Re: gEDA-user: Barrie Gilbert

[Karel Kulhavy <clock@xxxxxxxxxxxxx>]

On Wed, Sep 06, 2006 at 06:55:35AM -0400, Dan McMahill wrote:
> Bob Paddock wrote:
> >On Wednesday 06 September 2006 03:36, Karel Kulhavy wrote:
> >
> >>Anyone knows what exactly is called Gilbert cell?
> >
> >
> >The Gilbert Cell is named after Barrie Gilbert of Analog Devices,
> >invented in 1968.

I like how in one article Barrie Gilbert says how modern design is plagued
with DSPs and digital thinking instead of people trying to figure what
wonders are lying in the analogue domain. He seems to be the guy who uses
elegance as the design rule of thumb. Do you know about more people like
him?

CL<

> >
> >http://www.ieee.org/organizations/pubs/newsletters/sscs/jan03/jssc_classic.html
> >
> >The Gilbert Cell has become common in RF designs, used as a double balanced
> >mixer.   It is a four quadrant multiplier.  Somewhere in my files I have a 
> >paper
> >by Gilbert  where he states that he never really meant it to be used the 
> >way,
> >that has been the most common usage.  He recommended an obscure
> >division technique instead.  I'll dig the paper up this evening.
> 
> Bonus points to anyone who can name the real inventor of the mixer in 
> question here.  Hint:  It wasn't Gilbert even though it's called a 
> Gilbert cell.
> 
> >>How many transistors does it actually have?
> >
> >
> >http://rfdesign.com/mag/503rfdf1.pdf
> >
> >
> >>Is it possible to make a well working Gilbert cell with ordinary 
> >>non-matched
> >>transistors? 
> >>And btw do you know what translinear mean?
> >
> >
> >One set of frequencies is translated linearly to an other set of 
> >frequencies.
> >Using non-matched transistors will not be linear, resulting in spurious 
> >outputs.
> 
> no.  That is not what translinear means.  Translinear circuits in this 
> context refers to the class of circuits where you find a loop consisting 
> just of bipolar junctions and have an equal number in each direction. 
> The idea is that  Ic = Is * exp(Vbe/Vt) where Is depends on the device, 
> Vt is the thermal voltage (kT/q), Ic is collector current, and Vbe is 
> the base-emitter voltage.  If you write out KVL around this loop of 
> base-emitter junctions you get:
> 
> sum( Vbe_cw ) = sum( Vbe_ccw)
> 
> where Vbe_cw = junctions where the voltage is positive in the clockwise 
> direction and Vbe_ccw = junctions where the voltage is positive in the 
> counter clockwise direction.
> 
> Now assume all the Is are the same and some simple math shows that
> 
> product( Ic_cw ) = product( Ic_ccw )
> 
> For example, you can build a circuit where I1 * I2 = I3 * I4
> 
> and you can build a squaring circuit or a square root circuit.
> 
> These circuits work on large signals.
> 
> I'll try to find a reference to post tonight and I'll sketch out a more 
> concrete example.
> 
> -Dan
> 
> 
> 
> _______________________________________________
> geda-user mailing list
> geda-user@xxxxxxxxxxxxxx
> http://www.seul.org/cgi-bin/mailman/listinfo/geda-user


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user