A railcar weighing scale is the only tool that stops an overload before the railroad finds it. Railroads weigh cars on their own track scales en route, and if a car is over its load limit, it gets set out, held and billed. The penalty side of that story is in our guide to overloaded railcar charges. This post is about the other end: the scale at your plant, how accurate it actually is, and the loadout routine that keeps every car under its number.
Most overloads are not caused by a crew trying to squeeze in extra tons. They come from a scale nobody has checked, a target that ignores car-to-car differences, or a margin that is smaller than the scale's own error. All three are fixable at the source.
What the Railroad's Scale Is Checking
A railroad scale checks whether a car's lading is over the car's load limit, or whether its gross weight is over a track limit on the route. The load limit is stenciled on the car as LD LMT and carried in the car's Umler record. It is roughly the gross rail load (GRL) minus the light weight (LT WT) of the empty car.
Two details matter for your scale program:
- A track scale reads gross weight. It does not know that 1,500 lb of what it sees is frozen residue from the last load, ice in the bottom of a gondola, or a liner added in the shop. If your crew loads to a net target and the car's real empty weight is heavier than the stencil, the car can go out heavy even when your net math was right.
- Every car has its own limit. Two covered hoppers in the same series can carry different light weights after repairs or reweighing. A fleet-average target puts the heaviest cars over.
Railroads also catch "billed" overloads, where the weight on the bill of lading is already over the limit when the waybill arrives. That one is a paperwork failure, not a scale failure, and the fix is in the routine below. See our rail car types guide for how capacities differ by equipment.
Types of Railcar Weighing Scales
There are four common ways to weigh rail freight at a plant: a static track scale, a coupled-in-motion track scale, an uncoupled-in-motion track scale, and weighing the product before it drops into the car with a loadout bin or belt scale. Each has a different accuracy and fits a different operation.
Static track scale
The car is stopped on the scale deck and weighed. Short cars fit in one draft. Longer cars may be weighed in two drafts, one truck at a time, and the readings added. Static weighing is the most accurate option and the reference other methods are tested against. The cost is time: every car has to be spotted, stopped and moved off.
Coupled-in-motion scale
Cars roll across the scale in a train at a set speed, still coupled, and the system records each car. This suits unit-train and large block loadouts where stopping each car would slow the whole operation. Accuracy depends on track approach conditions, speed and train handling, and Handbook 44 notes that approach track matters more for a system used for individual car weights than for one used only for whole-train totals.
Uncoupled-in-motion scale
Cars are cut loose and roll across the scale one at a time, often in a hump or gravity yard. Because each car crosses alone, the coupler forces of a moving train are not in the reading, and the single-car tolerance is tighter than for coupled weighing.
Loadout bin or belt scale
At many grain, coal, aggregate and mineral loadouts, the product is weighed before it reaches the car. An automatic bulk weighing system batches product through a weigh hopper (Handbook 44, Section 2.22). A belt-conveyor scale totals product as it moves on the belt (Section 2.21). These are fast and can be very accurate, but they measure what left the bin, not what is sitting in the car. A track scale check on the loaded car is what confirms the result.
| Scale type | How it weighs | Single-car tolerance basis (HB 44) | Best fit |
|---|---|---|---|
| Static track scale | Car stopped on the deck, one or two drafts | Class III L maintenance tolerance (Table 6) | Low-to-medium volume, cars near the limit |
| Coupled-in-motion | Cars rolling coupled in a train | Up to 3× static maintenance tolerance; train total within 0.2% (T.N.3.6) | Unit trains and large blocks |
| Uncoupled-in-motion | Cars rolling one at a time | Within static maintenance tolerance (T.N.3.7) | Yards and gravity tracks |
| Loadout bin / belt scale | Product weighed before it enters the car | Separate codes (Sections 2.21, 2.22) | High-volume bulk loadouts |
How Accurate a Railcar Scale Has to Be
A static railway track scale is a Class III L device under NIST Handbook 44. Its maintenance tolerance is counted in scale divisions (e): 1 division for loads up to 500 divisions, 2 divisions up to 1,000, and 1 more division for each additional 500 divisions or fraction of 500. Acceptance tolerance, which applies to new or newly repaired scales, is half the maintenance tolerance.
Here is what that means on a loaded car. Illustrative example: a 286,000 lb gross car on a static scale with 100 lb divisions is a load of 2,860 divisions.
- Static maintenance tolerance: 6 divisions, or ±600 lb.
