[Solid_ec] calculation

Mon Sep 11 00:17:25 EDT 2017

dear all

we have found the reason of my problem. As prof. feng mentioned, the PMT anode current is saturate for vertical test.
now we are doing the horizontal test. The input signal is smaller than vertical test, so we didn't use attenuator. here are the results:

pmt1000V(Gain=3.1E6)      QDC=500(most probable value)     pedstal=90

Npe=[(500-90)*0.1*1E-12] /[3.1E6*1.6E-19]=(500-90)/(3.1*1.6)=82

pmt1100V(Gain=6.5E6)      QDC=1000(most probable value)    pedstal=90

Npe=[(1000-90)*0.1*1E-12] /[6.5E6*1.6E-19]=(1000-90)/(6.5*1.6)=87

For horizontal test, NPE does not vary with HV, this is normal. So we can be sure that the problem is due to the vertical test signal is too large to lead to PMT saturation.
the number of event for two HV points just is 20. Our test is still continuing to get more events.

Therefore, npe would be 748 if the PMT is not saturated under 1000V in the vertical test. The npe would be greater than 748 if the PMT is saturated under 1000V.
after horizontal test, we will do vertical test again. we will test the NpE under the 900V. If Npe also is 750, we would ensure that it is not saturated under 1000V in the vertical test.

On the other hand, Beijing Hamamatsu test staff told me that the results of the two test methods （spe & sp/sk）will not be much different.And they think that the SP / Sk method is more accurate.

best regards,
chendi

--

申忱迪   Chen-di Shen
Department of Engineering Physics, Tsinghua University
418, Liuqing Building, THU, Haidian District, Beijing, China, 100084
Tel：86-15600915869 & 86-18101259887
email: shenchendi404 at 163.com

At 2017-09-10 18:05:23, "Cunfeng Feng" <cunfeng.feng at cern.ch> wrote:
Hi Xiaochao,

The key point for us is how to count the NPE. SDU_3 module light yield is about 400 PEs, it's far lower than THU new module (more 700 PEs). The main reason maybe the define of Gain is different.

The first attached picture is the SP and SK values for my ten PMTs which provided by Hamamatsu company, few of them used for our detector. The second form puts the sp and sk of Hamamatsu and our test result together.

We can find that our GAIN at 1500V is not proportional to Sp/Sk. We also provide the voltage for gain at 1E6, 2E6 and 4E6.  We use single photon electron method to define the Gain, not Sp/Sk. Which method is reasonable?  As we know, most particle experiment use single photon electron to define Gain instead of Sp/Sk.

Cunfeng

On 2017/9/8 19:56, Xiaochao Zheng wrote:

Hi Cunfeng:

Please see attached.  At this moment I don't think it's important to discuss what the value of b is and whether it is reasonable, although I think data shows the power of b works.  (We could test another PMT (perhaps SDU lab can do this) and see if it exhibit the same behavior. )

The important issue is how do we understand Chendi's shashlyk data.  Any idea?

Thanks,

Xiaochao

From: "Xiaochao Zheng" <xiaochao at jlab.org>
To: "cunfeng feng" <cunfeng.feng at cern.ch>
Cc: "申忱迪" <15600915869 at 163.com>
Sent: Friday, September 8, 2017 7:49:51 AM
Subject: Re: [Solid_ec] calculation

within certain error bars. 8.1 works pretty well for most of points.

From: "cunfeng feng" <cunfeng.feng at cern.ch>
To: "Xiaochao Zheng" <xiaochao at jlab.org>
Cc: "申忱迪" <15600915869 at 163.com>, "solid ec" <solid_ec at jlab.org>
Sent: Thursday, September 7, 2017 11:47:28 PM
Subject: Re: [Solid_ec] calculation

simple calculate:

1.1^8.1=2.16, Gain at 1100V should be 3.1*2.16=6.7>6.5(chendi list)(3.1is the gain at 1000V)

1500V : 1.5^8.1=26.7, =>G=3.1*26.7=82.7<102(chendi's)

So the relation not normally as G=a*V^b.

On 2017/9/8 11:23, Xiaochao Zheng wrote:

Hi Cunfeng:

There is a slight difference here:  When the curve is linear on the double-log scale, one could have:

  gain = a * (HV)^b, where a and b are constants.  Looking at Chendi's data (see the spreadsheet file on the minutes), the value of b seems to be about 8.1, which is likely to be the number of stages for this PMT. In this case, the increase in the gain per dynode is linear with HV, and the total gain is HV raised to the power of number of stages, and thus Chendi's results are reasonable and is consistent with what we expect.

At the meeting JP said what CHendi showed agrees with JP's experience.  He said for the Hall A HRS detectors, when the HV is increased by fixed amount (such as 100V each time), the gain increase by a fixed scaling factor (for example, doubles every 100V).  THis can be either explained using the single log scale function, or a power function as above.

So I think the problem may be elsewhere:

1) Can the PMT anode current saturate?

2) Can the attenuator that Chendi used behave non-linearly?

3) If we do a HV scan for the SDU3 module, is it possible to show the same (weird) behavior?   Can the SDU group do such a scan to prove the method to be valid? (That we should get a fixed Npe for different HV values)?

Thanks,

Xiaochao

From: "cunfeng feng" <cunfeng.feng at cern.ch>
To: "申忱迪" <15600915869 at 163.com>, "Xiaochao Zheng" <xiaochao at jlab.org>, "solid ec" <solid_ec at jlab.org>
Sent: Thursday, September 7, 2017 10:56:18 PM
Subject: Re: [Solid_ec] calculation

Chendi,

I agree with Xiaochao that the Gain-V relation is linearity under double-log axis as shown on the PMT specification sheet, not single-log axis. You should contact with the PMT company to explain why the Gain-V results is not linearity under double-log axis.

Your PMT Gain raises much more, it should be the reason the NPE decrease.

Cunfeng

On 2017/9/7 22:26, 申忱迪 wrote:

hello

here is my calculation.

best regards

--

申忱迪   Chen-di Shen
Department of Engineering Physics, Tsinghua University
418, Liuqing Building, THU, Haidian District, Beijing, China, 100084
Tel：86-15600915869 & 86-18101259887
email: shenchendi404 at 163.com

_______________________________________________
Solid_ec mailing list
Solid_ec at jlab.orghttps://mailman.jlab.org/mailman/listinfo/solid_ec

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/solid_ec/attachments/20170911/80dc9ae9/attachment-0001.html>