# Circuit to produce Vout= V1 x V2

#### Robert Onabeach

Joined Apr 5, 2021
7
Hi Everyone new guy here who needs your help. I need a simple (?) circuit to produce an output voltage that is the product of two input voltages. I thought it would be easy using some op-amps but have been unable to come up with one. A search produced the standard summation and differiation circuits but no product circuits. The closest I found was using an analog computer to do the product.

Actually I would love to build an analog computer from a kit but found none available.

Any sugestions?

Thks Robert

#### crutschow

Joined Mar 14, 2008
32,904
Multiplication is a non-linear function, while op amps are linear and require additional circuitry to perform any non-linear operation.
The easiest and most accurate way to multiply two voltages is to use an analog multiplier chip as MrChips suggested.

#### Papabravo

Joined Feb 24, 2006
20,386
Multiplication is a non-linear function, while op amps are linear and require additional circuitry to perform any non-linear operation.
The easiest and most accurate way to multiply two voltages is to use an analog multiplier chip as MrChips suggested.
For AC signals, like sine waves, you can use a diode mixer. It is what they do in radios. When you multiply two sine waves you get energy at the frequencies of the two inputs but you also get the sum and the difference. You can sorta do multiplication with potentiometers, but that doesn't help much when things are changing.

#### Robert Onabeach

Joined Apr 5, 2021
7
For AC signals, like sine waves, you can use a diode mixer. It is what they do in radios. When you multiply two sine waves you get energy at the frequencies of the two inputs but you also get the sum and the difference. You can sorta do multiplication with potentiometers, but that doesn't help much when things are changing.
Thanks for the help. Yes, I found out an analog multiplier is what I need. Just ordered an AD633JN. Don't need bandwith; the application is almost DC.

#### crutschow

Joined Mar 14, 2008
32,904
What are the peak-to-peak voltages of the two signals?

Below is the LTspice simulation of the AD633 used as a simple multiplier:

#### Attachments

• 103.1 KB Views: 12

#### Ian0

Joined Aug 7, 2020
8,392
Gilbert cell
https://en.wikipedia.org/wiki/Gilbert_cell
Op amps may be linear, but transistors are exponential. It is (relatively) easy to perform exponentials and logarithms using bipolar transistors.
Adding two logarithms is the same as multiplying.

#### crutschow

Joined Mar 14, 2008
32,904

#### Ian0

Joined Aug 7, 2020
8,392
I believe that's what the AD633 uses internally.
So if you want an accurate one that those nice people at Analog Devices have calibrated for you, it will cost you a tenner, or you can make one yourself out of three BCM847B matched dual transistors that will cost 60p.

#### Robert Onabeach

Joined Apr 5, 2021
7
So if you want an accurate one that those nice people at Analog Devices have calibrated for you, it will cost you a tenner, or you can make one yourself out of three BCM847B matched dual transistors that will cost 60p.
The application is quite rudimentary. I want to make a power meter to measure the power produced by a model gas engine hooked up to a little generator, I.e. P=I x V. I will measure the current with a power resistor. Haven't checked the waveform from the brushed generator yet. I hope it is mostly DC. Accuracy is not important; just need a ballpark figure to monitor the output of the engine.

#### crutschow

Joined Mar 14, 2008
32,904
A power monitor IC, such as one of these may be a cheaper alternative to an analog multiplier.

#### Robert Onabeach

Joined Apr 5, 2021
7
A power monitor IC, such as one of these may be a cheaper alternative to an analog multiplier.
The output will go to a analog voltmeter, so I don't need any digital output or programable capability.
I just looked at the signal output of the generator; has about a 10V dc output with 1-2V ripple into a 20Ω load.

#### crutschow

Joined Mar 14, 2008
32,904

#### crutschow

Joined Mar 14, 2008
32,904

#### crutschow

Joined Mar 14, 2008
32,904
I've had mostly good luck with eBay, but has there been problems purchasing IC's from them?
Sometimes they are fake or knockoffs that don't function as well as the originals.
I've not purchased any IC's from them, but I've heard some bad stories from some who have.
It's especially problematic if the IC markings don't match Analog Devices'.

#### Robert Onabeach

Joined Apr 5, 2021
7
Sometimes they are fake or knockoffs that don't function as well as the originals.
I've not purchased any IC's from them, but I've heard some bad stories from some who have.
It's especially problematic if the IC markings don't match Analog Devices'.
The markings from the eBay supplier do indeed look very questionable, so I will not buy from them. Thanks for the heads up!

#### crutschow

Joined Mar 14, 2008
32,904
If you want to try a roll-your-own, here's a multiplier circuit that uses PWM to do a 1-quadrant multiplication (positive input voltages only) and is limited to a couple hundred Hz input signals, but that should be okay for your application.
The LTspice simulation is below:
It use a low-cost, single-chip, quad LM339 comparator to multiply the two signals, Vin1 (yellow trace) and Vin2 (red trace).
The PWM duty-cycle represents the value of the In2 input, and the PWM output voltage amplitude is determined by the value of the In1 input.
The average of this output is thus determined by the multiplied value of the two inputs.

The two transistors form a current mirror to give constant-current capacitor C1 charge, and thus a linear sawtooth generated by U1, for reasonable good linearity of the multiplication.

The In2 voltage range is about 0-4V (for a 5V supply), but In1 can go from 0-8V.

As you can see, the PWM filtered output (blue trace) shows the curved output as expected for the product of the two ramp inputs, Vin1 and Vin2.
The output is approximately (0.25 * Vin1 * Vin2).

Last edited: