[sdiy] Linearized Vactrol?

Thomas P. Gootee tomg at fullnet.com
Wed Mar 12 23:27:46 CET 2008


Hi Harry,

Thank you for responding.

I have placed a download link for my VTL5C2 LTspice files on
my 'spice-modeling' web-page, at:

http://www.fullnet.com/~tomg/gooteesp.htm

The four files are in a downloadable ZIP file, and include
the VTL5C2 model, a VTL5C2 symbol for LTspice, a
test-circuit which plots the R vs I curve, and a
'plot-settings' file that automatically sets up the plotting
when the test-circuit is run.

The VTL5C2 model is given in the form of a schematic, which
is typically used as a symbol on an LTspice schematic.

The files are best viewed, and run, by using the LTspice
software (which is free from linear.com, and is an excellent
circuit-simulator), for which there is a direct download
link at the URL above.

After installing and starting LTspice, open the
VTL5C2_TEST.ASC schematic file, and then click on the
'running man' symbol in the toolbar, to run it.  It plots
the R vs I characteristic, like the one from the datasheet. 
To see the VTL5C2 model, either right-click on its symbol in
the vtl5c2_test schematic and select 'Open Schematic', OR,
open the VTL5C2.ASC schematic directly, using either the
menu or the toolbar's 'folder' button.

To extract the 'standard' spice netlist: When the VTL5C2.ASC
schematic is the active window, the netlist can be viewed by
selecting 'View' and then 'Spice Netlist', from the standard
menu bar at the top.  To copy the netlist to the Windows
clipboard, highlight the contents of the netlist window and
strike CTRL-C.  Then you can go to Notepad, or wherever, and
use 'Paste'.

I could also post the same (more-or-less-standard) spice
netlist for the VTL5C2 model, here, if anyone prefers that.

---

Yes, the linearization circuit is interesting.  It provides
two floating Vactrol LDR leads, for use in the user's
circuit.  The resistance between those two LDR leads is
linearly controlled by a control-voltage input.

In my VTL5C2-based prototype circuit, changing the control
voltage, Vctrl, from 0.0025v to 5v changes the resistance
between the two LDR leads, Rldr, from 500 Ohms to 1 megOhm. 
If the two Vactrols were perfectly matched, the Rldr vs
Vctrl relationship would be, theoretically, perfectly
linear.

I wish I could find a way to use one of the 'dual' Vactrols,
with the hope that they would be well-matched.  But, since
they appear to have 'center-tapped' LDRs (3 leads), instead
of two independent LDRs (4 leads), I don't see a way to make
them work, for this.

I designed the linearization circuit by starting with the
fact that putting a 'function' into the feedback loop of an
opamp results in the opamp circuit having the 'inverse
function'.

Before I started, as a sort-of 'proof-of-concept' for that,
I decided to see if I could linearize a standard 1N4148
diode's non-linear I vs V function, in a simulated circuit,
by using a first diode (and an opamp) to create its 'inverse
function', to provide a non-linear drive  for a second
identical diode, in order to try to 'cancel out' the second
diode's non-linear i vs v characteristic.  It worked
perfectly.  Then I realized that I had 're-invented' the
standard 'ideal diode' circuit.  :-)

A Vactrol is not as simple as a diode, however. And it took
me quite some time to figure out how to use one vactrol to
linearize the R vs I response of a second vactrol (but maybe
I'm just slow), and then a lot more time to get the whole
thing to respond well to dynamic Vctrl inputs.

Before I forget, just in case anyone else is interested,
below is the URL of a new paper by Bob Pease, of National
Semiconductor, entitled "A Comprehensive Study of the
Howland Current Pump", which is dated January 29, 2008. 
It's now AN-1515 at national.com:

http://www.national.com/an/AN/AN-1515.pdf

Just for 'completeness', I'll include the other two
current-source papers that I like:

http://apex.cirrus.com/en/pubs/appNote/Apex_AN13U_C.pdf

http://apex.cirrus.com/en/pubs/whitePaper/199210-Apex-Versatile_current_source_circuits.pdf

Regards,

Tom

Tom Gootee
tomg at fullnet.com
http://www.fullnet.com/~tomg/index.html



----- Original Message Follows -----
From: harrybissell at wowway.com
To: tomg at fullnet.com, synth-diy at dropmix.xs4all.nl
Subject: Spam:*************, Re: [sdiy] Linearized Vactrol?
Date: Wed, 12 Mar 2008 13:07:59 -0600

