C3 load cells, rated for 3,000 verification intervals under OIML R60, cannot legally or metrologically be installed in a scale approved for 6,000 divisions (Class III or IIIL with 6,000d); the scale's approval envelope is defined by the load cell class, and C3 falls short by a factor of two.
An OIML R60 accuracy class encodes the maximum number of verification intervals (n_max) the cell can sustain while meeting combined error, repeatability, creep, and temperature-drift tolerances, and the C-marking (C1 through C6) is the contractual ceiling, not a recommendation [S1][S2]. A scale that displays 6,000 divisions requires its load cells to be marked C6 or higher; a C3 cell physically functions in the rig, but the system fails conformity assessment.
OIML R60 Accuracy Classes and the Division-Count Contract
OIML R60 defines load cell accuracy classes A, B, C, and D, with class C subdivided by the number of verification intervals: C1, C2, C3, C4, C5, and C6, where the digit following the letter is the n_max divided by 1,000 [S2]. A C3 cell therefore has an n_max of 3,000, a C4 cell 4,000, and a C6 cell 6,000; the marking is the binding ceiling on the verified performance envelope the certification laboratory actually tested [S2].
The C-subclass also carries a defined combined error band as a fraction of rated output: C3 sits at 0.0230%, C4-C5 at 0.0174%–0.0140%, and C6 at 0.0116% [S4]. That error budget is the envelope inside which non-linearity, hysteresis, temperature effect on sensitivity, and creep over 30 minutes must all fit simultaneously, and the laboratory test sequence that produced the certificate was run at 3,000 intervals, not 6,000.
What Happens When C3 Cells Drive a 6000d Indicator
A scale approved for 6,000 divisions expects its load cells to support that interval count across the full weighing range, including corner-load, creep, and warm-up drift, and substituting C3 cells voids the conformity assumption that the indicator relies on for legal-for-trade output. The error envelope at 6,000d exceeds what the C3 cell was verified for, so the displayed weight is not traceable to the original R60 certificate. [S2]
Thames Side's T66 bending-beam cell illustrates the exact upgrade path: the same mechanical body is offered in C3 (3,000 divisions) and C6 (6,000 divisions) variants, with temperature coverage from -30 to +70 degrees C identical between versions, proving that going from 3,000 to 6,000 divisions is a sensor-class change, not a wiring or calibration tweak [S6]. Selecting the C3 variant on a 6,000d indicator configuration would replicate the fit but invalidate the metrology file.
Where C3 Is Actually the Right Choice
C3 is the workhorse class for non-trade industrial scales, including belt scales, conveyor scales, tank and hopper scales with 3 or 4 load cells, and platform scales up to roughly 3,000 divisions, and it remains the most common accuracy class shipped for general machine-construction weighing [S3][S4]. For a 1,500d or 3,000d process scale, C3 is correctly specified, and over-specifying to C4 or C6 simply increases unit cost with no metrological benefit.
The C3 class also covers most legal-for-trade retail and process applications that do not exceed 3,000 divisions, including bench scales and portable platforms where the resolution requirement is modest. A load cell module built around C3 sensors is the standard build for batching, filling, and inventory-by-weight operations that do not require the tighter 0.0116% error band of C6.
C3 vs C4 vs C6 Compared on the Four Decision Criteria
The decision between C3, C4, and C6 reduces to four operating criteria: required division count, combined error budget, temperature span, and cost. C3 supports up to 3,000d at 0.0230% combined error, C4-C5 supports 4,000-5,000d at 0.0174%–0.0140%, and C6 supports 6,000d at 0.0116% [S4]. On cost, each step up the class ladder typically adds a substantial multiplier for the same mechanical body, because tighter error bands require more extensive gauge matching, longer stabilisation, and individual calibration records.
For a 6,000d application, the comparison is binary: only C6 satisfies the metrological floor, and C3 or C4 are not substitutes regardless of price advantage. For a 3,000d or lower process scale, the comparison favours C3, because C4 or C6 buys tighter error than the scale's display resolution can resolve, and the higher-class cell becomes stranded precision.
Selection Criteria: Reading the Indicator, the Cell, and the Application
The correct class is dictated by the highest-resolution division count the indicator and the scale's verification certificate will declare, not by the load cell's rated capacity. Selection should begin at the system level: identify the number of verification intervals (n) the scale will be declared at, then specify load cells whose OIML class n_max equals or exceeds n, with margin for multi-cell loading configurations that reduce the effective interval count per cell. [S2]
For a four-load cell tank scale with a gross capacity split equally across the four supports, each cell effectively sees one quarter of the load, but the OIML class still applies per cell, and the system's verification interval count cannot exceed the cell's n_max regardless of how the load is shared. Bench scale builds with a single point cell follow the same rule: the cell's n_max is the ceiling on the scale's legal-for-trade division count.
Limits, Failure Modes, and Common Misreadings
Two common misreadings drive C3-on-6,000d mistakes. First, the C-marking is read as a 'minimum' rather than a verified ceiling, leading to the assumption that a C3 cell can simply be asked to do 6,000d of work. The OIML R60 certificate covers only the tested n_max, and the laboratory did not characterise the cell at 6,000 intervals, so there is no metrological basis to claim 6,000d performance [S2].
Second, capacity headroom is confused with accuracy headroom. A 1,000 kg C3 cell used at 200 kg dead load has 5:1 utilisation headroom for mechanical safety, but no accuracy headroom for 6,000d display; the cell can carry the load without overstressing the strain gauges, but the error band at 6,000d is unverified and likely violated. Environmental drift, corner-load asymmetry, and creep over 30 minutes all eat into the C3 envelope faster than into the C6 envelope, which is exactly why the C6 cell exists as a separate certified variant [S6].
Standards and Sourcing Trail
OIML R60 (Recommendation 60-1) governs the load cell certification scheme and defines the C-class taxonomy used across Europe, Asia, and most export markets [S2]. For US legal-for-trade applications, NIST Handbook 44 governs the scale level, and NTEP Class IIIL covers single-cell and multi-cell load cells for larger-capacity scales with 2,000 to 10,000 scale divisions; NTEP Class III corresponds to typical 3,000d retail and process scales [S7].
The selection path stays clean when the metrology trail is explicit: pull the scale's verification certificate, read the declared n_max, then require the load cell data sheet to show an OIML R60 class with n_max equal to or greater than that number. If the application demands 6,000 divisions, specify a C6 (OIML R60) or NTEP IIIL 10,000d cell; C3 is rejected at the datasheet stage, before any hardware is on the bench.
Trackable next signals: confirm the indicator's maximum displayable divisions against the OIML class of the installed cells at commissioning, and require the cell manufacturer to issue an R60 test certificate matching the serial number of every cell delivered; a certificate of conformance without a serial trace is not a metrology document.
This topic is covered further in Concrete vibrator head motion: rotation vs oscillation.