- Static acceptance tolerance: 3 divisions, or ±300 lb.
- Uncoupled-in-motion, any single car: within the static maintenance tolerance, ±600 lb.
- Coupled-in-motion, any single car (trains of five or more cars, using individual car weights): no error over 3× static maintenance tolerance, ±1,800 lb. No more than 5% of cars may be off by more than 1,200 lb, and no more than 35% by more than 600 lb.
- Coupled-in-motion, whole train or group: the sum of the in-motion weights must be within 0.2% of the sum of the static weights.
Your scale's actual division size and capacity are on its marking plate, so run the numbers for your own device. The point holds either way. A coupled-in-motion scale is built to get the train total right, and it is allowed to be well off on any single car. If you use it to decide whether one car is under its limit, your margin has to cover that.
Two more rules shippers should know. Class III L railway track scales must have a recording element, so a printed or stored weight record exists for every weighment (UR.1.5). And a coupled-in-motion system used only for dynamic weighing must have a certified reference scale available for testing (UR.5).
Testing and Certification
Handbook 44 sets how a railway track scale is tested. The weights and measures authority with jurisdiction over the scale sets how often. Under N.3.1, an initial or subsequent test of a railway track scale needs a test-weight load of at least 80,000 lb, and a strain-load test up to the used capacity is recommended. That is why track scales are tested with scale test cars, heavy cars of precisely known weight, rather than stacks of hand weights.
After a repair that could affect accuracy, a scale can go back into temporary service on an interim test, but only if all of these are met:
- a test-weight load of at least 30,000 lb;
- a shift (section) test with at least 30,000 lb;
- a strain-load test to at least 25% of scale capacity;
- all results within tolerance; and
- the official with statutory authority notified right away.
In-motion systems are tested in motion. An uncoupled-in-motion scale is tested statically first, then with at least three reference weight cars passed over three times at normal speed and direction (N.5). A coupled-in-motion system weighing trains of ten or more cars is tested with a test train of at least ten cars run over at least five times in each mode of operation, or with a distributed-car test train that represents normal operation (N.4).
For a shipper, certification is about more than compliance. On Norfolk Southern, overload charges tied to rain, snow or ice are waived only when you send a certified weight certificate showing the car was under its stenciled load limit when loaded, within 24 hours of the overload notice, not counting weekends and holidays (NS 8002-A, Item 5000). A scale with a lapsed test cannot produce that certificate.
Setting a Load Target That Covers Scale Error
The safe load target for a railcar is its own load limit minus a margin that covers your scale's error, the railroad scale's error, and any weight the load can pick up in transit. A margin smaller than your own scale's tolerance is not a margin.
Illustrative example: a covered hopper stenciled at 286,000 lb GRL with a light weight of 63,500 lb has a load limit of 222,500 lb. One percent of that is 2,225 lb, which is where the NS Item 5000 penalty begins.
| How the car is weighed at the plant | Single-car error allowed (100 lb divisions) | Room left under the 1% line |
|---|---|---|
| Static, new or newly repaired | ±300 lb | About 1,900 lb |
| Static, in service | ±600 lb | About 1,600 lb |
| Coupled-in-motion, in service | ±1,800 lb | About 400 lb |
The railroad's scale has its own tolerance too, and the two errors can stack in the same direction. On top of that, an open-top car of aggregate, scrap or wood chips can take on rain or snow after it leaves. NS counts weather-related overloads as overloads. A car loaded to the stencil in dry weather can arrive over after a storm.
So build the target from your own data:
- Scale error: use the last test report, not the legal tolerance alone. A scale that tested 200 lb heavy is telling you something.
- Moisture pickup: for open-top service, compare origin and destination weights on a sample of cars across wet and dry months.
- Tare drift: track how far actual empty weights run from the stencils on your fleet.
Loading a few hundred pounds light on every car costs freight efficiency. Going over costs a penalty plus demurrage, switching, contractor time and a late, short delivery. On most lanes, the light car wins. Run a lane through the railcar capacity tool to see what the margin costs in tons or bushels.
A Loadout Routine That Stops Overloads
The routine below stops overloads at the source. It works for a single-car spur and for a high-volume loadout. The difference is how much of it the loadout control system does for you.
- Pull each car's load limit before loading. Take it from the stencil or the Umler record, car by car. Flag any car whose stencil and Umler record disagree.
