# Amplify a sine wave 0.5mV to 100mV

#### Barryallen

Joined Apr 17, 2017
20
Hi Everyone,

I'm building a circuit to read frequency of 0.5mV sine wave with range from 2KHz to 3Khz.

Now I need to amplify 0.5mV to 100mV so that I can read the frequency.

#### shteii01

Joined Feb 19, 2010
4,644
Gain=100/0.5=200

Gain Bandwidth Product=3000 Hz*200=600000 Hz=600 kHz=0.6 MHz

Now go look for op amps with Gain Bandwidth Product of 600 kHz or more.

#### Barryallen

Joined Apr 17, 2017
20
Gain=100/0.5=200

Gain Bandwidth Product=3000 Hz*200=600000 Hz=600 kHz=0.6 MHz

Now go look for op amps with Gain Bandwidth Product of 600 kHz or more.
Dear Shteii01,

I am very bad at this field. Please give me some more details like specific circuits.

Thank you again.

#### AnalogKid

Joined Aug 1, 2013
9,296
Can you post your circuit so far? That will tell us many things about the input stage needed.

ak

#### WBahn

Joined Mar 31, 2012
26,398
Dear Shteii01,

I am very bad at this field. Please give me some more details like specific circuits.

Thank you again.
If you are so bad at these things, then why are you doing it? What is the objective here? We often see people that are stuck on trying to implement a circuit that is not a very good way to solve a problem just because they don't know of any other way to go about it. So what is the problem you are trying to solve?

#### Barryallen

Joined Apr 17, 2017
20
If you are so bad at these things, then why are you doing it? What is the objective here? We often see people that are stuck on trying to implement a circuit that is not a very good way to solve a problem just because they don't know of any other way to go about it. So what is the problem you are trying to solve?
Hello,

Actually, I'm not too bad, like I have basic knowledge but It's not enough in this case.

I'm building a circuit to read the frequency from a piezometter. The sine wave output signal amplitude is on the order of .5 to 5 mV. It's too small to interface with an microcontroller.

Here is my idea: amplify the output signal from 0.5mv to 1000 mv ( gain = 1000/0.5=2000 ) using two stages amplifier, then square the signal by using an comparator op am to have squared pulses in order to read the frequency by an microcontroller.

Do you have any ideas?

#### Barryallen

Joined Apr 17, 2017
20
Can you post your circuit so far? That will tell us many things about the input stage needed.

ak
Hi, I posted my circuit above, Please give me some suggestions !

#### AnalogKid

Joined Aug 1, 2013
9,296
Both gain stages are DC coupled, so even a tiny 1 mV of input offset voltage in the first amp will appear as 10 V of offset at the output. Change both gain stages to non-inverting amplifiers with a capacitor in the feedback network shunt leg.

The output of U1B is a saturated square wave, so you do not need U3A.

ak

#### Ramussons

Joined May 3, 2013
1,032
If you all you need a square wave, then why so many stages?
Use a Hi Speed Comparator like LM193 to get a Square wave from the input Sine. You will need to level shift the input sine if it's not bipolar.

Or, this may be even simpler, though not sure whether it's ok for a .5 V input.

C2 will not be required

#### AnalogKid

Joined Aug 1, 2013
9,296
If you all you need a square wave, then why so many stages?
Use a Hi Speed Comparator like LM193 to get a Square wave from the input Sine. You will need to level shift the input sine if it's not bipolar.
Nope. The common mode input range of a 193 extends below the negative rail (GND in this case), so the part will function as a half-wave rectifier with gain without any added input offset.
Or, this may be even simpler, though not sure whether it's ok for a .5 V input.
It is not. The nominal transition levels for CMOS Schmitt trigger inputs are 1/3 and 2/3 Vdd, so this circuit will not see a 1 mV signal.

ak

#### Barryallen

Joined Apr 17, 2017
20
Both gain stages are DC coupled, so even a tiny 1 mV of input offset voltage in the first amp will appear as 10 V of offset at the output. Change both gain stages to non-inverting amplifiers with a capacitor in the feedback network shunt leg.

