Digitally-actuated potentiometers, or digipots, are commercially available from multiple suppliers but not analog-actuated potentiometers, or anapots. The following isolated anapot, or isopot, design is intended as a design “template” for general application wherever a manually-actuated pot is to be replaced and actuated by an analog voltage.
A manual potentiometer has three terminals, labeled VP+, the highest-voltage end, VP− the lowest-voltage end, and VPW, the wiper. The pot is emulated by the following simplified circuit.
The goal is to duplicate the pot wiper voltage, VPW as a fraction, a, of the voltage across the pot,
where a ranges from 0 to 1. The above circuit outputs VPW as the voltage,
At full-scale (a = 1), iO x RO = VP and a is varied by varying iO using a multiplier. The differential V/I converter converts VP to a current input to the multiplier, MULT. This is the full-scale iO that is scaled by the other MULT input, from a V/I converter that converts the control voltage (controlling a) to a multiplier input current.
The following design is limited to non-negative terminal voltages, though by using a negative power supply (instead of ground for −VEE), pot voltages can be readily extended to negative voltages. The scheme is shown below.
The conversion of VP to a current is performed by an op-amp Howland current source. It drives one of two translinear amplifiers of a LM13700 IC, used as a two-quadrant multiplier with current-source output. The current is input to a current mirror connected to VP+. The other half of the LM13700 is used as an input of the ground-referenced analog control voltage (0 V zs, 4 V fs) and outputs to the IX input pin of the multiplier. The Howland circuit sources a (positive) current into the IY input of the multiplier. The isopot circuit diagram is shown below.
Click to view enlarged image
Tracking current sources are implemented using half of dual op-amp, U2, as a V/I converter. The 5.1 V Zener diode is biased at a current that makes its voltage near 0 TC by R3. The op-amp places a reference 5.1 V across R2 in series with trimpot VR1. This pot should be adjusted for correct full-scale output at VPW with 4 V applied to the control input. The total current generated is
It is split among the three BJTs. The currents are apportioned by R13, R14, and R16. Q4 and Q6 conduct 100 μA; Q1 conducts 200 μA.
Howland Current Source
U2A is the op-amp of a Howland circuit that converts an input voltage, VP, to an output current that flows into pin 16 (IY) of U1B. Because op-amp inputs are maintained at the same voltage by feedback, the non-inverting input (pin 2) will be at the output voltage (pin 3). Then VP is across R8 + R7. Because VP+ and VP− are treated as independent voltages, the currents in R7 and R8 will not necessarily be equal unless VP is an isolated voltage source placed across the VP inputs. VP+ will cause a current to flow in R7 that combines with the current from the op-amp output through R4 to be the desired output current. This current will result in some output voltage, VL, at the op-amp inputs causing a current to flow in R6 that also must flow in R5. This current determines the op-amp output voltage, Vo. If VL increases, the inverting circuit of the op-amp will cause Vo to increase, thus bootstrapping R4 to keep the output current the same.
The inverting and non-inverting feedback paths must track to remove the effect of VL on output current. The Howland op-amp circuit behaves as a current source only when the following conditions are satisfied:
The output current, which is sourced from the circuit for the polarity of VP indicated, is
In the next part of this article, the core of the isopot - the translinear circuits - will be described.