Hello,

I've been searching the net for a good low frequency preamplifier, and have come up short. I was wondering if there was any the users of this site could recommend?

I'm interested in the range of 2-200Hz, most preamps are only really good down to 100Hz, 10Hz at a push!

I really like this preamp, having used it before, but as you'll see from the frequency response the roll off is pretty bad below 100Hz.

http://www.zen22142.zen.co.uk/Circuits/Audio/lf071_mic.htm

Ideally, I'd like something with little roll off at very low frequencies (around 2Hz), lower power requirements than the one posted (preferably around 3V), and fairly simplistic to change the gain.

Cheers

 

You need to say a bit more about what you are amplifying.
Why not increase the value of the input and output coupling capacitors?
You can eliminate the capacitors altogether and go all the way down to DC.
However you will need a split power supply to remove the DC offsets.

 

You need to say a bit more about what you are amplifying.
Why not increase the value of the input and output coupling capacitors?
You can eliminate the capacitors altogether and go all the way down to DC.
However you will need a split power supply to remove the DC offsets.

I'll be boosting the signal of a MEMS microphone. Trying to keep power consumption low so not sure split power supply is the way forward.

Sorry for my ignorance, but what will increasing the input and output coupling capacitors do? I understand that the design that I posted a link to I can change the value of gain by changing R2, but that's not my concern at the moment, I just want a suitable gain increase at low frequencies (which I would guess is achieved by changing the coupling capacitors, otherwise you wouldn't have suggested it!).

Cheers for your response.

 

The first MEMS microphone I come up with has a low limit of 100Hz. No point designing an amplifier for 2Hz when that is 1/50 times the microphone response.

Next problem those amplifier chips won't work at 3 volts.

Show me the microphone that goes to 2Hz and I'll get serious.

http://www.analog.com/en/audiovideo-products/mems-microphones/products/index.html

 

The first MEMS microphone I come up with has a low limit of 100Hz. No point designing an amplifier for 2Hz when that is 1/50 times the microphone response.

Next problem those amplifier chips won't work at 3 volts.

Show me the microphone that goes to 2Hz and I'll get serious.

http://www.analog.com/en/audiovideo-products/mems-microphones/products/index.html

I have custom microphones from Knowles which have a very small roll off down to 2Hz.

 

If you show me how to power the mic I might be able to go to DC (0Hz).
I'm thinking you have a 2 wire mic that uses 20 ua and the power supply is 3 volts. Use a 75k resistor to get the mic to use 1.5 volts at 20 microamps and the centering for the amplifier is done. If you can set the resistor to power the mic so it has 1.5 volts DC at its feed terminal you won't need capacitors.

PS, here's a chip that will work. 22 cents plus shipping.

 

If you show me how to power the mic I might be able to go to DC (0Hz).
I'm thinking you have a 2 wire mic that uses 20 ua and the power supply is 3 volts. Use a 75k resistor to get the mic to use 1.5 volts at 20 microamps and the centering for the amplifier is done. If you can set the resistor to power the mic so it has 1.5 volts DC at its feed terminal you won't need capacitors.

PS, here's a chip that will work. 22 cents plus shipping.

Thank you for your response. Min and max supply voltage at 1.5 and 3.6, respectively. Maximum current consumption is 250uA. Three pins; ground, VDD, and output.

So, if I understand correctly, you're suggesting using the circuit I posted, with your selected op amp (which has a very low power consumption), and a 74k resistor giving the mic a separate power connection? Obviously your op amps selling point is low power, but will I have issues with it not explicitly being low noise as the op amp is in the original circuit?

 

Here's a chip with less than a third of the noise of the TL071 and an AC circuit.

 

Here's a chip with less than a third of the noise of the TL071 and an AC circuit.

Thank you very much, sir. I appreciate your time.

This certainly achieves the low-power requirement, may I assume it also solves the original 'need for gain at low frequency' problem?

I shall pass it though a Multisim simulation soon and see how it performs.

Thanks again.

 

If it doesn't meet the frequency needs, it is merely a mistake on my part and can be cured by changing the size of the capacitors. Bigger capacitor = lower frequency.

