V-I characteristics of CdS photo resistors at constant ilumination

[IconicPCB]

Looking for information V-I characteristics of CdS photo resistors at a constant illumination.

Specifically how non linear is the resistance as the function of voltage applied to the photo resistor?
How temperature sensitive is it?

[David Hess]

I know their V-I characteristic is more stable than a FETs as the operating point is changed and when stable temperature characteristics are desired, they are used in pairs.  Compared to FETs when used as gain control in an oscillator, they contribute essentially no distortion so replacing them with FETs is not easy.

[IconicPCB]

Thanks David,

I have seen them used in ultra low distortion Wien Bridge oscillators as gain control elements.
The oscilator boasted distortion down in 0.0001% figures.

A reference oscillator to be used in testing 20bit ADCs.

[BrianHG]

If you are doing what I think you are doing, you can use a 3 pin output vactrol.  Use the center-tap of the photocell tied to GND.  One ouptut should be pulled up by say, 10k, with that output tied to the feedback of an op-amp with the + input setting to set the resistance you want while the output of the opamp would feed the led/lamp.  The third pin of the photocell would be used as your custom, precision compensated resistor to the common center GND.

In this configuration, you are guaranteed the set resistance you want, across multiple vactrol photocells, across multiple temperatures, across multiple source led source differences and it also will compensate for aging of the components.  I used to use these photocells for ultra clean precision audio pre-amp volume controls.

[T3sl4co1l]

They're pretty good, but if you need ppm or better linearity, I don't know.  You probably want similar consistency, too?  You'll be better off pulling that into the digital domain...

Tim

[Kleinstein]

The amplitude stabilization of a Wien bridge oscillator usually sends only a small part of the signal through the variable part (FET, CdS cell). The coarse adjustment is done with fixed resistors and maybe a pot. Only with a variable frequency the adjustment range needs to be little larger.
So the distortion of the generator can be much better than the linearity of the control part.

[PChi]

My experience of them was that the V-I characteristics aren't linear.
The Silonex web site, now gone, had some measurements. If the voltage across the photo resistor is kept low then they are relatively linear, better than FETs.

[IconicPCB]

Fixed frequency, preset gainto bring the oscilations on, CdS cell in parallel with a low value resistor to provide final AGC/levelcontrol.

[BrianHG]

Like I said: http://www.alliedelec.com/m/d/a9086e5f49f2112449f28d976a47bf0b.pdf
or http://www.datasheetlib.com/datasheet/577849/vtl5c2-2_excelitas-technologies.html
These dual channel linear photocells are old and you'll only find them on ebay.

Use 1 side of the cell to negative feedback an op-amp driving the LED, send you control into the + of the op-amp.  This guarantees the internal resistance of the photocell to withing the accuracy of the op-amp and your reference resistor.

Use the third terminal of the photocell to apply load/gain to your source signal.  You never need worry about the age & drift of the photocell since 1/2 of it's surface is being measured and the led brightness is being compensated by the op-amp in real time creating a 'PERFECT' voltage - to - resistor element no matter which photocell you happen to purchase.  It will be a perfect linear voltage to resistor converter.  The only negative part of this solution is that one side of your voltage programmable resistor will be connected to GND, or, a fixed voltage of your choosing.

[David Hess]

If a photocell is used in a closed loop application like AGC (automatic gain control) or gain control of a Wien bridge (same thing), then large signal compensation using a second photocell is not required.  Dual photocell linearized couplers are useful for bridging galvanic isolation barriers or maybe doing level shifts.  I have seen the same linearization scheme used for FETs to make open loop gain controls.

Thanks David,

I have seen them used in ultra low distortion Wien Bridge oscillators as gain control elements.
The oscilator boasted distortion down in 0.0001% figures.

A reference oscillator to be used in testing 20bit ADCs.

A two operational amplifier Wien bridge using an incandescent lamp for gain control can get close to that level or maybe better.  The second operational amplifier is used to suppress the common mode signal of the Wien bridge amplifier.  Linear Technology application note 43 discusses this on page 29.

Check out the oscillator design starting on page 62 of Linear Technology appliaction note 67 for a design intended for high resolution ADC testing.

Besides another low distoration design, this article has a bunch of dead links to follow.