The output of U1B is a saturated square wave, so you do not need U3A.

ak
Dear Sir,

The gain of the first op am is 100,000/1000=100, the second is the same. So the total gain here is 10000 --> the output voltage is 0.5mV*10000= 5000 mV =5V. But when simulating I just get 1V. How do I get 10V in this case plz?
Change both gain stages to non-inverting amplifiers with a capacitor in the feedback network shunt leg.
what is "the feedback network shunt leg"? So sorry for my bad. Plz just show me where exactly to place the capacitor.

Thank you.

#### Barryallen

Joined Apr 17, 2017
20
If you all you need a square wave, then why so many stages?
Use a Hi Speed Comparator like LM193 to get a Square wave from the input Sine. You will need to level shift the input sine if it's not bipolar.

Or, this may be even simpler, though not sure whether it's ok for a .5 V input.

C2 will not be required
Hello,
As I know, The offset voltage of comparator op am is around 2mV. My input voltage is 0.5 mV. So it seem to be impossible with this idea. Do you have any ideas else ?

#### DickCappels

Joined Aug 21, 2008
7,569
The simulation might be giving lower gain than you expect because the LM358 has a marginal gain-bandwidth product at best. A faster opamp will likely improve things.

To improve your real world experience you might want to AC couple the signals into both stages so you don't multiply the offset of the first opamp by 10,000X, and bias the input at 50% of power supply so you don't clip the negative halves of the cycles.

Do you want help selecting an opamp? If so, let us know where you live, whether you can buy parts through a distributor and if so, which distributor(s).

#### Barryallen

Joined Apr 17, 2017
20
The simulation might be giving lower gain than you expect because the LM358 has a marginal gain-bandwidth product at best. A faster opamp will likely improve things.

To improve your real world experience you might want to AC couple the signals into both stages so you don't multiply the offset of the first opamp by 10,000X, and bias the input at 50% of power supply so you don't clip the negative halves of the cycles.

View attachment 127023

Do you want help selecting an opamp? If so, let us know where you live, whether you can buy parts through a distributor and if so, which distributor(s).
Dear Sir,

Thank you so much for the answers.

Please help me select an opamp. I live in Vietnam. Just give me some opamps. I will find all the way to buy it.

By the way, Please give me the value of capacitors C3, C4.

Thank you again.

#### DickCappels

Joined Aug 21, 2008
7,569
A TLC2272 will do a great job. It is in stock at Electronic Source in Bangkok.

The 10 uf capacitors can be smaller if you want a quicker settling time after turn-on of if there is some low frequency noise you want to reject. I recommend the two 1N916 diodes across the second feedback resistor -they clip the output signal to about 1V P-P and prevent oscillation as the 2nd amp comes out of saturation.

#### Ramussons

Joined May 3, 2013
1,032
Nope. The common mode input range of a 193 extends below the negative rail (GND in this case), so the part will function as a half-wave rectifier with gain without any added input offset.

It is not. The nominal transition levels for CMOS Schmitt trigger inputs are 1/3 and 2/3 Vdd, so this circuit will not see a 1 mV signal.

ak
My mistake. Did'nt notice "m" in the input level

#### Barryallen

Joined Apr 17, 2017
20
A TLC2272 will do a great job. It is in stock at Electronic Source in Bangkok.

View attachment 127058

The 10 uf capacitors can be smaller if you want a quicker settling time after turn-on of if there is some low frequency noise you want to reject. I recommend the two 1N916 diodes across the second feedback resistor -they clip the output signal to about 1V P-P and prevent oscillation as the 2nd amp comes out of saturation.
Dear Sir,
Thank you so much for the circuit,
I will test it as soon as possible and let you know the result then.
Best regards,