Low power dicrete amplifier

I need good point on low power discrete amplifier with bandwidth up to 20 KHz , on 2 -3 BJT or FETs running on variable power 1.8 - 3 V with good bias stabilization in temerature range -40 +85 C, gain 30-40 DB ,output range 0.2 -:- (Vcc - 0.2) , Icc < 0.1 ma.

The amp is needed for subdollar design , but i have failed to find cheap amplifier IC (even 50 cents is expensive ) .

Look for cheap opamps! They might do the trick!

maxim and analog have much such op.

Suitable opamp costs about 0.50 $ and it is quite expensive because of project requirements .

essential information: what is the source and load impedance?

From your power supply limits, BJTs are the way to go if the frequency response does not need to go very low. Otherwise, the coupling capacitors and bypass capacitors will run the cost up.

You may be stuck with operational amplifiers.

Actually I have misssed some information in requirements . THe amplifier will be used as front to photodiode based fiber optic security loop . So signal is to be digitial . So, amplifier output will work towards
mcu input .

I have simulated the circuit (see attachment) based on ON's bias stablization chip schema with enable and get instablity in range 100 mV (measured on Q1 C) for -40 +80 industrial temperature range .
Well , as signal will be pure digital , not of that stability will be needed , so I can replace with serially connected resistors and diodes - providing
floating bias for biasing Q2 transistor . That variant gives Q1 C voltage chnage in range 0.5- 0.6 for -40 - +80 temperature range .

I1 is going to be replaced by photodiode . But before photodiode dark current must be taken into account is for mentioned temeprature range it can vary a lot . It could be case that biasing scheme will not be able to compensate diode dark current at high temperatures . So amp gain is to be reduced - eitehr by diode input current reduction or Q1 coll resistor decrease - that is unwanted i would like to keep number of discretes as as low possible .

From other side I did not evaluate the photodiode and fiber optic cable completely . I am plannig to use plastic cable with wave length 660 , so cheap didoes can be used . Does anybody have price information for photodiodes in that wave range ? The BPW-34 is good one but little bit expensive due to project requirements mass production .


Thanks for help.

MCP6001 /1.8-5.5V /140uA /1MHz
costs $0.25!

What you really need is a transimpedance amplifier (converts current to voltage). The transimpedance is the impedance value that is defined as the output voltage divided by the input current.

By simply connecting a resistor R from the output of the opamp to the inverting input of the opamp, and connecting the photodiode also to the inverting input, you realize a transimpedance equal to R!

But beware, if you need DC coupling, then be sure that the photodiode cannot be damaged by some overvoltage because of switching transients!

radiohead said:

What you really need is a transimpedance amplifier (converts current to voltage). The transimpedance is the impedance value that is defined as the output voltage divided by the input current.

By simply connecting a resistor R from the output of the opamp to the inverting input of the opamp, and connecting the photodiode also to the inverting input, you realize a transimpedance equal to R!

But beware, if you need DC coupling, then be sure that the photodiode cannot be damaged by some overvoltage because of switching transients!

Click to expand...

Well , it is not only the transimpendance issue . Photodiode has dark current which is dependent on temperature. Amplifier above provides autcorrection for diode dark current if temperature or power voltage will change , and i will be able to reduce led power consumption working in pair with pohotodiode - as amp will work in mode A and at the same time will emulate comparator for input signal source (pulse). Providing