AutomateDirectUSAMotion LibraryShop the parts catalog

Troubleshooting

Make a Servo Jog Follow a Joystick Velocity Demand

Servo and stepper motion Published 6 min read

Short answer

A velocity jog reads its speed operand only when the instruction executes, so a running jog ignores the joystick value until you re-execute it. Strobe the jog every scan with the scaled stick velocity and keep a deadband at neutral; any lag left after that is the acceleration limit, not the command path.

On this page
  1. The jog reads its speed once, at execution
  2. Strobe the jog every scan
  3. Separate a missed command from an acceleration ramp
  4. Worked example: full-stick step from 0 to 2.5 in/s at a 1 m/s per second limit
  5. Neutral, sign and units before any tuning
  6. The decision, case by case
  7. Questions asked

A jog instruction is the right tool for a joystick, and it is not a discrete-only command. What makes it feel wrong is one behavior: a velocity jog reads its speed operand when the instruction executes, not continuously. Once the jog is running, a new number in the tag changes the tag, not the command, so the axis holds the speed it started with. Re-execute — strobe — the jog every scan with the current stick velocity; a published two-rung pattern does exactly that with one motion control tag. Everything left after that is dynamics: the axis changes speed at its acceleration limit, so compute the ramp time before blaming the command path or touching tuning.

The jog reads its speed once, at execution #

The application behind this symptom is a Logix motion controller jogging a servo axis, with the stick scaled to a signed velocity. Motion Axis Jog (MAJ) is the velocity jog in that instruction set, and like the other motion instructions it acts on a rung transition: it runs when its rung goes false to true, reads its operands at that moment, and hands the axis to the motion planner. A rung held continuously true executes once. The speed tag stays live; the command does not. That is the whole fault.

The 0–10 V habit from an older actuator points the wrong way here. On a servo axis driven by a motion controller, the analog input becomes an operand and the controller issues the velocity — the same controller-and-drive split behind why a Kinetix 5500 needs a motion-capable Logix controller.

A compare-and-restart workaround — measure the running jog speed, compare it with the stick demand, and stop and restart through the change-dynamics instruction (MCD) when they differ — does update the speed, but only when the compare fires, so quick handle movements land late or land together. Per-scan re-issue removes the compare entirely.

Strobe the jog every scan #

  1. Scale the stick to signed axis velocity. The case at hand: left gives 0 to +2.5 in/s, center is 0, right gives 0 to −2.5 in/s.
  2. Wrap neutral in a deadband — a LIM instruction works — so a handle that does not return perfectly to center commands zero, not creep.
  3. Re-issue the jog every scan with the current demand. The sample-code pattern puts two MAJ rungs under one motion control tag: the upper rung stays false, and the rung below passes the new speed, or the same speed when nothing changed, so the jog's internal state resets every scan and the operand is accepted. The jog's merge behavior governs how a re-issued command blends into motion already running, which is what keeps per-scan re-issue smooth.
  4. Match the direction mode. A signed demand carries direction in its sign and speed in its magnitude; if your configured mode wants a magnitude plus a direction selection, split them before the operands.
  5. Set the stop you want: with the handle centered the strobed demand is 0 in/s and the axis decelerates at its deceleration limit, so that limit is your manual stopping ramp from full speed.
  6. Verify with a trend of the scaled demand and the axis commanded velocity while you sweep the handle: commanded velocity should follow within a scan or two.

Do the direction and feel tests at low speed with the work area clear and the stop function in reach; the safeguards for a manual mode come from the machine's risk assessment (ISO 12100), not from this article.

Separate a missed command from an acceleration ramp #

Trend four signals together: the scaled stick demand, the axis commanded velocity, the axis actual velocity, and the jog rung's state. Two facts sort every case. A servo cannot step to a new speed instantly even when the planner accepts the command immediately. And no acceleration setting makes an unexecuted instruction read a changed operand.

What you see First cause to check Confirming check
Speed tag changes; jog keeps its start speed No new instruction execution No execution event after the operand moves; rung state never transitions
Handle movements land late or in jumps Update gated by a speed-delta compare or a status wait Accepted commands lag demand changes by the compare's dead time
Commanded velocity steps at once; actual ramps on a fixed slope Acceleration, deceleration and jerk limits Ramp time close to velocity change ÷ acceleration limit
Axis creeps with the handle released Neutral drift or noise inside the deadband gap Raw input moves at center while the post-deadband demand stays zero — or does not
Direction or speed magnitude disagrees with the handle Sign mapping or axis scaling Low-speed test in both directions; read the scaling configuration's units

Keep hardware replacement out of this branch. Commanded-versus-actual evidence settles it — the same evidence-first order that decides how to tell if a servo motor is actually faulty.

Worked example: full-stick step from 0 to 2.5 in/s at a 1 m/s per second limit #

Inputs:

  • Joystick scaling for this axis: left 0 to +2.5 in/s, center 0, right 0 to −2.5 in/s, so snapping the handle to full left from rest demands 2.5 in/s.
  • Acceleration limit: 1 m/s per second, an illustrative value — read yours from the axis dynamics (planner) configuration. In axis units, 1 m/s = 39.4 in/s, so the limit is 39.4 in/s per second.

