1/X circuit

urosuros urosuros at bits.net
Thu Jan 27 18:15:00 CET 2000


----- Original Message -----
From: Tony Allgood <oakley at techrepairs.freeserve.co.uk>
To: synth diy <synth-diy at mailhost.bpa.nl>
Sent: Tuesday, January 25, 2000 11:09 PM
Subject: 1/X circuit


> Hi all,
>
> I'm looking for a circuit design that has a 1/x output where x is a
> voltage input. A bit like a expo convertor, but not quite. Any ideas?
> I'm stumped.
>
> Regards,
>
> Tony Allgood  Penrith, Cumbria, UK
>
> MidiDAC Superior midi-cv, SuperLadder, TB303 clone and Filter Rack
>
> http://www.techrepairs.freeserve.co.uk/projects.htm
>
>
>

           ------------- Vcc         Hi Tony sorry for this terrible "
drawing " and for my terrible english .
                |                          I am just a ee student so it is
highly possible that my idea does not work
                I                          and i would be thankfull if you
point me out where did I go wrong .
                I                         Story is : on a picture we have
voltage divider formed with resistor R and
                I                         mosfet . Obviously we want mosfet
to run in triode region of it's characteristic
                I                         If mosfet is running in triode
region then :
                /                               Vout=(Rfet/(R+Rfet))Vcc
                \                               Rfet=1/(B(Vgs-Vt))
                /
                \  R                     conditions for mosfet running in
triode region would be Vgs>Vt and Vds>Vt
                /                         so to do this we should aply
biasing voltage Vt on input as shown on picture
                \                         here we have first draw back :
input voltage must be allways positive but that
                |                         should be no problem if you are
doing with CV's ( they are allways positive,right ?)
  Vout------|                         so if conditions are met , and we
asume that R>>>Rfet ( which is met to if we take
                |                        somethig like 10k ohm resistor )
than R+Rfet=R ( aproximation )
                |                              Vout=Rfet(Vcc/R)
                |                             Vout=Vcc/(RB(Vin + Vt -Vt) ,
Vout=(Vcc/(RB))/Vin
         drain|____
                        |  gate                                     At the
and you should scale Vout and that would do high imput
                        |------------------ (Vin +Vt)          impedance
amplifier stage with gain of A=RB/Vcc that will
     source<------I                                             buffer too .
that should give a circuit with 1/x transfer .
                |
drawbacks : one mentioned above , bad dinamic range ( maybe
                |                                                    very
bad ) , and nonlinearity in triode region . I know some
         ____|____                       people linearise whole triode
region with asumption that mosfet acts as
           ----------                        resistor all the way up to the
saturation region but i'm not shure about that
              `````  gnd                   Anyway , as I sad I'm student( i
dont know much) and this idea just came

across my mind so it is  probably bulls**t and I would be glad if you point
me out what is not working if something is not working .
One last thing . I live in a bad place ( Belgrade , Yugoslavia ) for
electronic diy ( this is a bad place for various other stuf as a matter of
fact ) so finding BBDs' here is pure SF . So I must ask somebody to buy it
for me abroad , but I dont want to get ripped of . So what are usual prices
of chips from MN serie .
Thanks in advance , sorry again for english and clumsy letter ( I dont have
software to draw circuits with
Respects for all on SDIY
Uros

























More information about the Synth-diy mailing list