Load cell junction box wiring for five parallel cells

Wire all five load cells in parallel — every EXC+ joined to every other EXC+, every EXC- to EXC-, SIG+ to SIG+ and SIG- to SIG- — across two junction boxes cross-connected by one six-core cable. Five 350 Ω cells then read about 70 Ω between the excitation rails and draw about 71 mA at 5 V, and your multimeter confirms both numbers before the weighing module is ever powered. The second box adds terminals, not a second weighing channel: all five bridges stay one network, and the module sees one averaged signal, the arrangement compared in combined versus individual load cell readings. Here is the split, the terminal schedule, the cross-connection and the checks that prove it.

What do you need to wire five cells into two junction boxes?

Gather five cells of one model and one capacity, two boxes carrying 30 and 18 free terminals, six-core shielded cable for every run, gland inserts sized to each cable OD, and a multimeter that resolves 0.1 Ω.

The terminal count comes straight from the schedule. Box 1, nearest the module, lands 3 cells at 6 conductors each (18), the interbox cable (6) and the home run (6) — 30 terminations. Box 2 lands 2 cells (12) plus the interbox cable — 18. Four-wire cells without sense conductors use 4 cores each, which trims six terminations off both numbers.

For the box itself: on a washdown platform take a 316L enclosure rated NEMA 4X / IP66 by the manufacturer; not UL listed, and remember the assembly carries the rating of its weakest part, so fit glands rated to match the box. Lay out the gland positions in the cutout designer — one hole per cell cable in the bottom wall, one for the interbox run, one on a side or the top for the home run. If you would rather have one enclosure than two, a custom cutout pattern with five glands is delivered in 10 business days, counted from your approval of the drawing. The box and terminal choices are covered in instrument junction box selection for signal wiring.

Order the five cells in one batch. Datasheets quote output in mV/V — most commonly 2 or 3 mV/V — and cells from one order sit close enough in output that they parallel cleanly.

How do you split five load cells across two boxes?

Put three cells in the box nearer the module and two in the second box — 3 + 2 = 5 — then join the boxes with one cable so both carry the same six rails: two enclosures, one electrical network.

The split is purely physical. Shortest cable runs win: place each box where its cells reach it without coils on the floor, and take the home run from the box closer to the module trunk. Dividing cells between boxes never divides the measurement.

Electrically, five identical bridges in parallel equal one bridge with every resistor divided by five: 70 Ω in, 70 Ω out, sensitivity unchanged at 2 mV/V, and the module reads the average of the five. Cells whose outputs differ circulate current between themselves at rest, and that current drifts with temperature — the reason for one model, one capacity, one order. Leave about 20 % of your terminals spare for a sixth cell; spare-terminal planning for instrument boxes runs the same arithmetic.

How do you build the terminal schedule?

One row per function and one column per cell — EXC+, EXC-, SIG+, SIG-, SEN+, shield — so five six-wire cells fill a 6 × 5 grid that serves as both your wiring drawing and your check sheet.

Labels vary between cell makers — EXC± or E±, SIG± or S±, SEN± or FS± — so copy the wording from your cells' connection drawing onto the labels. A schedule beats a sketch because every conductor has a home and every home has a check. The full list of what such a drawing should carry is in the junction box terminal drawing checklist.

How do you wire the cells and cross-connect the boxes?

Land every cell on its scheduled terminals, join the two boxes function for function — EXC+ to EXC+, SIG+ to SIG+ — and run one six-core home run from box 1 to the module, all of it with the supply isolated.

Lock out the weighing module's supply before you land the first conductor; every check in this article is made de-energized.

1. Label both ends of every cable with cell number and function: a label of '4 SIG+' finds a fault in seconds, a color code will not.
2. Land cells 1-3 in box 1 and cells 4-5 in box 2 on their scheduled terminals.
3. Form six common rails in each box, one per function, using the spare positions you reserved.
4. Pull the six-core interbox cable and join rail to rail: EXC+ to EXC+, EXC- to EXC-, SIG+ to SIG+, SIG- to SIG-. Cores five and six are your spare pair or a future sense extension.
5. Pull the six-core home run from box 1 to the module: excitation and signal cores to the measuring inputs, SEN+ and SEN- to the sense inputs.
6. Land SEN+ and SEN- on the box 1 EXC rails — the node all five bridges actually share.
7. Carry the shields unbroken through both boxes and bond them at the module end only; enclosure grounding and bonding covers the panel side of that bond.
8. Tighten every terminal to the block's marked torque and tug each conductor. A signal core that falls out later moves your reading by about a fifth at uniform load, and the meter will show it as 88 Ω.
9. Seat the gland inserts and close the boxes: the insert makes the seal that holds the rating, not the nut.

