Bill Sloman said:
Modulating your light source is an excellent way to distinguish it from
stray light, but enough stray light can always saturate your detector and
lower levels can still generate enough random noise to make it quite
difficult to detect your modulated source - narrow-band interference filters
can prevent most of the stray light from getting to the detector in the
first place.
Dual gate MOSFET can be attractive as front end amplifiers if you are
modulating at high frequencies, but when I last looked at them (which was a
long time ago) the 1/f noise corner tended to be close to 1MHz, and the DC
stability was rotten.
For most purposes a good FET-input op amp is better, and much easier to
use - the Burr-Brown (now Texas Instruments) OPA655 comes to mind.
If you are working with modulation frequencies of the order of 50MHz, an
inductor can be an attractive load, but you have to make sure that your
inductor's self-resonant frequency is higher than you modulation frequency -
the interwinding capacitance for a single-layer wound inductor is usually of
the order of 1pF, equivalent to a parallel impedance of only 3k at 50MHz.
Alavanche photo-diodes can detect single photons if operated in the Geiger
mode, and can provide a very prompt output, but the recovery time is
relatively long, which restricts them to looking at very low light levels.
Photomultiplier tubes are a lot more flexible, and can offer steady state
gains of up to about ten million, where the avalanche photo-diode is limited
to about 150 when not operating in the Geiger mode.
im using a visible light laser so i can see its pointing at the right
target, not sure how much a narow band filter would help me but snds
interesting, id only thought of infra red filters till now, main problem i
think is signal to noise/interfenernce at the detector amplifier stage, i
might look and see how much noise is being cuased by natural light,
obviously if i shine a torch on the sensor it swamps it as it has a high
resiatance in the hv supply, but dont expect it to have to cope with that.
initialy i was using a low resitance load to get good wide frequency
response and using a rf bipolar tranny as they best noise figure at low
source impedance, that wasnt good enough so i had to go for a single
frequency and tuned input stage to get the impedance up and the performance
diference i get now from a bf998 dual gate mosfet is quite remarkable. im
measuring very small delays so phase change with signal level is a problem
such as ive found are cuased by an op amp with negative feedback, but
otheriwse that would be easier.
im just using a 15 turn (found by trial and error) air cored inductor with a
tuning slug and no extra capacitance to resonate, so this is quite an easy
solution at this frequency. i dont have any tables to work out capacitance
of inductors, but as you say as long as the srf is higher than what u using,
but if you can use a single reasonably high frequency i think a tuned input
stage is an optimal solution. although i still have a 100k resistor to limit
the q as it was unstable but i think that was becuse i had a tuned drain
load too wich i since did away with.
a gain of 150 is still very significant to me, however im not sure if the
range inbetween this and geiger mode is usuable i think its just rather
unstable, but ive operated it what must be close to its vbr (250v ish) and
it seems to perform very well indeed, infact the latest problem was that a
small un noticable amount of light was being reflected of the
mirror/lense/case and finding its way back to the detector, so il have to re
orient them. (not to mention the interference from the 250v inverter)
APD are quite expensive tho, but i expect PM tubes are more so and bulky too
(this is hopefuly going to be a small hand held device), and do they still
take a while to recover from exposure to excesive light levels ? cant see
why APD are so expensive maybe cos they just not used in quantity yet. not
sure how noise level compares, im using the C30902E from RS wich specifies
gains greater than 150 or upto 250 or greater.
Colin =^.^=