[sdiy] returning to the chronic disappointment of the 1496...

[alfred.pear at gmail.com]

Alright, apparently I was premature in denying the existence of AD633
in an otherwise chaotic and indifferent universe.

However, the $7 IC is prohibitively expensive when compared to the
lure of a 50 cent LM1496.

Jurgen, the doubled carrier frequency is the exact problem. I will
take a much closer look at the modulation stage in your Frequency
Shifter.

I'm also very interested in your solution, David.. I'll commence
pestering shortly.

Thanks SDIY!

-AP


On Thu, Jan 14, 2010 at 1:31 PM, David G. Dixon
<dixon at interchange.ubc.ca> wrote:
>> Has anyone found any workarounds for the signal bleed-through problems
>> with 1496s used as 4Q-Multipliers?
>> I've read and re-read mailing list archives, message boards, and tried
>> almost all the schematics I could find...I cannot seem to get the two
>> signals to actually *balance* with real precision.
>>
>> With the 633s and 1495s off the market, there has to be a cheap,
>> effective solution to ring modulation, right? Have OTA approaches
>> yielded worthwhile results?
>
> Alfred, I have developed, built and tested a 4Q multiplier based on the
> linearized SSM2164 (actually I use the Coolaudio V2164 -- $3.50 each at
> Small Bear, or $3.00 for 10 or more).  It works very well.  I originally
> developed the circuit based on the LM13700, and this also works very well,
> but is a little bit trickier to set up.  In either case, I arranged the VCAs
> in a "push-pull" arrangement.  One processes the carrier signal as is, and
> the other processes the inverted signal.  Also, one is controlled by the
> modulating signal as is, and the other is controlled by the inverse of the
> modulating signal.  In the case of 13700, the carrier signal must be
> dramatically reduced to avoid distortion, and the modulating signal must be
> shifted carefully, but things are much more straightforward with the
> linearized 2164.  I also have a pot which pans between the two carrier
> signals such that at either extreme one obtains 2Q multiplication, in the
> middle one obtains very well balanced, symmetrical 4Q multiplication, and in
> between one obtains various degrees of unbalanced, asymmetrical 4Q
> multiplication.  Finally, I've also brought both signals in through
> variable-gain ideal rectifiers fitted with 3-way switches which allow either
> +, +/-, or - polarity.  This adds a lot of unique functionality in CV
> modulation, and creates some very subtle and interesting effects in audio
> modulation.
>
> This circuit grew out of the desire to avoid spending beaucoup bucks on
> "proper" multiplier chips like the 633, and also from the painful
> realization that the (very simple) ring-mod circuit in the LM13700 datasheet
> sucks really hard!
>
> Please email me if you are interested, and I can send you my design files.
>
>