Re: FDC/CDC ASIC meeting minutes

[Yves Van Haarlem <yvhaarle@ernest.phys.cmu.edu>]

Hi Gerard,

Thanks for your feedback, here are my answers:

On Fri, 5 Sep 2008, Gerard Visser wrote:

> Hi guys,
> 	Yves, can you clarify, the gains listed in table 2 are determined 
> from the scope photos shown in fig. 8, right?
Right


> Where exactly on the histogram 
> for instance in 8(b) did you call the peak amplitude?
I took the position of the highest peak (= the argon peak).


> It seems to me rather 
> subjective, and perhaps there is no reason to expect this procedure to 
> asymptote to a constant gain as you lower the bias voltage.
Those spectra correspond pretty well with the fADC spectra taken 
with the Fe-55 source. The only thing is that I fit a gaussian by eye. 
So fitting a Gaussian the right way (which is as far as I know not 
possible on the scope) can change the gain values a bit but not the trend 
as a funtion of HV.



> I agree with your 
> conclusion that the plainly evident change of preamp pulse shape at higher 
> bias voltages affects the effective gain. Is it not possible that the shape 
> continues to change at lower voltages? (You could capture a set of waveforms, 
> normalize _each_ individually to same peak amplitude, and then average them, 
> and compare the results of this procedure at different bias voltages. This 
> could provide some quantitative indication of how much the shape is changing 
> from saturation of the preamp or whatever other effects.)
Yes it could be (I was thinking that at low HV the signal to noise comes 
into play and change the signal shape). This behavior needs to be 
understood. Lets see if I understand your proposal: you want a set of 
waveforms (let's say 10) for each HV setting, normalize each waveform to 
one and than take the average of the 10 waveforms. Shape-compare those 
waveforms for different HV. Those traces need to be taken with the scope, 
saved and than analysed offline. I will look into this. I will also try to 
measure the gain of signals (cosmics) wich are in the linear range of the 
preAMP (<500 mV amplitude) vs HV.


> Finally, one other 
> point, the scope trigger threshold in fig 8 is pretty high (compared to the 
> noise level), it seems. This must have some effect on the effective gain that 
> you get from reading these histograms. Can you try to repeat at much lower 
> threshold and see if it gives better results?
I can try that but for the moment I do not have that nice scope 
anymore...but I can set that threshold on the discriminator. 
I used the Fe-55 source to do the measurement and the Ar-peak is wel 
above the threshold.


> p.s. Are the histograms (I presume in a narrow time window) centered on the 
> peak time of the pulse in each of the 6 cases independently, or at fixed 
> time, not therefore picking up the peak amplitude? I presume/hope you have 
> done the former.
I do not remember the time window that was set (it was not very narrow) 
but the scope displays the measurment every time it measures an amplitude 
and that looked correct.


> p.p.s. One more thing, the double-humped spectra, did you guys discuss could 
> it have anything to do with the double-humped spectra you see from the preamp 
> out at higher bias voltages in fig 8? Therein, it looks like some feature 
> saturation behaviour of the preamp ASIC...
We did not discuss that but it is difficult to say - it disappears after 
shaping but is interesting.


Cheers,
 	Yves



>
>
> Simon Taylor wrote:
>> Hello.
>> 
>> Fernando, Yves, and I had a meeting on Wednesday to discuss the FDC and
>> CDC ASIC documents.  The minutes of this discussion are posted on the
>> GlueX wiki at
>> http://www.jlab.org/Hall-D/software/wiki/index.php/CDC/FDC_ASIC_Discussion_Meeting%2C_September_3%2C_2008
>> 
>> Simon
>