Design Article
Comment
anonymous user
1.The opto-coupler linearization task looks like solved well, but a task for ...
anonymous user
This technique is hardly new. Bell Labs employed a Western Electric ...
Use a photoelectric-FET optocoupler as a linear voltage-controlled potentiometer
Sajjad Haidar, University of British Columbia, Vancouver, BC, Canada; Edited by Paul Rako and Fran Granville
1/19/2012 11:00 AM EST
You can use a photoelectric FET
as a variable resistor or a potentiometer
in combination with a fixed
resistor. The H11F3M photoelectric
FET has an isolation voltage of 7.5 kV,
enabling you to safely control highvoltage
circuit parameters. The nonlinear-transfer characteristics of these
devices are problematic, however
(Figure 1). To correct the nonlinearity,
using a simple feedback mechanism as
a potentiometer yields a linear response
(Figure 2). This circuit uses two photoelectric
FETs—one for feedback and
the other for applications requiring an
isolated potentiometer. You connect
the inputs of the two photoelectric
FETs in series to ensure the same
amount of current for the input LEDs.
Place 50-kΩ resistors at the FET
outputs to mimic the response of a
potentiometer. The circuit amplifies
the difference between the set input
voltage, which you adjust using potentiometer
R7, and the feedback from
photoelectric FET 1. The resulting output
controls the current in the photoelectric-FET LEDs until the feedback
voltage equals the input voltage. The
output voltage follows linearly with the
input voltage (Figure 3). You might
think that photoelectric FETs bearing
the same part number are identical, but
small manufacturing discrepancies can
be present. Five H11F3M parts have
offsets within 3%.
|
|
|
|
|
Navigate to related information
anonymous user
1/19/2012 5:05 PM EST
Good as far as it goes but the transfer function of a photofet is a lot more linear as a conductance than as a resistance. In some applications you might not even need the matching device and closed loop to create an isolated variable conductance.
Sign in to Reply
anonymous user
1/20/2012 11:20 AM EST
Good Idea, however, the circuit is stable in any condition?
Sign in to Reply
Luca Bruno
1/20/2012 11:20 AM EST
Good Idea, however, the circuit is stable in any condition?
Sign in to Reply
Jeffrey L
1/23/2012 1:24 PM EST
The circuit seems look like a buffer OPAMP which signal should input to "+" and the feedback to "-". However, the Photo-FET is an inversion as used in this circuit, i.e. more input (current,) less output (resistance). So, the author is smart enough to swap the inputs to cancel out the feedback polarity issue.
Sign in to Reply
anonymous user
1/23/2012 1:46 PM EST
One of the reader commented that, the conductance of the photofet is linear. However my measurements show that conductance is linear in some limited range. That is what I found with the device used.
Sign in to Reply
bcarso
1/23/2012 6:40 PM EST
I believe CP Clare made a dual optoFET coupler with things arranged so the LED illuminated both devices, and they probably had a spec on the matching. Might be more trustworthy than using two different packages.
JFETs do make pretty poor voltage-controlled resistors as far as things that handle a.c. signals, although there are ways to configure them for better linearity. However, for this application it doesn't matter much.
Sign in to Reply
anonymous user
1/24/2012 4:04 AM EST
The original idea is basically interesting, and arranging things properly could appear to be successful even with phototransistor couplers.
There's however a problem, with the transfer characteristics deviance, from part to part: different couplers can have a significantly different current transfer ratio. So the feedback will adjust output voltage on the feedback opto, while on the output optocoupler current and voltage won't be as expected.
Here's why, back in the 90s, was made the IL300 by Siemens, as a linear device. Today is produced by Vishay, Avago has a similar part (HCNR200), and Clare as well (as mentioned in another comment, LOC111).
Having emitter and receivers (in this case photodiodes) in the same package helps with coupling, since the current to emission ratio is just one.
Sign in to Reply
anonymous user
1/24/2012 9:31 AM EST
This technique is hardly new. Bell Labs employed a Western Electric single-emmiter, dual detector optical coupler and op amp to linearize optical coupler transfer characteristics in the late 1970s.
Sign in to Reply
anonymous user
1/27/2012 2:36 PM EST
1.The opto-coupler linearization task looks like solved well, but a task for matching opto-couplers characteristics is not completed.
2.It hard to believe that H11F3M which was designed at about 20 years ago could provide chip to chip matching in a few percent.
3.To do that needs to add an additional circuitry that will adjust difference between a master (photo FET1) and a slave (photo FET2).
Sign in to Reply