- Weigh or inspect the empty. If you have a track scale, weigh the car empty and compare it with the stenciled light weight. A big gap means residue, ice or a stencil that is out of date.
- Set a per-car target. Target equals that car's load limit minus your margin. Never use a fleet average.
- Load to the target. Use the loadout bin or batch weigher to land close to the target, filling in stages for the last part of the load.
- Weigh the loaded car on a certified scale. Keep the printed or stored ticket with the car number, time and scale ID.
- Hold and trim anything over target. Pulling 2,000 lb off a car on your own track takes minutes. Pulling it off on a railroad reduction track takes days and a contractor.
- Key the bill of lading weight from the ticket. Have your system reject any net weight above the car's load limit and catch pounds-for-kilograms errors before the BOL goes out.
- File the tickets where you can reach them fast. The NS weather waiver window is 24 hours. Know where the certificate is before a notice arrives.
Weight data on the bill of lading also matters for who owns the number. Under shipper load and count terms, the weight is your statement. See shipper load and count vs. carrier load for why that matters in a dispute, and our rail BOL builder for the weight fields.
When You Don't Have a Track Scale
Without a track scale, you are estimating the car weight from something else, and every estimate adds error. Common workarounds:
- Summed truck tickets at a transload. Five or six truck scale tickets add up to the car weight, but each carries its own tolerance, and product left in a trailer or spilled at the transfer is missed. Our guide to transloading covers how those facilities run.
- Volume and density. Measuring outage in a tank car or hopper and multiplying by density works only as well as your density figure. Product density changes with moisture and temperature.
- Railroad weighing. Some railroads will weigh cars on request for a fee. BNSF lists a $500 reweigh fee in its Overload Process FAQ. Check the current tariff before you count on it.
If you load cars near their limit with no track scale, the margin has to be bigger. Ask whether a nearby facility or the serving railroad has a certified scale you can use for a sample of cars to measure your real error. A few weeks of data usually settles how much margin you need.
Weight control is one part of running a rail program. Our rail logistics services cover car supply, loading coordination and car flow, and our rail freight audit checks overload and demurrage bills against your scale tickets. To learn how railroads build these charges, the Railroad Pricing course module is free. If you have a specific loadout problem, contact us.
Sources: NIST Handbook 44, 2026 edition, Section 2.20 Scales (N.3.1, N.4, N.5, T.N.3.1, T.N.3.2, Table 6, T.N.3.6, T.N.3.7, UR.1.5, UR.5), nist.gov, doi.org/10.6028/NIST.HB.44-2026; Norfolk Southern Accessorial Services Tariff NS 8002-A, Item 5000, effective August 1, 2026; BNSF "Overload Process FAQ's," bnsf.com. Worked examples are illustrative. Tolerances depend on each scale's marked division size and capacity. This is general information, not legal or regulatory advice.
Frequently Asked Questions
What kind of scale is used to weigh railcars?
Railcars are weighed on railway track scales. A static track scale weighs a stopped car, either in one draft or in two drafts for long cars. In-motion scales weigh cars as they roll across, either coupled in a train or uncoupled one at a time. Bulk loadouts also weigh product before it drops into the car using automatic bulk weighing systems or belt scales.
How accurate does a railcar scale have to be?
Under NIST Handbook 44, a static railway track scale is a Class III L device. Its maintenance tolerance grows with the load: 1 scale division up to 500 divisions, 2 up to 1,000, then 1 more for each additional 500. Coupled-in-motion systems must total within 0.2% of static weights, and a single car can be off by up to three times the static maintenance tolerance.
How often must a railroad track scale be tested?
Handbook 44 sets how a test is done, not a single national schedule. Initial and subsequent tests need a test-weight load of at least 80,000 lb, and a repaired scale can return to temporary service on an interim test of at least 30,000 lb. The weights and measures authority with jurisdiction over the scale sets the inspection frequency.
How much below the load limit should I load a railcar?
Set the target below each car's own load limit by at least your scale's tolerance at that weight, plus any water or snow the load can pick up in transit. For an open-top car of wet aggregate that margin is larger than for a covered hopper of dry grain. Measure your own scale's error and your own moisture pickup rather than copying a rule of thumb.
Can I use truck scale tickets for a railcar weight?
You can add up certified truck tickets to estimate what went into a car at a transload, but every ticket carries its own tolerance and the errors add up across five or six trucks. Product left in a truck or spilled at the transfer is also missed. If the car is close to its limit, weigh the car itself on a track scale before release.