4 quadrant analog multiplier with plain components, DC accurate

[Davor]

As announced at 4 quadrant receiver topic, I invented a simple and DC accurate analogue multiplier. 3 opamps, 4 diodes of the most common kind, and several resistors. It is arguable that I could get some Gilbert cell chip for the purpose, but those are not DC accurate, and for sure opamps are much cheaper and common.

It came to me as an epiphany. It happens to me sometimes when I'm relaxed and do what I like ...as opposed to working under pressure.

There are 2 voltage inputs, of course, and V1 would correspond to a LO input because it saturates at high input levels. V2 would be a signal input. U1 and U2 are an instrumentation amplifier, and the rest ... you'll recognise a diode ring. There is also some clever play with balances in diagonals, and a result is ... not perfect, but surely good enough.

I'm sure someone else would spend the next at least 6 months in patenting this idea. I on the other hand am not at all impressed by patents, and I'm appalled by patent trolls. So it is published here and you are free to enjoy it - unless you are a patent troll.

[deemon]

Good idea , Davor ! Congratulations !

I'd never seen such implementation of multiplier .... the only drawback what I see is low speed due to big resistor values in the signal chain .
But for MD purposes it can be enough ...

[Davor]

Yes, it is high impedance all over. These can be scaled down, but the V1 input goes into saturation sooner. I wanted a DC accurate amplifier for 1Vpp at inputs, and this is precisely it.
Surely some other diodes would perform different, but one of my goals was to use only non-specialty never-to-fall-into-obsolence garden variety components available in Goundam near Timbuktu if necessary, and again - this is precisely it.

I'm not sure what will be my next quest, maybe an OTA stand in. It is horrible to observe how some components just die into oblivion, as if no one ever needed them.

[Aziz]

Nice work Davor! I agree with you and appreciate your patent policy.
How about tempco (temperature dependency)? I think, this is the crucial point in all analog multipliers.

Maybe it's really time to multiply the whole thing in DSP software?
a = b*c

*LOL*
Cheers,
Aziz

[Davor]

Take a closer look, and you'll see - it is the same component that does log and anti-log functions, so temperature does not matter that much.

Regarding DSP, I'm not inclined to give in for the portions crucial for the rig functioning. I'll settle for a VDI some of these days. But if I can fix it in analogue - analogue that is!

Your sound card project, or in any case a DSP project is the other side of the story, and it is precisely the way to do it commercially. Very little analogue bits and pieces, and a serious number cruncher.

[rickb]

Davor:

Has this multiplier or something like it found a place in one of your detectors?

[kt315]

....