CV/gain relationship of a JFET?

[Ribbledox]

Hello!

I came across a website (by Jurgen Haible) were he analyses the behaviour of a 1176 clone.

He mentions that the JFET element in the 1176 is neither linear nor exponential VCA, but an element with its own kind of control non-linearity. And he mentions that this is one of the reasons why the Time Constants on a limiter like the 1176 change with input level.

I am just curious, what sort of relationship this is, between CV and GR, and if its in the nature of the JFET itself? I read somewhere earlier, that this is a squared relationship, rather than linear or exponential. Is there any thruth in this?

Would appreciate a clarification=)


/R

 

[tmbg]

Wouldn't a 'squared' relationship be an exponential relationship?

 

[clintrubber]

Wouldn't a 'squared' relationship be an exponential relationship?

In Jip&Janneke-taal (please forgive me the Dutch jargon): squared is that thing you do with '2' and 'exponential' the thing you do with 'e' (2.71..).
Further difference: 'squared' is something to the power of '2', and 'exponential' is 'e' to the power of 'something'.

Have a look at the various electronics-links in the Meta-Meta; the JFET- & bipolar-relationships will explain it further.

 

[SSLtech]

To my way of thinking a square is an exponent, where the exponent is two. A cube is an exponent of three. They are both exponential in behaviour.

The term exponential is a general term for an accellerative numeric behaviour, or which square, cube, quad etc. are all examples.

Therefore anything which is squared is exponential, but not necessarily the other way round.

Keith

 

[clintrubber]

from Keith:

The term exponential is a general term for an accellerative numeric behaviour, or which square, cube, quad etc. are all examples.

Therefore anything which is squared is exponential, but not necessarily the other way round.

That describes it a lot better ! :thumb:

 

[volki]

ribbledox, what's the url of that analysis?

 

[Ribbledox]

The URL is: http://www.oldcrows.net/~jhaible/compressor/jh_compress.html

But I'd be very happy if someone could explain the gain/bias relationship of the JFET...

As far as I've understood it's because of this (the non-linear gain/bias relationship) that vintage compressors can't be used in feedforward, and it is not possible to have Time Constant that are independent from the input signal.

Or am I wrong?

Cheers!

/R

 

[CJ]

JH hangs around here, no?
http://www.groupdiy.com/forum/profile.php?mode=viewprofile&u=417

 

[volki]

[quote author="Ribbledox"]
As far as I've understood it's because of this (the non-linear gain/bias relationship) that vintage compressors can't be used in feedforward, and it is not possible to have Time Constant that are independent from the input signal.

Or am I wrong?[/quote]

thanks for the link, still need to go through it, though...

as for feed-forward, one classical example of that topology would be the U73 by TAB/Telefunken/maihak... that one's a tube design, even.

http://www.kubarth.de/cgi-bin/diy/german/getfile?f=4,3$P,#8P-34P,5]U+W4W,RYG:68`

IMO dependency of time constants on the signal is actually a good thing...?