 

I don't think anyone mentioned it but, regarding your capacitor questions, the rolloff point (-3dB) of an RC DC block is 1/(2\(\pi\)fRC) where f is the frequency. Thus you can see that in increase in either R or C will lower the rolloff point. (R for the referenced amp is the two 47KΩ resistors in parallel for the input, and whatever the output load resistance is for the output.)

 

I used 15k because that chip increases its input bias current when it gets hot.
But I didn't even think about adding a capacitor for the unknown load.
Those equations should help.

 

If it doesn't meet the frequency needs, it is merely a mistake on my part and can be cured by changing the size of the capacitors. Bigger capacitor = lower frequency.

Ah you're a legend. Thank you very much! I'll have to have a play around with this design shortly.

 

The 10uf input capacitor causes a cutoff frequency of 0.68Hz.
The 10uF capacitor in series with the 1k resistor to ground causes a cutoff frequency of 16Hz.

 

So I used Multisim to plot and simulate the design, making a change of a 10uF input capacitor as I think 22uF was a bit too low in frequency. I was under the impression I would get a cut off (which using the equation crutschow posted) of around 1Hz.

However, as attached. I'm getting some confusing results. Using an AC signal voltage source at 1mVpk and various frequencies as a microphone substitute, the results look good at 20Hz, giving a clear 100x gain with no phase shift, attachment '20Hz'.

However, at 2Hz, I'm getting a very strange cut-off jump, as you'll see in the attachment '2Hz'. There is also a large phase-shift. I'm to understand this may be unavoidable, but is there anyway to keep the phase at 0?

I'm not sure what's causing this. I've tried changing the input capacitor size but no change. Could this be because I'm using a 741 op amp as opposed to the 971? Perhaps the Vpk that I'm using for the signal generator/mic substitute?

Cheers.

 

Don't you know the very simple calculation of the frequency cutoff of a coupling capacitor feeding a resistance??
Why are the biasing resistors (15k) for the opamp with a value that is extremely low??

The two 15k biasing resistors (they should be 150k ohms) make a total resistance of 7.5k ohms and the 10uF capacitor that feeds them form a highpass filter with a cutoff frequency of 2.1Hz (45 degrees phase shift and -3db in level).

 

I already explained the 15k resistors in post#12. If you use 150k the bias point could shift by as much as 10% when the chip gets hot. If you change to 150k and reduce the capacitor proportionately, the phase shift will be the same. All RC networks cause phase shift. If you just change the input capacitor to 10 times its present value, the phase shift will be less and the low frequency limit will be very low.

 

I don't think anyone mentioned it but, regarding your capacitor questions, the rolloff point (-3dB) of an RC DC block is 1/(2\(\pi\)fRC) where f is the frequency. Thus you can see that in increase in either R or C will lower the rolloff point. (R for the referenced amp is the two 47KΩ resistors in parallel for the input, and whatever the output load resistance is for the output.)

You need to move the f to the left side of the equation.

 

Don't you know the very simple calculation of the frequency cutoff of a coupling capacitor feeding a resistance??
Why are the biasing resistors (15k) for the opamp with a value that is extremely low??

The two 15k biasing resistors (they should be 150k ohms) make a total resistance of 7.5k ohms and the 10uF capacitor that feeds them form a highpass filter with a cutoff frequency of 2.1Hz (45 degrees phase shift and -3db in level).

Unfortunately, no. I don't know the 'very simple calculation of the frequency cutoff of a coupling capacitor feeding a resistance(??)'. Although I think we can agree that if I did know such a calculation I wouldn't be asking for help.

Thank you for your input, the resistors have been changed to 150k and the capacitor to 220uF giving satisfying results down to 2Hz.

I already explained the 15k resistors in post#12. If you use 150k the bias point could shift by as much as 10% when the chip gets hot. If you change to 150k and reduce the capacitor proportionately, the phase shift will be the same. All RC networks cause phase shift. If you just change the input capacitor to 10 times its present value, the phase shift will be less and the low frequency limit will be very low.

Thank you.

 

Multisim shows (severe distortion) that a lousy old 741 opamp (45 years old) does not work when its supply is less than about 10V.