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

[Terry]

Just built a DIY germanium diode ring modulator. Had to match the diodes  
and use good quality audio transformers. Used Edcor WMS10K-10K  
transformers, listed at 20-20,000Hz frequency response. Sounds really nice  
to my ears and built with all discrete parts. Added a discrete Moog CP3  
type after the ring modulator to bring the ring audio level up.
Pics of the build:
http://i62.tinypic.com/25r0d3m.jpg
http://i62.tinypic.com/2ldiyo9.jpg
http://i58.tinypic.com/zxn18n.jpg

Put a video up to demo the sound of the germanium diode ring modulator. I  
think it sounds really sweet. Might be the higher quality transformers  
used. Can hear in the video when I unpatch either source the bleedthrough  
is minimal. An interesting quality I noticed was the carrier input, (the  
one that goes to the center taps on both transformers) has a lag. Can see  
at one part of the video when I change the square wave to pulse, it lags  
behind on the scope shot.
Best sounding DIY module I ever made, sounds like magic.

There is an article written by Harold Bode describing his ring modulator  
in the first issue of the Electronic Music Review from the Moog  
Trumansburg days. He mentions some of the differences between the  
switching-type and  mulitplier-type ring modulators. Starts on page 9,  
good read from the old days.
Link here:
http://www.ubu.com/emr/periodicals.html

>> As mentioned in the video, the diode ring mixer has the issue that
>> there are dead spots when the carrier is near zero, because of the
>> diode turn on threshold. This got me thinking, what if you instead
>> used FETs, perhaps driven by a comparator, instead of diodes? This
>> should, if done well, get rid of that non-linearity, as far as I can
>> tell.
>
> Indeed, the dead spots are the major reason that the diode ring  
> modulator doesn't sound very good.  The diode nonlinearities are also a  
> problem.
>