Problem Details
Symptom: on a multihead (combination) weigher, one or more hopper channels will not hold zero. You execute a zero/tare on the affected module, the displayed value settles briefly, then wanders — drifting up or down, jumping, or oscillating around zero with no product in the hopper.
Because a multihead weigher combines several hopper weights to hit a target, a single unstable channel corrupts every combination that includes it. The controller either rejects combinations (throughput collapses) or passes out-of-tolerance packs. Operators frequently mask this by re-zeroing repeatedly or by widening the acceptance window, which hides the fault instead of fixing it.
Root Cause Analysis
A weigh hopper channel is a simple chain: hopper body → hopper hanger/hook → load cell beam → load cell mounting block → module frame → main frame. The strain gauge bridge only reports true weight when the only force path from hopper to frame runs through the load cell beam. Any parallel path, or any joint that changes stiffness as the machine vibrates, produces drift.
| Cause | Typical evidence | Class |
|---|---|---|
| Load cell body or free end resting/rubbing on the module housing, cover, guard or cable | Drift changes when you lightly move the cable or press the cover; value shifts after cover is refitted | Mechanical |
| Loose mounting screws at the load cell fixed end or the hopper hanger | Value steps when the frame is tapped; screw turns by hand; witness marks / fretting around the hole | Mechanical |
| Product build-up, film, sugar/oil residue or a stray screw bridging the load cell gap | Drift appears after a production run and disappears after wash-down | Mechanical |
| Hopper hung incorrectly, twisted, or contacting the neighbouring hopper / dispersion cone | Fault clears if the hopper is removed and the empty channel is stable | Mechanical |
| Overloaded / permanently deformed load cell beam | Zero offset will not return after unloading; hysteresis on repeat test weights | Mechanical / cell |
| Water ingress into the cell body or connector after wash-down | Slow one-directional drift; recovers after drying | Electrical |
| Damaged per-module electronics board (A/D converter, excitation supply, amplifier) | Fault follows the module position when the load cell is swapped; excitation voltage unstable | Electronic |
| Cable chafing, broken shield, or connector corrosion between cell and module board | Value jumps when the harness is flexed with the machine running | Electrical |
| Vibration coupling from the linear feeders or a nearby unbalanced machine | Drift is periodic and correlates with feeder amplitude / another machine starting | External |
Most multihead weighers of this architecture carry one electronics board per weigh module, with the load cell wired locally to that board and only a digital link back to the main controller. That layout is what makes the swap test below decisive: it lets you separate the transducer from its front-end electronics without touching the rest of the machine.
Diagnostic Procedure
- Log the raw count, not the displayed weight. Most weigher HMIs expose a service/diagnostic screen showing raw A/D counts or internal mV per channel. Record the raw value for the suspect channel and for two known-good channels over at least 60 seconds, machine stopped. Noise on all channels points to a common cause (power, frame vibration, grounding); noise on one channel points to that module.
- Empty-channel stability test. Remove the weigh hopper from the suspect module and re-zero. If the bare channel is stable and drift returns only with the hopper fitted, the fault is in the hopper, hanger or hopper-to-neighbour clearance — not in the cell.
- Tap test. With the hopper refitted and the channel zeroed, tap the module frame lightly with a soft mallet. A stable channel returns to the same zero. A channel that settles to a different zero each time has a loose joint or a friction (stick-slip) contact.
- Fastener audit. Check every screw in the load path: load cell fixed-end bolts, free-end/hanger bolts, module-to-frame bolts. Re-torque to the OEM value. Do not guess a torque figure — an over-torqued load cell mounting distorts the beam and creates permanent zero offset.
- Clearance inspection. Verify the free end of the load cell and the hanger move without touching anything: housing, splash cover, harness, cable tie, seal, or product residue. Slide a thin feeler/paper strip through the gaps. Any contact at all is a fault.
- Cable flex test. With the channel zeroed and displayed live, gently flex the load cell cable and its connector. Any value change indicates a broken conductor, damaged shield or a corroded connector pin.
- Excitation and bridge check. With the cell disconnected from the board, measure the bridge resistances at the cell connector: input (EXC+ to EXC−), output (SIG+ to SIG−), and insulation from each conductor to the cell body. Compare against the values on the load cell datasheet or calibration certificate for that specific cell. Then reconnect and confirm the excitation voltage at the cell is present and steady.
-
Swap test (the decisive step). Swap the suspect load cell with a known-good cell from another module, leaving the boards in place. If the fault stays with the module position, the module electronics board or its harness is at fault. If the fault follows the cell to the new position, the cell is at fault.
