Opamp based amplifier design question

[Magtheradon]

I'm constructing a low frequency AC amplifier using a LM358 dual op-amp and i want a gain of 100. For this im using a 1M resistor in the feedback path and a 10k resistor in the input. But some people advised me against taking a high valued resistor in the feedback path saying that it will lead to parallel loading of the op-amp and also since the current will be in the micro amp range the circuit will be prone to noise. Is this true? Do i have to scale down the resistances? The amplifier is a part of a bigger circuit and the same op-amp is used as a low pass filter. It will be difficult for me to scale down the resistor and capacitor values now. Please advice guys.

Andy

 

[keith1200rs]

A 1M resistor is noisier than a 100k resistor (thermal noise). Also, a high impedance is more likely to pick up noise than a lower value. The opamp input noise current will be multiplied by the feedback resistors. All of these point to using lower values in general. Whether it makes any difference in your circuit depends on specific details.

Keith

 

[LvW]

Another argument against high feedback resistors is the parasitic voltage drop caused by the input dc current leading to a voltage offset.

 

[Magtheradon]

Another argument against high feedback resistors is the parasitic voltage drop caused by the input dc current leading to a voltage offset.

This is an AC amplifier so i should not have that problem right?

 

[mdocampo1]

Increasing your resistances will make your currents decrease, thus approaching the leakage currents of your op-amp. Due to this, the leakage currents of your op-amp will not be negligible. Better use smaller values of resistances

 

[Magtheradon]

Guys this is my circuit

Its actually a heart beat monitor. It works on the principle of density modulation. When the heart pumps there is an influx of blood into the tip of the finger. This leads to a decrease in the light received by the infrared sensor. Im amplifying the voltage drop across the IR sensor 100x100=10000 times and feeding it to a microcontroller and sending it to the computer to plot it.

Since the human pulse rate varies between 60-150 beats per minute i wanted to band pass filter the signal between 1Hz and 2.5Hz. In the actual circuit using the available resistances and capacitances i could only band pass between 0.66Hz and 2.3Hz. I had to use a 1M feedback resistance in the circuit. Otherwise i would not get the 2.3Hz cutoff frequency by scaling down the resistances due to non availability of the precise value of capacitance required.

Guys please comment if my circuit is prone to noise and what can i do to remove it.

 

[FvM]

I think with a LM358 amplifier, you don't need to worry about resistor noise contribution. But you can scale down the first stage resistor network by a factor of 10 (1k/100k) and check, if the total output noise is reduced. I guess, it isn't. Would be different with a low noise OP in the first amplifier stage. But I also think, the sensor noise will beat amplifier and resistor noise anyway.

 

[Magtheradon]

I think with a LM358 amplifier, you don't need to worry about resistor noise contribution. But you can scale down the first stage resistor network by a factor of 10 (1k/100k) and check, if the total output noise is reduced. I guess, it isn't. Would be different with a low noise OP in the first amplifier stage. But I also think, the sensor noise will beat amplifier and resistor noise anyway.

Scaling down the op-amp resistance isn't a good option for me cause i'll have to use a 0.6uF capacitor with the scaled down resistances to maintain the cut-off frequency but 0.6uF is not available at any store.

But why do you say that with LM358 i don't have to worry about resistor noise contribution? Is there any special reason?

 

[FvM]

Yes, because the OP input noise is much higher than the expectable resistor noise. For the amplifier circuit without photo detector, the noise of the most critical 10k resistor corresponds to about 14 nV/√Hz, but the OP noise can be expected in a 50 to 100 nV/√Hz order of magnitude according to the LM358 transistor currents.

The photodetector shot noise will contribute anyway with 100 nV/√Hz at 100 µA photocurrent. So you can effectively ignore the resistor noise and won't have much problems with LM358 noise.