Wiring a Potentiometer to a 4-20mA Actuator Input Loop

Which quick fixes fail on a 4-20 mA positioner?

Four approaches get proposed every time a hand backup is needed for a positioner that lives in a confined space. Three of them break in the field.

The first is a plain speed pot wired the way it is wired next to a VFD HOA switch. That works because a drive accepts 0-10 V from a resistive divider fed by its own reference supply. A positioner with a 4-20 mA-only command input has no such reference and no voltage input, so a three-terminal divider hanging in the panel produces nothing the actuator will act on. The pot has to sit in series with a supply and the positioner input, as a rheostat, or feed a converter.

The second is adding a second analog channel and switching sources in PLC logic. That is the cleanest architecture right up to the point where you write down the design assumption: the backup exists because the controller may be dead, isolated, or compromised. Any scheme whose transfer decision, power, or scaling lives inside that controller inherits its failure. Reject it on the requirement, not on the wiring.

The third is buying a handheld loop calibrator and leaving it in the panel. It solves the signal problem and creates a maintenance problem: battery state, someone walking off with it, no permanent terminations, no labeled operator interface at 2 a.m. Handhelds are commissioning tools. A backup control element belongs on the DIN rail or the door.

The fourth is a bare DPDT switch dropped between the PLC analog output and a second current source. It is the right idea and the wrong hardware, for reasons that have nothing to do with the pot. See the transfer section below. And the reflex mitigation - a capacitor across the outgoing 4-20 mA pair - does not fix it. It slews the step, stores energy that discharges into whichever source is connected next, and on a HART positioner it loads the modem band. Look at what the switch does to each source's compliance before you add components to the loop.

What does the actuator actually see?

Work the chain from the command source to the final element. Every wrong value in this loop shows up as valve position, not as an alarm, unless the positioner is smart enough to flag it.

Signal	Source	Symptom when wrong
Position command, 4-20 mA	PLC analog output (auto)	0 mA on PLC power loss or output fault: positioner executes its loss-of-signal action, typically driving to 0% and closing the valve
Position command, 4-20 mA	Panel loop source or pot + series resistor (manual)	Below ~3.8 mA a smart positioner declares under-range and may hold or fail rather than track; above 20.5 mA it saturates at 100%
Loop supply, 24 VDC	Panel power supply	Current scales directly with supply voltage in a passive pot scheme: 5% supply error = 1 mA error at the 20 mA end
Loop return / common	PLC common vs panel 0 V	Switching only the positive leg leaves both returns tied: parallel return path, shunted current, position reads low and drifts
Positioner input impedance	Positioner terminals	Left out of the resistance budget, the 20 mA end never reaches full scale and the valve stops short of full open
Position feedback	Positioner 4-20 mA output / local indicator	With the PLC down there is no readback at the HMI: the operator is driving blind unless a local pointer, limit lamps, or a panel mA display exists

Does the series pot math survive a real loop?

Ohm's law sets the whole design. With a 24 V supply and the pot in series with the positioner input:

I = V / R_total
R_total = R_fixed + R_pot(set) + R_positioner + R_wiring

24 V / 0.004 A = 6000 ohm   -> 4 mA endpoint
24 V / 0.020 A = 1200 ohm   -> 20 mA endpoint

So the loop resistance has to be adjustable from 1.2 kohm to 6.0 kohm, total, including everything already in the circuit. A 10 kohm pot with a 1 kohm fixed resistor covers that band, but badly: full counter-clockwise lands near 24 mA, full clockwise near 2.2 mA, and the usable 4-20 mA span occupies roughly the first half of rotation. Sizing the fixed resistor at the 20 mA endpoint and the pot at the difference is tighter - about 1.2 kohm fixed and a 4.7 kohm pot puts 20 mA at one mechanical stop and just under 4 mA at the other. Measure the positioner's input resistance (or take it from its datasheet terminal specification) and subtract it from the fixed value; a couple hundred ohms of positioner input is a real part of that 1.2 kohm.

