Linear potentiometer provides nonlinear light-intensity control
Circuit matches light control to eye’s intensity-response curve.
Stephan Goldstein, Analog Devices, Wilmington, MA; Edited by Brad Thompson -- EDN, April 28, 2005
The human eye's highly nonlinear response to light levels poses problems for designers of adjustable lighting. Simple hardware or software linear-control methods compress most of the apparent intensity variation into a relatively small portion of the adjustment range. A strongly nonlinear control characteristic is necessary. Such a characteristic spreads the intensity adjustment over a wider range and offers a more natural feel. This Design Idea shows how to use an inexpensive linear potentiometer to develop a satisfactory hardware technique. In an experiment in a darkroom, one of the room's corners was too dark because a fixed barrier shielded safe light. Using spare parts from a junk box, you could assemble a simple red LED-based auxiliary safe light, but if the light level were adjustable, you could balance the light levels and minimize the risk of fogging the printing paper. However, the experimenters in this case lacked an audio-taper intensity-control potentiometer and wanted to avoid paying for one.
Figure 1 shows a simplified version of the technique. Diode-connected transistor Q1 and an AD589 1.235V reference, IC1, produce a reference voltage of 1.235V+VBE(Q1) at Node A. Connected between Node A and Q2's emitter, linear potentiometer R2 and resistor R3 cause Q2's emitter and collector current to vary as 1.235V/(R2+R3). The relationship isn't exact because the VBE voltages of Q1 and Q2 vary slightly as you adjust the potentiometer, but, in practice, this nonlinear—if not logarithmic—characteristic works well.
Transistor Q2's collector current generates the control voltage across R4, and, whereas Q2 always operates close to saturation, the components limit Q2's collector-base forward bias to an acceptable 200 mV. When you set R2 to its minimum resistance for maximum light intensity, resistor R3 limits LED current, and, when you set R2 to its maximum resistance for minimum intensity, R1 limits the current through IC1.
The reference voltage produced at Q2's collector drives a standard integrating servoamplifier comprising an AD8031 rail-to-rail op amp, IC2; an IRFD010 low-power MOSFET, Q3; R5; R6; and C2. The servo sets the current through R5 to R4/R5 times the current through R4. Resistor R7 isolates Q3's gate capacitance to prevent load-induced instability in IC2. A 12V-dc module supplies power to the circuit and allows the use of four LEDs per string, for a total voltage drop of approximately 8V across each string. To prevent current hogging and provide a maximum of approximately 20 mA for each series-connected LED string, resistors R8 through R11 divide Q3's drain current into four. Voltage drop across each resistor is 1V, leaving Q3 to support a 3V drain-source voltage and an approximately 250-mW power dissipation. If you increase the number of LEDs or the power-supply voltage, you may need to replace Q3 with a higher dissipation MOSFET.
Check out our Best of Design Ideas section!
-
In the 28 April 05 issue of EDN Magazine you published a Design Idea for a quasi-logarithmic light control using a linear potentiometer. In the article, the authors wanted a better feel and more natural change of light vs. the turning of the linear potentiometer that they had on hand. The eye, similarly to the ear, has an essentially logarithmic sense to increasing or decreasing input levels. The authors stated, however, that they did not want to have to buy a log pot for the purpose. Thus they developed the circuit shown in the Design Idea.
I think they could have saved themselves a lot of work. One can easily create very close to logarithmic action with a linear potentiometer by connecting a fixed resistor with a value equal to the full value of the potentiometer between one end of the linear potentiometer and the adjustable tap of the control. The action is not exact but is surprisingly close to a true log pot.
Granting that their idea was a fun exercise, but at the same time noting the parts count and the specific parts used, I think they might have been better off just buying the log pot. Or using my suggestion.
Thanks for your time.
Gary Sharpe - 2005-6-5 11:34:00 PDT -
Very neat circuit! What is the purpose of R6 in the circuit, why is it 10K? Thanks!
Richard - 2005-3-5 15:17:00 PDT
