[sdiy] On the topic of ring modulation (Using FETs instead of diodes?)

[Donald Tillman]

On Aug 14, 2014, at 1:01 PM, Ove Ridé <nitro2k01 at gmail.com> wrote:

> On 14 August 2014 20:43, Donald Tillman <don at till.com> wrote:
>> 
>> The diode ring modulator depends on the diode nonlinearity, which the precision rectifier circuit doesn't have.
> 
> Could you explain this further? Are you saying the circuit wouldn't
> work with a theoretical diode which is fully on when forward biased
> and fully off when reverse biased?

Radio Dude is demonstrating the radio application of a ring modulator, not the audio application.  In his case he's operating with the LO signal at about 6 times the level of the RF signal, so the bridge diodes are effectively being used as switches, switching the RF signal between polarities.  And later LC circuits will filter out all but the difference signal.

For the audio application we use germanium diodes, and use the forward diode curve, so the diode bridge does a smoother fade between polarities.  Audio applications want something closer to an analog multiply operation.  

Radio Dude describes the multiply operation at 0:44, but he implements "some form of multiplication" which ends up being closer to switching between polarities.

So neither the audio or radio application is actually helped by a "theoretical diode" or a "precision rectifier".

If you wanted Radio Dude's operation for audio, it's a simple matter to switch an opamp circuit between a gain of +1 and -1. No precision rectifier needed.

  -- Don

--
Don Tillman
Palo Alto, California
don at till.com
http://www.till.com