As some will remember i developed a laser range finder last year for my master thesis... it was a succes, got resolution of 1.5cm over 10 meter range. I got 87% for it and placed second in the philips lighting thesis contest for industrial engineering in belgium. I want to thank all and especially Colin for his help on the receiver end, without him and also Winfield Hill it would have costed me a lot more time and probably not have worked as good...
Now i am trying to achieve mm resolution with optoelectronic mixing as Colin also explained to me last year. I made a complete receiver were i inject a frequenty into the bias voltage of the avalanche photodiode. This injection occurs with a high voltage coupling capacitor right on the kathode of the photodiode with a resistor of 50ohm to the dc reverse bias power supply. The dc bias voltage is 207 V and the used avalanche photodiode is the C30902E. Next i used a 100K smd resistor to give sufficient signal at the input of my amplifier. The higher this resistor the better for the signal-noise ratio(for thermal resistor noise) (as noise increases with root of the resistance and the signal lineair) but too much would decrease the bandwidth with a total estimated input capacitance of about 20pf. Next i use a opa657 for now (i am going to change this to the opa655 but i am waiting for the samples) with voltage feedback(inverting amplifier) with a AC voltage gain of 2000 (bandwidth 200KHz with feedback capacitor) and DC unity gain (for exact offset at the output).The power supply is between 0 and
12V , and the signal is DC offsetted at 6V (7806 with 100K -> this give 50K small signal gain (2 times 100K parrallel)) after the amplifier i use a one stage RC low pass filter too get high frequentie noise and sum frequency out of the signal bandwidth.Now i injected between 30-60Mhz signal with a amplitude between
100mV-3V on the reverse bias with a laser send frequency 0-300Khz apart. On the output of my amplifier i clearly see the mid frequency of 0-300Khz with an amplitude of 8V pk-pk !!! if I use retroreflection and increase the injection signal pk-pk upto 1.5V. But the output signal is also very noisy => 1v pk-pk. I assume this is due the photodiode shot noise wich is also amplified 2000 times. This would easily be solved with using a tuned feedback but then rapid phase change would occur for a small frequentie error wich would give a large distance error (distance is measured through phase measurement).Soo with using retroreflection signal / noise ratio is positive, but on a white wall the signal is almost as low as the noise wich is 1V pk-pk!!!!
Too better understand the reason where the noise came from i put a spectrum analyser at the output of the photodiode and completly removed the amplifier. Then it seems that the noise is really in function of the DC reverse bias voltage(wich is perfectly normal as darkcurrent increases and thus shot noise with reverse bias) and especially in the lower frequency range 0-1Mhz. (It must be noted that the input impedance of the spectrum analyser is 50ohms soo there is not much signal voltage amplification) Also the DC reverse bias voltage was lower then 180 V no extra noise occured (in comparison with no dc voltage)soo this power supply i made with 2 zener stages and low pass filtering is for noise no problem.
The big question is now, how should i get rid of the noise without tuning , if u read this colin, did you also had much noise due the large voltage gain of your amplifier wich amplifies the shot noise and ifso how did you solve it???
Thx for any help on this one.