Rule in words: time to reach a new velocity equals the velocity change divided by the acceleration limit; the deceleration limit governs stops the same way.

Arithmetic: t = 2.5 in/s ÷ 39.4 in/s per second = 0.063 s.

Result: about six hundredths of a second — a calculated example for these inputs, not a measured ramp.

Decision: an axis that needs a second or more to follow the handle, while commanded velocity has already stepped, is not limited by an acceleration value this size — the demand never became a command, so fix the execution. An actual velocity that trails commanded by roughly 0.06 s is the limit doing its job; raise the acceleration limit in the dynamics configuration for a snappier manual feel.

Neutral, sign and units before any tuning #

  • Confirm the deadband rejects real drift: trend the raw input and the post-deadband demand with the handle released, before enabling continuous manual jogging.
  • Verify direction at low speed in both directions. Reversed direction is a mapping fix, not a drive fix. Correct direction with wrong magnitude is a scaling fix: read the axis scaling configuration, which converts counts to engineering units, and the stick conversion feeding it.
  • Keep three numbers separate: the operator-facing percent of maximum, the software velocity cap on the axis, and the physically permitted manual speed from the machine's requirements.
  • Tuning is the last branch, and it is per-axis work — axis loop tuning differs between axes enough that one set of gains does not transfer.

The decision, case by case #

  • Commanded velocity stuck at the start speed while the demand moves: the jog is not re-executing. Strobe it every scan.
  • Updates land only on large handle moves: delete the compare-and-restart gate; per-scan re-issue replaces it.
  • Commanded steps and actual lags by about velocity change ÷ acceleration limit: set finite acceleration, deceleration and jerk values for the ramp you want, and leave the command path alone.
  • Creep at center: widen the deadband past the measured drift.
  • Wrong direction: fix the sign mapping. Wrong speed magnitude: fix the axis scaling.
  • Commanded velocity tracks the stick and the ramps match the configured limits, but the response still disappoints: now, and only now, look at loop tuning — per axis.

Questions asked

Should the MCD-based speed update be kept once strobing works?

Delete it. That pattern only executed when the speed difference crossed your compare point, which is why updates were coarse and infrequent. The change-dynamics instruction still has a job, altering the dynamics of an active move of another type, and it too takes effect only when it executes.

How wide should the neutral deadband be?

Trend the raw stick input for several seconds with the handle released, then set the limit band wider than the peak-to-peak drift and noise you recorded. Going wider than that only gives away fine control near center. Re-check the band after any joystick or input-module replacement.

What acceleration value gives good manual feel?

Choose the ramp time first and divide: acceleration equals full manual speed divided by that time. For 2.5 in/s reached in 0.25 s, that is 10 in/s per second. In metric, a change from 1 to 3 m/s at 1 m/s per second takes 2 s by the same rule. Finite acceleration, deceleration and jerk values make the ramp repeatable where auto-calculated ones do not.

Should the operator see percent instead of units per second?

Percent of permitted manual speed works well for operators and keeps engineering units in the logic. Keep three numbers distinct: the display percentage, the software velocity cap in the axis configuration, and the physically permitted manual speed from the machine's requirements. The cap is what limits the axis; the display scale limits nothing.

Can one signed joystick value drive both direction and speed?

Only if the configured direction mode accepts it. Where the mode wants a magnitude plus a direction selection, split the sign out first: a negative number in a magnitude operand gives wrong-direction motion or an instruction fault. Read the help for the exact mode you configured before wiring operands.

Keep reading

  • Selection guide

    Choosing Servo or Stepper for a 1% Speed Regulation Pinch Wheel

    Pick the servo: behind a 45:1 planetary reducer, a velocity-mode servo meets a 1% wheel-speed tolerance with margin, because 200-300 lb-in at 80 rpm needs only 4.9-7.4 lb-in at the motor and the loop corrects load-induced error. A stepper qualifies only when its torque-speed curve clears the load with margin at the required motor speed and a measured ripple test holds the wheel inside 0.8 rpm.

    Servo and stepper motion

  • Troubleshooting

    Why a Servo Motor Oscillates Between Two Positions

    A servo that hunts in a stable 30-50 degree band and restarts on a light shaft nudge has a working position loop running on wrong motor or load data. Fix the configuration — engineering units, load description, pole count, resolver excitation — before swapping phases or retuning: in the documented case, converting metric values into the drive's lbf-ft fields and correcting the load entry stopped it.

    Servo and stepper motion

  • Troubleshooting

    Brushed DC Servo Tach Dropout Diagnosis: Why It Jumps

    Jump at one shaft angle plus a speed-synchronous lope after tach work is a dropout in the tach's voltage, not an encoder or alignment fault. Prove it by spinning the feedback shaft and looking for a once-per-revolution notch in tach voltage; a cracked armature conductor needs a replacement armature or feedback unit, not another coat of insulating varnish.

    Servo and stepper motion