Photon Transfer Curve

Hi Jens,

Yes, but in the case that the read-noise increases with a longer
exposure time (and the brighter images were captured with longer
exposures), it would maybe pointing to a wrong darklevel acquisition.

Agreed. But in this setup, the integration time was held constant and
the exposure time (the time the light source was on) was increased to
get more signal. So, the dark level should be pretty consistent.

I would expect the characteristic steep fall off of the curve at
saturation.

Yes, if the converter is saturating bevor the sensor gets close
to the fullwell.
Otherwise it shouldnt fall off so quick, dependend on the sensors
characteristic. Ok, 760e noise_rms is nothing against 500ke fullwell,
so it wouldnt flatten the curve a lot, too.

So if I'm at full well and the convertor isn't saturating, the curve
still won't drop off quickly?

--------
If I can go back to something you mentioned earlier about non-linear
response: you mentioned that non-linear pixel response (output voltage
/ incident photon) could also affect the photon transfer curve. Well,
I found an interesting article that describes this in detail. In fact,
they derive a method to determine things like conversion gain of a
non-linear sensor that appears to be quite accurate. Maybe it is of
some interest to you:

Bedabrata Pain and Bruce R. Hancock, "Accurate estimation of
conversion gain and quantum efficiency in CMOS imagers", Proc. SPIE --
Volume 5017, pp. 94-103, 2003

As always, Thanks.

-Patrick

Hi Patrick,

Agreed. But in this setup, the integration time was held constant and
the exposure time (the time the light source was on) was increased to
get more signal. So, the dark level should be pretty consistent.

This is a good method.

Yes, if the converter is saturating bevor the sensor gets close
to the fullwell.
Otherwise it shouldnt fall off so quick, dependend on the sensors
characteristic. Ok, 760e noise_rms is nothing against 500ke fullwell,
so it wouldnt flatten the curve a lot, too.

So if I'm at full well and the convertor isn't saturating, the curve
still won't drop off quickly?

I had to measure it again, noise drops off quickly too.
I had seen it before, but forgot it ;-/.
If the pixel-charges overcomes the potential-barriere to the ABG,
the pixel-voltage is something like clamped to the Voltage of the ABG.
With higher currents to the gate, the resistance decreases and noise
drops. This would be my explanation.
Ive measured it on a ABG CCD with only ~7ke fullwell, but it should
be the same with higher fullwells.


--------
If I can go back to something you mentioned earlier about non-linear
response: you mentioned that non-linear pixel response (output voltage
/ incident photon) could also affect the photon transfer curve. Well,
I found an interesting article that describes this in detail. In fact,
they derive a method to determine things like conversion gain of a
non-linear sensor that appears to be quite accurate. Maybe it is of
some interest to you:

Bedabrata Pain and Bruce R. Hancock, "Accurate estimation of
conversion gain and quantum efficiency in CMOS imagers", Proc. SPIE --
Volume 5017, pp. 94-103, 2003

Thank you Patrick, the whole Book looks very interesting.
By now, i am no more involved into these things, i think this became
obvisible.

Best regards,
Jens