How to Test a Load Cell: Zero Balance, Resistance and Insulation

Reviewed by MNM Scales technicians · Updated October 27, 2026

A failed load cell is one of the three most common problems our technicians see, along with settings changed by mistake and cut cables. Load cells wear out after years of heavy use, and an overload or a shock load, such as a pallet dropped on a floor scale, can bend the metal inside. MNM Scales has tested and replaced load cells for over 30 years, since 1992, on scales and weighing systems used by aerospace, automotive and steel companies, food and chemical plants, and many others. This guide shows the same tests our technicians do, with the values to look for.

Load cell troubleshooting: symptoms and likely causes

A load cell that is not working rarely just goes dark. It shows up as a symptom on the indicator. Match what you see to the likely cause, then run the test that confirms it.

Symptom	Likely cause	Check first
Reading stays at zero or does not change under load	Open signal or excitation wire, cut cable, loose connector or a failed cell	Connectors and cable, then Test 3 (resistance)
Reading drifts or creeps	Moisture in the cell or cable, temperature change, something binding the platform	Test 4 (insulation) and the cable entry seals
Will not return to zero	Overload or shock load (permanent set), debris under the platform	Test 1 (inspection) and Test 2 (zero balance)
Jumping or unstable reading	Damaged cable, water in the junction box, poor grounding or vibration	Cable and junction box, then Test 4
Corners read differently	One damaged cell or a junction box out of trim	Corner test and the multi-cell steps below
Reading goes negative under load	Signal wires swapped, or the cell is loaded in the wrong direction	Wiring against the datasheet and the mounting
Indicator shows an error code with nothing on the scale	Signal outside the indicator’s range: bad cell, broken wire or short	Tests 2 and 3; see the error codes section below

Before you start: tools and safety

• A digital multimeter that reads millivolts (mV) and ohms (Ω).
• A megohmmeter (insulation tester) set to 50 V DC or less. Higher test voltages can damage the strain gauges.
• The load cell datasheet or certificate: rated output (mV/V), input and output resistance, zero balance and excitation.
• A known test weight, for the load test.
• Labels and a photo of the wiring before you disconnect anything.

Turn the indicator off and disconnect the load cell at the indicator or junction box. Test each load cell on its own: on a multi-cell scale, a reading taken through the junction box mixes all the cells together.

Know the wires

A strain gauge load cell is a Wheatstone bridge with two excitation wires (power in), two signal wires (output) and usually a shield. Six-wire cells add two sense wires. Colors vary by manufacturer, so always check the datasheet. A common scheme is shown below.

Wire (common scheme)	Function
Red	+ Excitation (power in)
Black	− Excitation
Green	+ Signal (output)
White	− Signal
Yellow or bare	Shield (ground), not part of the bridge
Blue and brown (6-wire)	+ Sense and − Sense, at the indicator

Test 1: visual and mechanical inspection

Many failures can be seen. Look for:

• A bent, twisted or cracked body, or a mounting that no longer sits flat.
• Damaged, crushed or cut cable, especially where it leaves the load cell.
• Corrosion, swelling or water around the cable entry or the gauge area.
• Weld spatter or arc marks: welding near a scale without protecting the load cells can damage them.
• Loose mounting bolts, missing hardware, or debris binding the cell.

Test 2: zero balance (no-load output)

Zero balance shows whether the load cell has been overloaded and permanently bent. Power the cell with its normal excitation, usually from the indicator, with no load on it, and measure the mV between the two signal wires.

1. Connect only the excitation wires to the indicator or a stable supply, and note the exact excitation voltage.
2. With no load on the cell, measure the mV between + signal and − signal.
3. Work out the full-scale output: rated output (mV/V) × excitation (V). A 3 mV/V cell at 5 V gives 15 mV at full capacity.
4. Compare the no-load reading with the datasheet zero balance, usually ±1 to ±2 percent of rated output. For the example above, ±1 percent is ±0.15 mV.

A reading well outside the zero balance spec, or one that keeps creeping, usually means the cell was overloaded or shock-loaded and has taken a permanent set.

Test 3: input and output resistance (bridge integrity)

Disconnect the load cell completely and measure resistance with the multimeter:

Measure between	What you are checking	Healthy result
+ Excitation and − Excitation	Input resistance	Matches the datasheet (often about 350 to 400 Ω for a 350 Ω cell)
+ Signal and − Signal	Output resistance	Matches the datasheet (often about 350 Ω, within a few ohms)
Each excitation wire to each signal wire	Bridge symmetry	Pairs read close to each other

An open circuit (no reading), a short (near 0 Ω), or values well off the datasheet mean a damaged gauge or wire. Uneven pairs in the symmetry check point to a damaged bridge arm, often from overload or moisture.