Two properties of this circuit deserve respect. Current versus resistance is hyperbolic, not linear: dI/dR is about 16.7 uA per ohm at the 20 mA end and 0.67 uA per ohm at the 4 mA end. That is roughly 25:1 more mA per degree of rotation near full open than near closed. A single-turn pot gives coarse control exactly where the valve is doing the most work; a 10-turn pot with a counting dial fixes it. Second, the current tracks supply voltage one-for-one, so feed it from a regulated 24 VDC rail and trim the fixed resistor against a meter rather than trusting the nameplate voltage.

Dissipation is not a problem but the part rating still is: peak power in the pot element is about 0.14 W with a 24 V rail and a 1 kohm series resistor, which rules out miniature trimmers with 0.1 W elements and comfortably suits a panel-mount pot. Wire the pot as a rheostat - wiper strapped to the unused end terminal - so a dirty or lifted wiper contact cannot open the loop and slam the valve to its fail position. Put a loop-powered 4-20 mA display or a pair of disconnect test terminals in series so the operator sets a number rather than a dial position.

One caution on smart devices: a HART positioner does not present a fixed resistance, and its input impedance is specified for a driven current source, not for a resistive divider. If the actuator is HART-enabled, use an actual current source rather than a passive pot string.

Why can the transfer switch kill the positioner electronics?

A current source is defined by what it will do to get its setpoint current through the load: raise terminal voltage up to its compliance limit. Two of them fighting over one load, or one of them abruptly open-circuited, is where the damage comes from.

Make-before-break contacts parallel the PLC analog output and the manual source for a few milliseconds. Each source now sees a load that is sinking the wrong current, both drive toward compliance, and the loser can be back-fed above its terminal rating. Break-before-make avoids that but leaves both sources momentarily open, and an unloaded analog output sits at full compliance voltage until it is reconnected - that stored step lands on the positioner input at the instant of transfer.

Mitigate it at the switch, not with a capacitor:

• Use a break-before-make, 4-pole switch and open both the positive leg and the return of the source being dropped. That kills the ground loop between PLC common and panel 0 V at the same time.
• Park each idle source on its own burden resistor - a 250 to 500 ohm resistor on the normally-open side of the switch - so neither source is ever driven open-circuit. The PLC keeps outputting into a load and never reports an open-loop fault.
• Fuse the manual branch and keep the manual source galvanically isolated from the PLC output.
• Confirm the positioner's loss-of-signal / fail action parameter before you transfer anything. If dropping to 0 mA drives the valve closed and that is acceptable for this process, document it; if it is not, the transfer must be fast enough that the positioner never times out into fail action.

Which hardware class fits a permanent manual station?

Approach	How it makes the mA	Permanent panel fit	Watch out for
Pot + series resistor off 24 VDC	Passive rheostat, Ohm's law	Yes, cheapest	Supply-voltage dependent, non-linear dial, unsuitable for HART inputs
DIN-rail 4-20 mA generator / loop simulator	Active regulated current source, 24 V powered	Yes	Adjustment is usually a trimpot behind the door, not an operator control
Panel-mount manual loading station	Digital setpoint to isolated 4-20 mA output	Yes, operator-facing	Best fit; see the Precision Digital PD460 family at predig.com/family/pd460
0-10 V to 4-20 mA signal converter fed by a pot	Pot divides a reference, converter drives the loop	Yes	Needs a stable reference for the divider; two devices to power and calibrate
Programmable field/DIN transmitter driven by a pot	Transmitter linearises a resistance input to 4-20 mA	Yes	Configuration software and a resistance-input model, e.g. Moore Industries THZ3 at miinet.com
Smart positioner with local/off/remote switch and open/close/stop buttons	Command generated inside the positioner	Positioner-side, no loop transfer at all	Requires actuator access - defeated here by the confined space
Small standalone PLC/HMI as a hand-auto station	Reads the auto mA in, outputs mA to the valve, bumpless transfer in logic	Yes	Reintroduces a programmable device: only acceptable if it is separate and network-isolated from the suspect controller

