce noise and carry 5V power to Theremino PMTs adapters, but that took up too much of my time and was always a hassle consistently finding a quality card I could modify. I was greatly happy when Steven added an internal USB soundcard to the GammaSpectaculars.

I think the software does a decent job of rejecting noise and sorting valid pulses once you have a good clean properly working 48kHz sound card or better. I believe higher sampling rates are more of a limit of using the sound card interface itself. At this point, it's more about a building (or finding) a better detector for your application or energy of interest, building a better shield, testing geometry, and consistent environmental controls over duration of measurements.

Regards, Tom

[ChrisRadium2023]

Thanks very much Tom,

thats very informative !! thanks a lot.

In fact this afternoon I went to the music shop where they sell these kind of usb sound card or preamp mic interface.. and i got a "scarlett solo Focusrite" @ 192 kHz and playing with both now at home on my laptop with Theremino. And I noticed the following :

With the 48Khz or the cheap chinese cards I could never see the shape of the pulse the same way my oscilloscope did. It always seemed to be around 50 to around 100uS long....
If I set theremino @ 192 or even 384kHz , then the pulse shape turns into a nice clean sinusoidal sort of wave sampled into numerous points ( at least much more than at 48kHz). And still I can get descent spectrum but smoother when I set @ 384kHz (with the 48kHz card).

WHen I use the 192kHz card its completely different and this time I'm able to see the same shape or a very close representation of the pulse I can see on the oscilloscope. But there is something odd : the spectrum is shifted towards the high energy keV. So its very likely that i'm not yet able to set the parameters properly yet and I will have to work on it . I cant say yet if the new acquisition is worth it or not.

This 192kHz card uses a usb cable that doesnt have a ferrite filter (unlike the 48kHz) and there's more noise compared to the 48kHz so I will buy a ferrite filter and check the effect on it.

I will make a usb filter to clean the usb 5V voltage supply and see if it improves the results.

Here is the shape of the pulse with a 48kHz card set @ 384kHz on Theremino :

And the same 48 kHz card at 96kHz via Theremino : there are much less point on the pulse

And here is the pulse with a 192 kHz card : the pulse is the shape that appear on the oscilloscope (I dont use preamp yet ; only a crude CRRC filter as a shaper)

and now the pulse with a 192kHz card at 384 kHz via Theremino :

the pulse contains more points

to be continued..

Chris.