Digital Scale Accuracy Guide 363: A Practical Question Answered

You place an item on one digital scale and record its weight. Then you move it to another scale and get a different number. Which result should you trust?

A disagreement does not automatically prove that either instrument is broken. The scales may use different units, display increments, tare references, or operating conditions. They may also differ in performance, calibration state, or suitability for the load.

The practical question in Guide 363 is: Why do two digital scales show different weights, and how can you identify the cause?

The best starting point is a controlled comparison with a suitable reference. Establish the same measurement task on each instrument, check each scale independently, and record repeated readings before making adjustments.

This guide explains what the disagreement can mean, what it cannot establish, and how to choose a useful next step. Find the instructions for your AWS model in the American Weigh Scales manual directory before changing settings or attempting calibration.

The practical answer

Two digital scales can disagree because they do not measure or display the result under identical conditions. Differences in units, rounding, zero, tare, load placement, environment, and sample behavior should be checked first.

Instrument performance is another possibility. One or both scales may have a consistent measurement error, inconsistent readings, or insufficient capability for the quantity being weighed.

Use a suitable known reference on each scale under appropriate conditions. Repeat the measurements and compare both agreement with the reference and consistency within each instrument.

Do not assume that the more expensive scale, newer scale, or display with more decimal places is correct. A direct comparison between two unknown instruments does not establish the accepted mass of the object.

First confirm that both screens use the same unit

A unit mismatch can create an apparent disagreement even when both scales are functioning normally. Check the unit symbol beside the number rather than relying on memory.

Grams, ounces, troy ounces, grains, and other units represent different quantities. Similar-looking abbreviations deserve careful reading on a small display.

Where both models support it, select the same unit for the comparison. Record the unit with every result so the values remain interpretable later.

If conversion is necessary, use the correct conversion and avoid rounding early. Do not compare a rounded converted value with a more detailed display and call the remaining difference an instrument fault.

Confirm normal weighing mode as well. Counting, percentage, or hold functions can present values that are not directly comparable to an ordinary live mass reading.

Different display increments can produce different numbers

Readability describes the smallest displayed increment. One scale might display whole grams while another displays tenths or hundredths.

An object near a display transition can therefore produce different-looking readings without establishing a significant performance problem. The finer display may show detail the coarser instrument cannot present.

That does not mean every difference is explained by rounding. Compare the actual increments, model specifications, and task requirements before drawing a conclusion.

Adam Equipment’s precision balance terminology distinguishes readability from accuracy and consistency.

More decimal places provide a finer display, not an automatic guarantee of a closer result. A detailed but biased reading can still be less useful than a suitable verified instrument with fewer digits.

Accuracy and agreement are different questions

If two scales agree, they can still share a similar error. If they disagree, either one or both may differ from the accepted mass.

For example, two instruments could display 100.20 grams for a suitable 100.00-gram reference. Their agreement would be good, but their indicated difference from that reference would remain.

A third scale that displays another value does not automatically settle the question. Without an adequate reference and controlled conditions, you have added another unknown instrument.

The useful question is not simply which number appears most often. It is how each result compares with an accepted reference under suitable conditions.

This distinction prevents a common mistake: adjusting one scale solely to match another whose performance has not been established.

Compare each scale’s repeatability

Before comparing instruments, check whether each gives consistent results by itself. Remove and replace the same stable reference under unchanged conditions.

Record several placements, using a procedure appropriate to the task. Reading the same unchanged display repeatedly does not test repeated loading.

If one instrument produces a wide spread while the other stays consistent, investigate the first instrument’s conditions and operation. A fixed adjustment will not necessarily resolve inconsistent behavior.

If both are consistent but separated by a fairly steady amount, a reference comparison can help assess their indicated errors.

Do not discard inconvenient readings simply to improve agreement. Record disturbances or invalid trials explicitly and follow any required procedure for repeating them.

Use an appropriate reference instead of a guessed object

The object you are comparing may not have a sufficiently known mass. A coin, product package, or familiar tool can help reveal a change, but it is not automatically a precise reference.

Choose an appropriate test weight for both instruments and the purpose of the check. Confirm its accepted value, condition, and relevant documentation.