Note the last row against the stated threat model. If the concern driving this project is the main PLC being unavailable or compromised, a second controller is acceptable only when it is physically and network-separate from the first. The two positioner-side options in the table are technically superior - a positioner with dual command inputs and a local/remote input, or one with an LOR switch and open/close/stop pushbuttons, removes the transfer switch entirely - but neither helps when the whole point is to avoid confined space entry.

How do you wire and commission the transfer station?

1. Read the positioner's loss-of-signal parameter and its input impedance and compliance requirement from the actuator manual. Record what the valve does at 0 mA before anything is built.
2. Measure the actual loop resistance the PLC output currently drives: open the loop at the panel, insert a DMM in series, and log the mA at 0%, 50% and 100% command.
3. Size the manual source. For an active loop source, set its output span to 4.00-20.00 mA. For a passive pot string, size R_fixed = 1200 - R_positioner - R_wiring ohm and pick a multi-turn pot of roughly 4.7 kohm.
4. Wire the pot as a rheostat with the wiper tied to the unused end terminal.
5. Install a break-before-make 4-pole transfer switch. Pole 1 and 2 switch the positive and return of the PLC output; pole 3 and 4 switch the manual source. Land a 250-500 ohm burden resistor on each source's normally-open contacts so neither is ever open-circuited.
6. Put a loop-powered mA indicator or a disconnect terminal pair in series with the positioner leg, ahead of the transfer switch, so the reading is valid in either mode.
7. Fuse the manual branch at the panel supply and label the switch AUTO / MANUAL with the direction of increasing valve opening on the dial.
8. Post the transfer procedure on the panel: which position closes the valve, what the operator watches, and how to hand control back.

How do you prove it works before you need it?

A backup that has never been stroked is a rumor. Bench the source first, then the loop, then the valve.

1. With the positioner disconnected, terminate the manual branch into a 250 ohm resistor and confirm 4.00 mA and 20.00 mA at the mechanical stops. Trim the fixed resistor, not the supply.
2. Sweep the pot slowly and watch for discontinuities - a dead spot or a jump is a worn element, and no amount of trimming corrects it.
3. Reconnect the positioner and repeat the endpoints under real load. If 20 mA no longer reaches full scale, the positioner input resistance is larger than assumed; reduce the fixed resistor.
4. With the mA meter in series, throw the transfer switch both directions and watch the excursion. Nothing above roughly 22 mA, and no interval below 3.5 mA long enough for the positioner to enter its loss-of-signal action.
5. Stroke the valve at 25%, 50%, 75% and 100% in manual and compare to the local position indicator or the positioner feedback output. Log the dial numbers against actual position and put that table on the panel card.
6. Transfer back to AUTO and confirm the PLC analog output channel reports no fault and the valve tracks the automatic setpoint.
7. Exercise the transfer on the same schedule as any other standby equipment. Contacts corrode and pots seize in panels that are never touched.

Look at the loop current before you touch the pot. If the valve is not where the dial says it should be, the resistance budget or the return path is wrong - trimming the dial only hides it.

When do you stop and call the manufacturer?

Stop building panel hardware when the actuator is a HART or bus-addressed positioner, when this valve carries any part of a safety instrumented function, or when the loss-of-signal behavior in the manual is ambiguous about what happens between 0 mA and 3.6 mA. Those cases are answered by the actuator manufacturer's official support with the positioner serial number in hand - dual-command-input and local/remote firmware options often already exist for the installed model and remove the transfer switch entirely. For the loop source itself, take the compliance voltage, isolation rating and output span questions to the instrument vendor's application support rather than inferring them from a catalog page.

FAQ