The short answer is that all three are valid, they just prove different things, and the right one depends on how accurate the number has to be and how the scale is installed. Test weights (also called billets) are quick and repeatable and are the workhorse of routine calibration. A calibration chain sits a step closer to a real belt load. A material test proves the whole system end to end on your own product. We manufacture our own belt weighers and also service and calibrate all the major brands, so what follows is how we actually pick a method on site, not a preference for one over the others.
A belt weigher (also called a weightometer, belt scale or conveyor belt scale) has to be set against a known reference so its span reads true, and these three methods are those references. For the broader picture of a calibration day, see what happens during a calibration visit.
What is each calibration method?
Static test weights (billets). Known masses applied directly to the weigh frame. There are two common arrangements, and they matter when you buy a scale. Plate weights are mounted by hand at each calibration. Store-in-place billet weights live on the scale permanently and are applied with a built-in lifting mechanism, so nobody carries anything. Either way, the weight hangs on the frame and acts through its levers, so the load the scale actually sees is the effective test weight, not the stamped mass. That geometry is the whole point of the next section.
Calibration chain. A length of chain of known mass per metre, laid across the weigh span so it loads the belt the way spread-out material does. It rolls along with the belt rather than sitting on one point, which makes it a closer imitation of a distributed load than a static weight hung on the frame. It is a middle path: more like real material than a billet, less effort than a full material run.
Material test. Run a known quantity of product across the scale, then weigh that same material independently, usually over a verified weighbridge or into a weigh hopper, and compare what the scale counted against what the material truly weighed. Nothing is built into the scale for this one: the material is the reference.
What does each method actually prove?
This is where honesty matters, because the three methods are not equal, and treating them as equal is how a scale reads fine on the day and wrong in production.
Static test weights prove the instrument responds correctly to a known force applied through the weigh frame. That is genuinely useful: it catches a drifted span, a dead load cell, a binding frame, most of what goes wrong. But a billet loads the frame through its own geometry, so it is an indicator of how the scale responds, not proof that the scale reads your material right at your flow. On a well-located, mechanically sound scale, that indicator is a very good one and we calibrate against it every day.
A calibration chain gets closer, because it lays a distributed load along the span and rolls with the belt, so it exercises more of what real material does. It still is not your product, but it is a better stand-in than a point weight.
A material test proves the most, because it tests the whole chain at once, belt, frame, load cells, speed sensor and integrator, on real product at a real rate against an independent scale. It is the truest answer and the most effort, which is exactly the trade-off. This is the same ladder we describe in the belt cut test, where you weigh material off a stopped belt instead of over a weighbridge.
The honest nuance to carry away: on a poorly located scale, or one that has been worked hard, static weights alone can mislead you. If the belt lifts off the bed rollers, the structure vibrates, or the weigh area sits too near a curve or a loading point, the billet may say "in spec" while the material says otherwise, because the billet never sees those effects. On scales like that, the only truthful calibration is a material run, and test weights should be used as an indicator only.
When is each method the right choice?
- Routine service on a sound, well-located scale: static test weights. They are fast, repeatable and honest for the job, which is why they carry most of our calibration work.
- Commissioning a new scale: static weights to set it up, then a material test to confirm the baseline against real product before anyone trusts the tonnes. A brand-new install has never seen its own material, so the first numbers should be proven, not assumed.
- Settling a dispute or a stock reconciliation gap: a material test, or a belt cut. When someone is arguing about the tonnes, an independent reference on real product is the only thing that ends the argument. A billet check will not.
- A poorly located or hard-worked scale: a material test is the truer check, with static weights kept as a between-times indicator. See the installation factors in choosing a belt weigher for why location caps accuracy.
- A quick health check between full calibrations: static weights, or an electronic check (below). Enough to catch a gross fault early, not a substitute for the real thing.
Store-in-place or manual plates: which should a buyer choose?
If you are specifying a new scale, this choice shapes every calibration visit for the life of the machine, so it is worth a minute now.
Plate weights, mounted by hand, are the cheapest up front. The trade-off is that someone lifts and mounts weights at every single calibration, a slower manual-handling task each time, and on an awkwardly placed scale that is a real safety and time cost.
Store-in-place billet weights with a lifting mechanism cost more up front, because the weights and their gear are built into the scale. They pay that back in faster, safer routine calibrations, since the load is applied without anyone carrying anything, which usually justifies itself on a scale calibrated often or sitting in a tight or high spot. A calibration roller chain is the third build option and suits some layouts and duties as a middle ground.
Our rule of thumb: the more often a scale is calibrated and the harder it is to reach, the more the store-in-place option is worth. A light-duty scale that is easy to get to may not need it. This is the same guidance as the calibration-method section in choosing a belt weigher, just in more detail.
What about the electronic (R-Cal style) check?
Most integrators have an internal electronic reference that injects a known signal to check the electronics. It confirms the integrator, cabling and load cell circuit are reading and scaling a signal correctly, and it takes minutes. But be clear about its limit: it proves nothing about the mechanics. It cannot see build-up on the frame, a worn roller, a bind, a lifting belt or a slipping speed sensor, because it never applies a real load. Treat it as an electronics check, never as a stand-in for weights, a chain or material.
What do we actually do on a routine visit?
On a normal maintenance calibration of a sound scale, we set span with static test weights and back that with our own judgement. We inspect first, because a weight hung on a dirty or bound frame just calibrates the fault in. If the scale is well located and the mechanics are clean and free, the test weights give us a calibration we trust, recorded as as-found and as-left figures on the report.
We escalate when the scale or the site tells us to. A history of drift, a stock gap the site cannot explain, a poor location, a belt that lifts, a commissioning baseline: any of those and we push up the ladder to a material comparison, because that is the only way to prove the number on real product. We would rather tell you a billet-only calibration is an indicator on a difficult scale than let a tidy report imply more certainty than the install can deliver.
Frequently asked questions
Are test weights good enough on their own?
On a well-located, mechanically sound scale, yes, for routine calibration. On a poorly located or hard-worked scale they are an indicator only, and a material test is the truer check. The install decides which situation you are in.
Is a calibration chain better than test weights?
It is a closer imitation of a real distributed belt load, because it spreads along the span and rolls with the belt. Test weights are quicker and more repeatable. Which is "better" depends on the scale and the duty.
Does a material test replace routine calibration?
No. A material test proves the whole system end to end but needs product, a trusted independent scale and time, so it is not a monthly job. Static weights do the regular work, with a material test at commissioning or when the numbers or the site warrant it.
Is any of this a trade verification?
No. These are practical maintenance calibrations to keep your tonnes trustworthy. Legal-for-trade verification is a separate, regulated process. Most of our work is routine maintenance, not trade.