What resistance should five parallel cells read?

With the module leads off, five identical 350 Ω cells read 350 / 5 = 70 Ω from EXC+ to EXC-, about 70 Ω from SIG+ to SIG-, and about 53 Ω from EXC+ to SIG+, because a bridge is resistive between any two of its nodes.

Worked example — five 350 Ω cells, 5 V excitation, 25 m [82 ft] home run of 20 AWG at 33 Ω/km:

• Equivalent resistance: 350 Ω / 5 = 70 Ω. Measured at the module end you add the two excitation cores, 2 × 25 m × 0.033 Ω/m = 1.7 Ω, so expect roughly 71-72 Ω.
• Excitation current: 5 V / 350 Ω = 14.3 mA per cell, and 5 × 14.3 mA = 71 mA for the bank. At 10 V — the top of the 2.5-10 V range most datasheets allow — that becomes 28.6 mA per cell and 143 mA for the bank, so check your module's drive limit before raising the voltage.
• Lead drop: 0.071 A × 1.7 Ω = 0.12 V. Without sensing, the bridges see 5.00 − 0.12 = 4.88 V and the scale reads 2.4 % low (0.12 / 5.00). With SEN± on the box 1 rails, the module lifts its output until the rails hold 5.00 V.
• The 3 m [10 ft] interbox run adds about 0.1 % more drop at two cells' 29 mA — inside the corner matching you already accept.
• Signal level: a 2 mV/V cell delivers 10 mV at 5 V and full load, and five in parallel still deliver 2 mV/V. Output impedance falls from 350 Ω to 70 Ω, so resistance in the signal pair distorts the reading five times less.

A low reading from EXC+ to SIG+ is normal — about 260 Ω on one cell, 52.5 Ω on five — so treat it as a health check, not a short.

What checks prove the wiring before power-on?

Three unpowered readings at box 1 — 70 Ω EXC+ to EXC-, 70 Ω SIG+ to SIG-, 53 Ω EXC+ to SIG+ — plus a core-to-core isolation test on the interbox cable prove the network before anything is energized.

Work in this order:

• Interbox cable, both ends free: each core continuous end to end, a few ohms at most, and no core touching another.
• Home run: the same two checks.
• Network assembled, module leads off: the three readings above, taken on the box 1 rails.
• Wiggle each cable at its gland while you watch the meter; a jumping reading is a strand or a clamp finding its seat, and you want that found now.

Keep an insulation tester's 500 V away from connected cells and modules: strain gauges are 10 V devices.

Then power up and prove it live:

• Unloaded output sits near zero, inside the zero-balance band printed on the cells' certificates; the tare closes the rest.
• Corner test: press each of the five loading points in turn, and the reading rises at all five. A corner where it falls has a reversed SIG pair or a swapped EXC pair — swap it at the box before any calibration.
• Place one known test mass over each support in turn; the five results should sit within the cells' own output spread. A wider spread points to a bound-up mounting foot, which is mechanical, not electrical.

What wiring mistakes show up later as weighing errors?

Four faults cause nearly every bad multi-cell scale — a reversed signal pair at one cell, an excitation polarity swap at one cell, the shield carrying current, and mixed cell models — and each leaves a signature the corner test or the meter finds.

• Reversed SIG+ / SIG- at one cell. Its output inverts and fights the other four; pressing its corner drops the reading. Swap the pair at the box and the corner test comes back positive.
• EXC+ / EXC- swapped at one cell. Same inverted output, so the same falling corner — and the resistance readings still look right, which is why the corner test is the one that catches it. Check polarity against the cell drawing before you re-terminate.
• Shield used as a bridge return. Readings drift and jump when nearby equipment starts, because those 71 mA change path with whatever else runs in the tray. One bond at the module end fixes it.
• Mixed models or capacities in parallel. Outputs differ, current circulates between bridges at rest, and zero walks with temperature. A straight parallel box is for matched cells; summing boxes with trimming resistors exist for the mixed case.

The decision in one line: matched cells, one network, 3 + 2 across two boxes, cross-joined function for function, sense at the box 1 rails, and pass on 70 / 70 / 53 Ω before you power the module. For everything about the enclosure itself — stiffness, mounting and temperature — carry the rules in load cell enclosure design without weighing errors, and the wiring above will hold its calibration.

Enclosures for this job

Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.

Questions