Test 4: insulation resistance

Moisture is a quiet load cell killer: it causes drift long before the cell fails completely. With the cell disconnected, tie all the bridge wires together (excitation and signal), and use the megohmmeter at 50 V DC or less to measure between them and the load cell body, and between them and the shield.

Reading (bridge to body or shield)	What it means
5,000 MΩ (5 GΩ) or more	Normal for a healthy, dry load cell
About 2,000 to 5,000 MΩ	Watch it; retest, and check the cable and seals
Below about 2,000 MΩ	Suspect moisture or contamination; drift is likely
Near zero	Short to the body; replace the load cell

Never use a high-voltage insulation tester on a load cell or connect one with the indicator still attached.

Test 5: response to load

Reconnect the cell to excitation and watch the signal while you apply a known weight. The output should rise smoothly in proportion to the load, hold steady, and return to the same zero when the load is removed. Jumping, creeping or a different zero after unloading all point to a damaged cell or cable. On a complete scale, the corner test does the same job: the same weight over each load cell should give the same reading.

Finding the bad cell on a multi-cell scale

Floor, tank and hopper scales usually have four load cells summed in a junction box, so the first job is to find which cell is at fault. Our technicians work through it like this:

1. Do a corner test with the same weight over each load cell and note any corner that reads high, low or unsteady.
2. Open the junction box, check for water and corrosion, and photograph the wiring before touching it.
3. Disconnect one load cell at a time and run the zero balance, resistance and insulation tests on that cell alone.
4. Compare the four cells with each other as well as with the datasheet: a cell that differs from its three neighbours is the likely culprit.
5. Check the homerun cable from the junction box to the indicator the same way, since a damaged cable can look like a failed cell.

Write the readings down. They make the replacement quote faster, and they give you a baseline for the next inspection.

Load cell error codes on the indicator

A load cell has no display, so it cannot show an error code itself. The indicator or scale head reads the load cell signal and shows a code when that signal is out of range. These codes, from the manufacturers’ manuals, often point to the load cell:

Code (scale)	What the manual says	What it can mean for the load cell
Err30 / Err31 (Brecknell B140)	A/D input over the maximum / below the minimum range	Signal out of range: a damaged cell, open wire or short. Run Tests 2 and 3.
Err04 / Err05 (Brecknell B140)	The zero point is too high / too low	Something resting on or binding the cell, or a cell with a permanent set.
Err 8.1, 8.2, 8.4 (OHAUS Scout)	Zero out of range, underload (pan off)	Usually the pan; if it returns with the pan seated, the cell may be damaged from a drop.
Err 1 (CAS ED)	Zero point shifted since the last span calibration	Recalibrate; if it keeps shifting, test the cell.
-E (A&D EK-i)	Weight sensor pushed upward	Check for anything trapped around the pan; the sensor may be faulty.
-OL- (Brecknell CS crane)	Overload	Repeated overloads can bend the cell: check zero balance.

Some indicators also show raw A/D counts in a service menu. The Brecknell B140 manual, for example, expects about 5,000 to 100,000 counts with the platter empty and more than 300,000 at full capacity; counts outside those ranges, when calibration cannot be done, mean the load cell has been permanently damaged.

For the full code lists by brand, see scale error codes explained.

Test results at a glance

Result	Likely cause	Next step
Zero balance out of spec	Overload or shock load; permanent set	Replace the load cell
Open or shorted bridge	Broken gauge or wire, cut cable	Replace the load cell (cable is part of it)
Low insulation resistance	Moisture in cell or cable	Dry and reseal if mild; otherwise replace
All tests pass, scale still wrong	Junction box, indicator or calibration	Check the junction box and calibrate (see how to calibrate a floor scale)

What is a Class 1 (C1) load cell?

Load cells are graded by accuracy class. Under the international standard OIML R60, the classes are A, B, C and D, and the number after the letter is the maximum number of verification intervals the cell supports, in thousands. A C1 load cell is rated for up to 1,000 intervals, a C3 for 3,000 and a C6 for 6,000. “Class 1” usually means C1.

Rating	Where you see it	What it tells you
C1 (OIML)	Datasheets of general-purpose cells	Up to 1,000 intervals: process, batching and check weighing where fine resolution is not needed
C3 (OIML)	Datasheets and OIML certificates	Up to 3,000 intervals: widely used in bench, floor and platform scales
C6 and higher (OIML)	High-resolution cells	Up to 6,000 or more intervals
Class III or IIIL, nmax, S or M (NTEP)	NTEP Certificate of Conformance in the US	The Handbook 44 accuracy class, the maximum number of divisions, and whether the cell is approved for single-cell (S) or multiple-cell (M) use

For a scale used in trade in the US, look for the NTEP certificate rather than the OIML class alone (see what is NTEP and legal for trade). When you replace a cell, match the class and nmax of the original. Ask us for the certificate details on any load cell we quote.

