[sdiy] New schematics

houshu at muj.biglobe.ne.jp houshu at muj.biglobe.ne.jp
Fri Jun 13 10:17:22 CEST 2003


Hello John, JH, and List,

To John Blacet and those who were confused with my schematics:
I'm sorry for my laziness.  But that's why I state at the top of 
my site "This page is not for beginners". 

2003/06/11 16:35:26 +0200, jhaible at debitel.net wrote:
>Nice stuff, as always!

Thank you very much.

>Two circuits have caught my interest in particular:
>
>>  - New style filter (with smooth asymmetrical distortion) :
>>    http://www5b.biglobe.ne.jp/~houshu/synth/VcfCtp0211.gif
>
>This looks promising - smooth assymetrical clipping, with each
>following stage dealing with the opposite half wave than the
>previous one, I presume? 

Exactly.  I think this structure has potential to be my best VCF.

>>  - A method to increase the input dynamic range:
>>    http://www5b.biglobe.ne.jp/~houshu/synth/Vca0302.gif
>
>I once thought about adding a long diode string to each emitter
>of a differential pair. But I've never really tried it.
>Your circuit should increase the SNR by 3dB (double signal
>amplitude, SQRT(2)* noise ). 
>EMS built voltage controlled resistors from 20 diodes in series
>(instead of 2, so the factor is 10, not 20), - that would be 10dB
>SNR improovement.
>Would this be possible with a differential pair, too?

The patent (not mine) said similar things.  In a figure, more than 
3 diodes were inserted to each emitter.  I can't remember exactly 
because I saw the patent over 7 years ago.  I'm sorry.

>Or would it be better to connect all the transistors in _parallel_,
>rather than in series? (Which is what the expensive low noise
>matched pairs from NS do) ?

I have very limited knowledge on noise theory.  As for the noise 
from outside, the higher the signal level is, the lower the 
influence of the noise is.  When the diff amp is used without 
feedback (open loop), the serial structure could have advantage.
When used with feedback, frankly, I can't analyze it.

Another possible advantage is input bias current, which is not 
so important in low noise applications though.

Sam Hoshuyama
Kawasaki, Japan



More information about the Synth-diy mailing list