2. Channel mapping in multi-load-cell systems

A single load cell is electrically simple: four wires, one bridge, one signal. A platform scale or weigh module with four, six, or eight load cells is a different problem. Each cell produces its own millivolt-level signal, and the indicator can only accept one. Between the cells and the instrument sits a junction box whose internal wiring determines how those signals are combined and, just as importantly, which physical corner of the scale corresponds to which channel. Getting that mapping wrong produces errors that are easy to misdiagnose: a scale that reads correctly at the center but drifts when a load moves to one corner, or a weigh module that reports a plausible but consistently offset weight because two cells were summed with opposing polarity.

This section covers the three things that must be correct before any calibration is attempted: the summing network inside the junction box, the assignment of cells to channels and corners, and the polarity convention that ties them together.

The summing network: parallel voltage summation

Most industrial junction boxes combine load cells by connecting their signal outputs in parallel across a common excitation supply. Because each cell is a low-impedance bridge (typically 350 Ω or 1000 Ω), paralleling the outputs produces a voltage that is the average of the individual cell outputs, weighted by their source impedances. With matched cells, the result is a clean average:

V_out = (V1 + V2 + V3 + V4) / 4        (matched impedances)
V_out = Σ(Vi / Zi) / Σ(1 / Zi)          (general case)

The averaging behavior is the reason a four-cell platform reads the same total weight regardless of where the load sits — provided the cells are matched and each contributes equally. It is also the reason a single miswired cell does not simply disappear: it pulls the average in the wrong direction, and the resulting error scales with the fraction of load that cell carries.

Excitation is distributed in parallel to all cells from the same supply pair. This matters for mapping because a cell with reversed excitation will produce an inverted output, which is indistinguishable at the summing node from a cell with reversed signal polarity. Both faults look like "one corner reads negative."

Junction box wiring: the physical layer of the mapping

A typical stainless-steel junction box for a four-cell platform contains a terminal block with positions for each cell's excitation (+E, −E), signal (+S, −S), and shield, plus a common bus that feeds the home-run cable to the indicator. Cells connect through individual cable glands, usually labeled 1 through 4 or J1 through J4. The mapping problem begins here: the label on the gland, the terminal position, and the physical corner of the platform are three separate facts that must be reconciled.

Three wiring conventions are common and must be identified before work begins:

• Home-run per cell: each cell has its own cable to the box; the box performs the summation. Mapping is explicit and easy to verify.
• Daisy-chain: cells are wired in sequence, with the summing occurring across intermediate terminals. A swapped pair mid-chain affects two channels at once.
• Integrated weigh module: the cells and summing board are potted or sealed inside the module, and only a single home-run cable exits. Mapping is fixed at manufacture and must be obtained from the datasheet or nameplate, not inferred.

For the first two conventions, the practical rule is to trace continuity from each cell's connector to the terminal block before applying excitation. A continuity check with the system powered down confirms which terminal pair belongs to which cell and catches the common error of a cable landed one position off.

Channel assignment principles

Channel assignment is the deliberate decision about which cell occupies which channel and which channel corresponds to which physical corner or section of the scale. Four principles govern it:

1. Geographic consistency: channel numbers should follow a fixed spatial convention — for example, 1 = front-left, 2 = front-right, 3 = rear-right, 4 = rear-left, viewed from the operator position. The convention must be written down and applied identically on every scale of the same design.
2. Polarity consistency: all cells must be wired so that a downward load produces a positive signal contribution at the summing node. Mixed polarity within a network cancels signal and produces a scale that under-reads by an amount proportional to the affected cell's share.
3. Traceability: every channel must be recorded against the cell serial number, cable tag, and physical location. This record is what makes later corner diagnostics possible.
4. Symmetry for matched groups: where cells are supplied in matched sets, keep the set together on one scale rather than mixing cells across scales.

The following figure shows a four-cell platform with its junction box, the parallel summing bus, and the channel-to-corner assignment recorded on the terminal block.

Worked example: verifying a four-cell platform

Consider a 1,000 kg platform with four 500 kg cells, a 2 mV/V sensitivity, and a 5 V excitation. Each cell at full rated load produces 10 mV. With the platform unloaded, the summed output should be near zero after the indicator's zero adjustment. The verification sequence below is illustrative; the numbers are representative values for a matched set, not measured results.

1. Record the mapping. Note each cell's serial number against its channel and corner. Confirm the gland labels match the terminal labels.
2. Check polarity individually. With the system powered and the platform empty, apply a known test load (for example, a 100 kg test weight) directly over each corner in turn. The indicator should move in the positive direction for every corner. A negative or near-zero response at one corner indicates reversed signal or excitation on that cell.
3. Check summation symmetry. Place the same 100 kg weight at the center and at each corner. For a matched set, the four corner readings should agree within the platform's corner-load tolerance — commonly a few tenths of a percent of the applied load. A corner that reads low by roughly 25% of the expected share suggests that cell is not contributing, which points to an open signal line or a reversed connection.
4. Confirm the average. With the weight at center, the indicated value should equal the applied load within the indicator's resolution. If the center reading is correct but corner readings diverge, the summing network is working but the cells are mismatched or mechanically misaligned.

Weigh modules and fixed mappings

Integrated weigh modules — common in tank, hopper, and conveyor weighing — often ship with the summing board sealed inside the module housing and a single cable exiting. In these systems the channel mapping is a manufacturing fact, not a field decision. The practical consequence is that the mapping must be obtained from the module datasheet or nameplate and recorded in the maintenance file. When a module is replaced, the replacement must carry the same internal mapping, or the indicator's corner-adjustment values (if any) will no longer correspond to the physical corners.

For multi-module systems — a large tank on three or four weigh modules — the same principles apply at the module level: each module is a channel, each channel maps to a physical support point, and the polarity convention must be consistent across modules. A tank with one module wired in reverse will read low by that module's share of the load, and the error will change as the tank's center of gravity shifts with fill level.

Mapping record: the deliverable that makes diagnosis possible

The single most useful artifact from this work is a mapping record that ties channel, terminal, cable tag, cell serial number, and physical location together. A minimal record for the example platform above looks like this:

With this record in hand, a future corner-load fault can be localized in minutes: the technician knows which terminal to probe, which cable to trace, and which physical corner to load. Without it, the same fault becomes a process of elimination across the entire network. The mapping record is not paperwork — it is the diagnostic interface to the summing network.

The next step after mapping is confirming that the summed signal behaves as the network predicts under known loads, which is the subject of the calibration and verification work that follows.