> Hi Tom
> 
> Sounds like an interesating circuit. A technique for
> 'linearizing' a photocell exists in National Semiconductor
> AN-20 "analog multiplier"
> 
> It needs matched cells as well, and is shown with an
> incandescent lamp which really limits the response at very
> high resistances.
> 
> The best deal would be to use a vactrol that is
> center-tapped so that the cell is at the same temperature,
> and reasonably well matched. Some vactrols like the ones
> used in the Mutron III envelope follower have two
> independant elements, this might be a good choice if
> available.
> 
> Vactrols are useful for many things at s-diy, even if they
> are slow. They are reasonable quiet, have no voltage
> offset etc.
> 
> I'd like to see the spice model you have, btw.
> 
> If your circuit is really novel, submit it to EDN magazine
> as a 'design idea' or Electronic Design magazine and get
> paid maybe $150 for it, and be famous. It usually takes
> about 6 months for publishing. They will not take it if it
> has been published (ir public domain) already...
> 
> H^) harry
> 
> 
> On Wed, 12 Mar 2008 12:27:44 -0500, Thomas P. Gootee wrote
> > Hello, everyone.
> > 
> > I am new to this list.  My name is Tom Gootee.  I live
> > in southwestern Indiana, in the USA.
> > 
> > I was an EE, originally, until 1984, then did mostly
> > C-language software for about 15 years, and have been
> > designing (and manufacturing & selling) analog
> > electronic systems for the last ten years or so (mostly
> > test-and-measurement equipment).
> > 
> > I hope that my question is not perceived as too unusual,
> > or inappropriate, for this group. If so, I apologize. 
> > But I don't know of anyplace else where Vactrols are
> > discussed as frequently, or by people whom I respect as
> > much as the members of this group.
> > 
> > -----
> > 
> > I am wondering whether or not there might be any
> > interest in a wide-range, 'perfectly'-linearized
> > voltage-controlled Vactrol-type device (i.e. a
> > wide-range linear voltage-controlled resistance).
> > 
> > I have developed a circuit that does that, which would
> > also work for any other LED/LDR type of device (e.g.
> > Silonex, et al).  I'm just not sure what to do with it. 
> > i.e.  Would it be worth trying to commercialize it? 
> > Would there be enough interest?  Would it be foolish of
> > me to simply 'donate' the circuit into the public
> > domain? 
> > I originally wanted to develop it for use in some of my
> > (non-audio) product designs.  And I also have audio
> > hobbies, and wanted a series/shunt LED/LDR-based audio
> > attenuator with precisely-constant input impedance.
> > 
> > But, for many applications, the vactrol-type devices are
> > way too slow.  So the applicability of a linearized
> > version seems somewhat limited.
> > 
> > Also, I think that there are probably other,
> > somewhat-similar circuits already available, having seen
> > the (fairly horrible) attempt at it that's published on
> > the Silonex website.  So it's probably not unique, or
> > even new.  However, my circuit HAS been structured for
> > optimization of things like maximum resistance-command
> > step-response slew rate, with near-zero overshoot,
> > maximally-wide resistance range for a given LED/LDR
> > device, high precision in general, etc etc, yada yada
> > yada. 
> > But, since it relies on an identical LED/LDR device for
> > its control mechanism, any mis-matches (other than a
> > constant resistance offset) between the target LDR
> > device and the control LDR device will negatively affect
> > its accuracy. 
> > In other words, I guess I'm also thinking that maybe
> > it's really just not all that great, and I might be
> > worrying about all of this for no good reason.
> > 
> > So I'm hoping that someone here will be kind-enough to
> > give their opinion about whether or not it's worth
> > pursuing further, in any way at all.
> > 
> > I'm sorry to have 'blathered-on', for so long, about all
> > of that.
> > 
> > [P.S.  If anyone wants a fairly-detailed LTspice model
> > of the Perkin-Elmer/Vactec VTL5C2 Vactrol, or a model of
> > the Silonex NSL-32SR2, just let me know.]
> > 
> > Thank you.
> > 
> > Regards,
> > 
> > Tom
> > 
> > Tom Gootee
> > tomg at fullnet.com
> > http://www.fullnet.com/~tomg/index.html
> > _______________________________________________
> > Synth-diy mailing list
> > Synth-diy at dropmix.xs4all.nl
> > http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
> 
> 
> Harry Bissell & Nora Abdullah 4eva
> 



More information about the Synth-diy mailing list