If the machine allows it, repeat the swap with the boards instead of the cells to confirm. - Repeatability with a test weight. After any repair, place a certified test weight in the hopper, record the reading, remove it, and confirm the channel returns to the same zero. Repeat five times. Non-returning zero = deformed beam or residual mechanical binding, not a calibration problem.
Corrective Actions
| Finding | Action | Post-repair verification |
|---|---|---|
| Loose fastener in load path | Clean thread, apply the OEM-specified threadlocker if the manual calls for it, torque to spec | Tap test + 5× test-weight repeatability |
| Cell fouling / contact | Remove obstruction, restore clearance, reroute or re-clip the cable clear of the moving end | Feeler check on all gaps, then re-zero and monitor 10 min |
| Product build-up in the gap | Clean per OEM wash-down instructions; review cleaning frequency in the SOP | Zero stability before and after the next wash-down |
| Deformed / overloaded cell | Replace the load cell; recalibrate that channel with certified weights | Span check at low, mid and full hopper load |
| Water ingress | Replace cell or connector; verify IP rating of the replacement matches the original and that seals/gaskets are refitted | Insulation resistance to cell body; stability after next wash-down cycle |
| Damaged module electronics board | Replace the board with the OEM part; re-enter or re-run the channel calibration as the OEM procedure requires | Raw-count noise comparison against known-good channels |
| Harness / connector fault | Replace harness; keep signal cable shield terminated exactly as the OEM wiring diagram specifies | Cable flex test with live reading |
| External vibration | Isolate the source: check feeder amplitude settings, machine levelling and mounting feet, and nearby equipment | Raw-count noise with the suspect neighbouring machine stopped vs running |
Verification and Return to Service
- Zero the repaired channel and monitor raw counts for at least 10 minutes with the machine idle. Peak-to-peak noise should match the known-good reference channels recorded in step 1.
- Run the machine empty at production speed with feeders active. Drift that only appears with feeders running points back to vibration coupling or a marginal mechanical contact.
- Perform the OEM calibration for the channel using certified test weights at low, mid and full scale. Record the values.
- Run a production check batch and compare the repaired channel's contribution rate against the other heads. A channel that is systematically selected less often than its peers is still reading high or low.
- Verify final pack weights on an independent checkweigher or reference scale before releasing the line.
- Where the weigher is used for legal-for-trade or declared-quantity packing, complete the metrological re-verification your site procedure and local weights-and-measures regime require after a load cell or weighing-electronics replacement.
Escalation and Spares Strategy
A multihead weigher is production-critical and its weigh modules are usually matched, calibrated assemblies. If the mechanical checks come back clean and the swap test points to the electronics, engage the OEM's technical support before substituting non-original parts: channel calibration constants, module addressing and firmware compatibility are commonly tied to the specific board and cell pairing, and a mismatched replacement can produce a channel that calibrates but does not track linearly.
Practical spares set for a line running this machine type:
- One spare load cell of the exact OEM part number and capacity — capacity is per-hopper, not per-machine, and a higher-capacity cell will degrade resolution.
- One spare module electronics board, kept with its firmware revision noted.
- One spare load cell harness and connector set.
- A certified test weight set covering low, mid and full hopper load.
Log every drift event with channel number, date, raw counts and the corrective action taken. Recurrence on the same physical position — even after the cell and board have both been replaced — indicates a frame, mounting or vibration problem rather than a component problem.
FAQ
Why does my multihead weigher hopper weight drift right after zeroing?
Because the load path is unstable, not because calibration has shifted. The most common causes are the load cell free end touching the housing, cover or cable, a loose screw at the cell or hanger mounting, or product residue bridging the load cell gap. Verify clearance and fastener torque before touching any calibration setting.
How do I tell whether the fault is the load cell or the module electronics board?
Swap the suspect load cell with a known-good cell from another module and leave the boards in place. If the fault stays at the same module position, the board or its harness is at fault; if the fault moves with the cell, replace the cell.
Can I just recalibrate the unstable channel instead of repairing it?
No. Calibration maps a stable A/D reading to a weight; it cannot compensate for a load path that changes with vibration or temperature. Recalibrating an unstable channel produces a value that looks correct at the moment of calibration and drifts again within minutes or hours.
What should I check first after a wash-down if one head starts fluctuating?
Check for water ingress at the load cell body and connector, and inspect the cell gap for trapped residue or a displaced seal. Measure insulation resistance from the bridge conductors to the cell body and compare against the load cell datasheet; a low value confirms moisture in the cell.
Should auto-zero tracking be enabled on a multihead weigher?
Only within the OEM's intended window and never as a workaround for a hardware fault. Aggressive auto-zero absorbs product residue left in the hopper and slowly developing mechanical faults, which shifts real pack weight without raising an alarm.