Replacing a load cell

If a load cell fails, replace it with the same type, capacity, output and approval. Our load cell types guide explains the shapes, and the load cell buying guide covers sizing. Common replacements we quote include the MNM Scales ZEMIC H8C 2,500 lb single-ended beam, NTEP, the MNM Scales GX-1-SSH-5K stainless steel washdown shear beam, 5,000 lb, NTEP, the GX-1-20K large envelope shear beam, 20,000 lb, the MNM Scales 3,000 lb alloy steel S-type load cell, the MNM Scales MS-1-SSH 5,000 lb stainless steel S-type, NTEP, the MNM Scales LC-4204-K300 300 kg single point load cell and the MNM Scales IDS-672-25K-CT pancake load cell, 25,000 lb. Mounting hardware is not included with bare load cells unless the listing says so. Replacing a whole set on an older floor scale? A new MNM Scales stainless steel junction box for 4 load cells is worth adding.

Do not cut or shorten a load cell cable to fit: on a 4-wire cell, the cable length is part of the cell’s temperature compensation. Coil the extra length instead.

After any load cell change, the scale must be calibrated. Our technicians replace load cells in our Cypress, TX shop and on-site, and bring their own certified test weights to calibrate floor and crane scales. On-site service in the Houston area is $295 plus $2 per mile, round trip, charged once per visit, with calibration priced per scale (see industrial scale calibration and scale repair). Testing a force gauge or compression setup instead? Shop compression scales or see compression scale vs force gauge vs load cell.

Why buy from MNM Scales

• Load cells, junction boxes, cables and indicators from one company, quoted by phone, with technicians who test and install them
• An authorized dealer for the brands we sell, in business since 1992 (over 30 years)
• In-house calibration with a certificate, or NIST calibration with a NIST-traceable certificate; mail-in calibration also covers scales bought elsewhere and is usually done within 1–3 business days after we receive the scale
• TDA-registered technicians; on-site service in the Houston area is $295 plus $2 per mile, round trip, charged once per visit, with calibration priced per scale
• Scale rentals of all types of scales
• Order pickup at our shop, 21249 FM 529 Rd, Cypress, TX, by appointment
• Phone answered 24/7 at (832) 290-3120

Frequently asked questions

How do I test a load cell with a multimeter?

Disconnect it, then measure input resistance across the excitation wires and output resistance across the signal wires and compare them with the datasheet. With excitation applied and no load, measure the mV output and compare it with the zero balance spec.

What should a load cell read with no load?

Close to 0 mV, within the datasheet zero balance, usually plus or minus 1 to 2 percent of rated output. For a 3 mV/V cell at 5 V excitation, 1 percent is 0.15 mV.

What is a normal load cell resistance?

It depends on the cell, so use the datasheet. Many cells read about 350 to 400 ohms across the excitation wires and about 350 ohms across the signal wires.

How do I check a load cell for moisture?

Measure insulation resistance between all bridge wires tied together and the body or shield, using 50 V DC or less. A healthy cell reads 5,000 megohms or more; below about 2,000 megohms points to moisture.

What causes a load cell to fail?

The most common causes are overload or shock loads, years of heavy use, moisture getting into the cell or cable, cut or crushed cables, and welding near the scale.

Can a damaged load cell be repaired?

Load cells are sealed, calibrated parts, so a failed cell is normally replaced with one of the same type, capacity, output and approval, and the scale is recalibrated.

Can I shorten a load cell cable?

No. On a 4-wire load cell, the cable length is part of the temperature compensation, so coil the extra cable instead of cutting it.

Do I need to calibrate after replacing a load cell?

Yes. Any load cell replacement changes the scale’s response, so the scale must be calibrated with certified test weights afterwards.

Why is my load cell not working?

The usual causes are a cut or damaged cable, a loose connector, moisture in the cell or junction box, or an overload that bent the cell. Check the cable and connections first, then run the resistance, zero balance and insulation tests to confirm.

Do load cells show error codes?

No. The indicator or scale head shows the code when the load cell signal is out of range, such as Err30 or Err31 on a Brecknell B140 or Err 8.4 on an OHAUS Scout. Test the cell to confirm whether it is the cause.

What is a Class 1 load cell?

Usually an OIML class C1 load cell, rated for up to 1,000 verification intervals. C3 cells support 3,000 intervals. In the US, trade scales use cells with an NTEP certificate that lists class III or IIIL, nmax and single or multiple-cell use.

How do I troubleshoot a load cell?

Start with the symptom: no change under load, drift, no return to zero or unstable readings. Inspect the cell and cable, then test zero balance, input and output resistance and insulation, and compare the cells on a multi-cell scale.