How to Verify a Rod-Pump Dynamometer Sensor Install

Back to Blog

Confirm the Load Cell Is in the Load Path and Square

The polished rod load cell only reads true if the entire rod load passes through it and nothing shares that path. On a horseshoe or donut cell that sits under the carrier bar, verify the cell is centered on the polished rod, that the carrier bar bears fully on the cell face and not on a corner, and that no clamp, spacer, or safety collar is carrying part of the load around the cell. A cell loaded off-center or in a partial load path reads low and nonlinear, and the error changes through the stroke, which distorts the whole card rather than just shifting it.

Check the mechanical clearances the same way you would before energizing any rotating or reciprocating equipment. The cell and its cable must have room to move with the carrier bar through the full stroke without pinching, rubbing the bridle, or fouling the stuffing box. A cable that snags at the top of the stroke will yank the connector and eventually open the circuit, and a cell that contacts the wireline hanger at bottom will pick up a spurious load spike. Confirm the fixed reference the position sensor uses is genuinely fixed and not something that flexes under load.

If the install uses a strain-gauge horseshoe transducer clamped to the polished rod itself rather than a carrier-bar cell, the verification shifts to the clamp. The transducer must be torqued to the manufacturer's value on a clean, undamaged section of rod, oriented so its measuring axis is along the rod, and positioned where it will not hit the stuffing box or the carrier bar at either end of travel. A loose or misoriented clamp reads a fraction of true load and drifts as it works loose.

Prove the Position Reference at Both Ends of the Stroke

The position signal tells the controller where in the stroke each load reading was taken, so a wrong position reference twists the card even when the load cell is perfect. Whether the position comes from an inclinometer on the walking beam, a Hall-effect sensor watching a magnet on the crank, or a potentiometer on the pitman, verify it reads a consistent zero at the bottom of the stroke and full scale at the top. Cycle the unit by hand or at slow speed and watch the raw position value track smoothly from one end to the other with no dead spots or reversals.

The direction matters as much as the endpoints. Confirm that the position value increases as the polished rod rises, because the controller assumes the upstroke and downstroke sit in a known half of the card. If position runs backward, the plotted card is mirrored and the software will misread where load is applied, turning a healthy full card into something that looks like a pump problem. Fix direction in the mounting or the configuration before you trust any diagnosis.

For crank-angle position sensing, verify the sensor sees exactly one pulse or one full cycle per pump stroke, not two or a fraction. A magnet placed where the sensor catches it twice per revolution, or a gap that lets it miss occasionally, gives the controller a stroke count and a card that jump around. This is also the place to confirm your strokes-per-minute reading is credible, because the same position pickup usually feeds the strokes-per-minute count the controller reports.

Capture a Card and Sanity-Check the Shape

With load and position both verified, let the unit run at normal speed and capture a surface card. The output is a dynamometer card, a closed loop of load versus position for one complete stroke, and a correctly installed system produces a recognizable shape: load is higher across the upstroke, drops during the transition at the top, sits lower across the downstroke, and rises again at the bottom. If the loop is inverted, open at one end, or collapsed to a line, the problem is almost always the install and not the pump.

Check the absolute load numbers against what the well should carry. The minimum load on the card should be near the buoyant rod weight and the maximum near rod weight plus fluid load, both of which the pumping engineer can estimate. A card whose whole load band is shifted far from those estimates points to a load cell that needs zeroing or a scaling error in the configuration, not a downhole condition. Verifying the numbers here prevents a calibration error from later being misread as a fluid-load change.

Finally, take several consecutive cards and confirm they overlay closely. A well-installed, correctly wired system repeats stroke to stroke; cards that wander in load level, drift in position, or change shape at random indicate a loose mount, an intermittent connection, or electrical noise on the signal. Only once the cards are stable and shaped right should you enable the controller's pump-off logic and start trending the well.

The checklist below is the pass/fail sequence to run before you trust the install and arm the controller.

• Load cell centered and square, carrying the full rod load with nothing bypassing it.
• Cell and cable clear through the full stroke with no pinching or snagging.
• Position reads a repeatable zero at bottom and full scale at top.
• Position value increases as the polished rod rises (correct direction).
• One clean position count per stroke, matching a manual strokes-per-minute timing.
• Resting load near buoyant rod weight; peak and minimum loads near the engineer's estimates.
• Card is a closed loop with load high on the upstroke, repeating stroke to stroke.

Common Mistakes

The most common error is accepting a card because it looks like a card, without checking direction and scale. An inverted position signal produces a mirror-image card that a hurried commissioner reads as a real pump condition, and the mistake propagates into every later diagnosis. Always confirm load rises on the upstroke before interpreting anything.

The second is leaving a load-sharing element in the path - a safety clamp resting on the carrier bar, or a bridle strand still partly loaded - so the cell reads a fraction of true load. The card looks fine in shape but every load number is low, which later reads as a phantom fluid-load problem. The third is not securing the cable, so it works loose over days and the signal degrades after the crew has left.

Frequently Asked Questions