From nathanw@msu.edu Thu Mar 2 11:36:08 2023 Date: Thu, 2 Mar 2023 16:35:33 +0000 From: "Whitehorn, Nathan" To: "Edmunds, Daniel" Subject: Re: PMT signal range - is this it ? On 3/1/23 17:11, Dan Edmunds wrote: > Hello Nathan, > > Is the following more or less your understanding of > the range of signals that the Disco-Kraken PMT ADC > will need to process ? > >       Thanks,  Dan > > > 1. The big 100 nsec signal is about 5000 photons arriving >    over the 100 nsec period.  If each of these photons >    results in one photo-electron at the PMT cathode   and >    if we need this to result in a 0.81 Volt pulse across a >    50 Ohm load this will require about 16.2 mAmp of current >    to flow for about 100 nsec  or  about 1.62 nano-coulombs >    of charge.  The required PMT gain is: > >     1.62 nano-Coulomb   X   6.24 x 10E18  electrons / Coulomb >    ----------------------------------------------------------- >                  5000 original photo-electrons > > >    which gives a required PMT gain of  2 x 10E6 >    Note that a gain of 2 million is a rational value for >    the Hamamatsu R14374 R14689 PMT. > This is basically right. Some nits: - Keep in mind that they are not evenly distributed across the ~ 100 ns, but peaked at early times. This is about a factor of 2 correction to the maximum current. - The PMT also saturates at a bit above 30 mA. - The actual peak doesn't always involve completely useful information for the physics -- the useful information is the rising edge and structure on the falling tail. For IceCube, we usually cut out the area around the peak of the waveform, within a factor of 2 of the PMT saturation current, in analysis. If it turns out we can't get both in the end, we want to prioritize single photons over the peak of ultra-high-energy waveforms. To give a sense of what these look like, I have attached some sample waveforms from one of the highest-energy neutrino events IceCube has ever seen (panopticon.pdf). Red regions have been excluded in analysis and the timing for far-away modules will be more compressed for us as a result of reduced optical scattering in sea water; the very nearby, very highest-amplitude ones won't be that different. Each plot is a different PMT, the Y axis is in PE/50 ns, and the X axis is in microseconds. I will try to get you some P-ONE simulation when I can. > > 2. The smallest signal is one photo  making I assume one >    photo-electron and if the resulting signal lasts for >    6 nsec  (which is the number that I've heard)  and if >    the PMT gain is 2 x 10E6  then this 6 nsec long signal >    will have an amplitude across 50 Ohms of about 2.67 mV. > >     2 x 10E6 electrons   X   1.60 x 10E-19  Colombs / electron >    ------------------------------------------------------------ >                        6 nsec > > >    gives a current of  53.4 uAmps  or  2.67 mV  across 50 Ohms. I think it might be a little smaller than that, actually, since, looking back at data, 8 ns might be better than 6 ns as an estimate for the pulse width. I've attached a plot for gain 10^7 with the pulse shape we were using in simulation (pmtsample.png). The ADC transfer function used in this model is a single-pole LPF at 100 MHz and does not quite match lab data -- the pulses in data are narrower and taller at the ~ 10-20% level -- but it is reasonably close. I have attached also a sample PMT pulse from a similar PMT into the ADC evaluation board in the lab (pmtpulse2.png). This is in 16-bit mode, with 4 ns sampling, and the gain was set so that a single photoelectron was ~ 1% of the ADC's dynamic range. > > 3. At a PMT gain of 2 million the biggest signal is >    0.81 Volts  and  the smallest signal is  2.67 mV >    which is a ratio of  303. > >    If the big signal is right at saturation 4095 ADC counts >    then the small signal will give about      13 ADC counts. > >    So we must design the Disco-Kraken to provice an >    ADC noise and cross-talk well below 13 counts. > Yes, I think this is about right. The screenshot from AD's tool (I have misplaced the actual data -- the plot was scraped from a talk I gave) corresponds in 12-bit mode to ~ 50 counts amplitude and S/N and crosstalk look pretty good (channel 3 is plotted as a crosstalk canary). I think they would still look good with gain scaled down by ~ 3-4 and the same noise, which is at least promising that this kind of performance should be achievable. -Nathan -- Nathan Whitehorn Assistant Professor Department of Physics and Astronomy Michigan State University Biomedical and Physical Sciences 3220 East Lansing, MI 48824 (517) 884-5563 [ Part 2, "pmtsample.png" Image/PNG (Name: "pmtsample.png") 31 KB. ] [ Unable to print this part. ] [ Part 3, "pmtpulse2.png" Image/PNG (Name: "pmtpulse2.png") 141 KB. ] [ Unable to print this part. ] [ Part 4, "panopticon.pdf" Application/PDF (Name: "panopticon.pdf") ] [ 195 KB. ] [ Unable to print this part. ]