[SOLVED] Selecting a Charge Sensitive Amplifier (CSA)

[bilal_oct]

I am wondering which charge sensitive amplifier is suitable for detecting the charge between 1 x 10^5 electronics to 135000 x 10^5 electronics. Anyone has idea ?

 

[Dominik Przyborowski]

20 bits or 124dB of dynamic range... I don't know any ASIC design which can do this even if it has switched gain, not to mention about COTS ones.

For what application You need it?

 

[erikl]

Means a resolution of 10µV out of 1.35V - on a 16 fF capacitor, if really electrons are meant. Surely not feasible, not with the required resolution. A single electron - probably tens or hundreds of them - would disappear in the noise floor.

 

[dick_freebird]

This is not an impossible problem given that you know
the constraints / application. Cryogenic-cooled imagers
operate down to 100 e- noise floor (across pretty long
frame times with "well capacities" in the 10K-100K e-
range. Supplement this for range with some sort of
hybrid scheme (when well gets to (say) half capacity,
increment counter and reset; ADC residue at end of
frame and add upper bits from counter). Strict 20 bit
linearity, maybe not - some likely major-carry type
DNL problems I'd expect - but you can get there.

Room temp noise floor being one key question for
non-cryogenic apps.

 

[BradtheRad]

In case your application has to do with this... It reminds me of the charge-coupled device which is used in detecting light. Low temperatures enhance detection.

Also a photon multiplier tube. I've heard that it detects a single photon, and cascades that one event into many photons, or a voltage, etc.

 

[bilal_oct]

Consider that the charge equals to 135000 x 10^5 electrons is produces at the output of MAPD (milti pixel avalanche photo diode) and the objective is to detect 100 electrons per ADC channel then which CSA and ADC is suitable for this purpose ?

 

[Dominik Przyborowski]

If I understand correctly You have APD matrix with 135k pixels. Your signal is a current spike with 100ke- of total charge, or this 10^5 is an electron multiplication ratio?
This 100e- is your typical signal from single pixel APD, or it is a number of primal electron-hole pairs generated in APD before multiplication?

Precise your question: what is your typical signal from single pixel APD, expected dynamic range, maximum acceptable ENC (Equivalent Noise Charge) and expected counting rate?

In general you are able to make readout chain for it using fast and low noise opamp as CSA, simple CR-RC^n shaping (depending on counting rate with APD tail cancellation or without) followed by ADC or peak detector and ADC. All parameter of readout chain can be simple calculated from well known formulas based on kind of filtering (order, center frequency, etc), sensor capacitance, CSA gain, input noises of opamp, to obtain best solution.

//edit: and of course last question - ASIC solution or COTS one?

 

[bilal_oct]

The actual charge at the output of APD when one pixel of APD is fired is 1 x 10^5 electrons, here 10^5 is the electron multiplication gain of the APD. In total there are 135000 pixels on the detector MAPD, which means there will be 135000 x 10^5 electrons if all the pixels are fired. Which type of CSA and ADC can detect the number of photons between 1 (equivalent to 1 x 10^5 electrons) and 135000 (135000 x 10^5 electrons)? The MAPD is capable of detecting 1 to 135000 incident photons.

 

[bilal_oct]

@ erikl and others, my objective is to detect 10 x 10^5 electrons (which corresponds to 1keV) and integer multiple of 10 x 10^5 electrons up to 135000 x 10^5 electrons. I just calculate that with 16 fF the voltage will be between 100 uV to 1.35 V, which needs a ADC with (1.35 V / 100 uV = 13500) channels. A 14 bit ADC can provide 16384 channels. So, the requirement for ADC is to detect an analog signal which swings from 100 uV to 1.35 V using a 14 bit ADC. Is there any commercially available ADC from Analog or other manufacturers which can meet this requirement ?

 

[bilal_oct]

Which CSA is most appropriate to detect the charge from 10 fC to 10^6 fC ?