NIST’s overview of mass standard calibration explains that weight calibration establishes a mass value and associated uncertainty.

For precise work, nominal mass alone may be insufficient. Use the information required by your procedure, including any calibrated value and uncertainty.

Do not select the reference solely because it makes one instrument look correct. Suitability comes from the measurement requirement and scale capability.

Prepare each instrument according to its own instructions

The scales may have different startup, warmup, or acclimatization requirements. Follow each model’s instructions rather than forcing both through an identical button sequence.

Place each on suitable firm support. Check that the housings do not rock and that nothing is trapped beneath their feet.

Use the intended platform and correctly positioned accessories. A transport cover or removable tray should be arranged as the manual specifies.

Confirm a normal ready state with an empty platform unless directed otherwise. Resolve warnings before recording comparison measurements.

The AWScales product FAQ provides general operating guidance. The exact model manual governs limits, controls, and calibration references.

Different locations can explain different behavior

Two instruments on opposite sides of a room may face different airflow, sunlight, vibration, or support conditions. Even nearby positions on a flexible bench can behave differently.

Look for vents, windows, fans, appliances, and activity on shared surfaces. Keep hands away from the table when recording readings.

If practical, compare the instruments in the same suitable location sequentially, allowing each to meet its preparation requirements. Do not move a sensitive instrument and read it immediately without considering acclimatization.

METTLER TOLEDO’s guide to weighing the right way discusses installation, operation, and sample-related influences.

A location comparison is useful only if you control the other relevant factors. Changing the reference, container, and location together does not isolate the environment.

Check zero and tare separately

One scale may be reporting gross weight while the other subtracts a stored container value. The numbers will differ even if both operate as instructed.

Start with an empty platform and clear tare through each model’s documented procedure. Confirm the expected zero or ready indication.

Then decide whether the comparison concerns the bare reference or a net load in a container. Use the same intended measurement on both instruments.

Do not assume a zero display means an empty platform. A tared container remains physically present and uses capacity.

If removing a container produces a negative value, consider the stored tare before diagnosing a fault. Record the sequence so the displayed reference remains clear.

Include container fit and external contact

A bowl may sit properly on one platform but touch the housing on another. That changes how its weight is supported and can make the comparison misleading.

Inspect the arrangement from the side as well as above. Check cables, walls, tools, and surrounding work surfaces for contact.

Use a smaller rigid container if appropriate, or compare the reference directly on the intended platform according to the procedure.

Keep hands, scoops, and packaging away during measurement. External contact can introduce a force without obvious container movement.

If direct-platform readings agree but container readings do not, focus on the container arrangement and tare process before changing calibration.

The sample may change between measurements

A comparison assumes the object remains sufficiently unchanged. Liquids can evaporate, materials can gain or lose moisture, and small particles can be lost during transfer.

A warm or cold item may also interact differently with the weighing conditions. Movement inside a container can delay a usable static reading.

Use a stable suitable reference to investigate the instruments before relying on a changing sample. Keep handling consistent and avoid unnecessary transfer.

Do not infer that the second instrument is wrong solely because its reading differs from an earlier sample measurement. Time and handling may have changed the task.

Record the measurement order and elapsed time where relevant. Those details can help explain a trend that is independent of either scale.

Different capacities can mean different suitability

A high-capacity shipping scale and a fine-resolution pocket scale serve different tasks. A small object may be unsuitable for meaningful comparison on the larger instrument.

Capacity is the maximum permitted load; it does not establish the smallest quantity that can be measured adequately for your purpose.

Likewise, fine readability does not prove that a scale meets a demanding relative accuracy requirement at every small load.

METTLER TOLEDO’s precision balance guidance explains that minimum suitable weight depends on performance, environment, and required accuracy.

Choose references and acceptance criteria that make sense for both instruments. If their intended applications do not overlap adequately, a direct comparison may answer very little.

Resolve power and warning conditions first

A low-battery warning, restart, overload message, or instability indication makes the session unsuitable for an ordinary comparison until addressed.

Use the specified power arrangement. For removable-battery models, replace depleted cells with a fresh matching set and confirm polarity.

Energizer’s battery care guidance explains appropriate replacement practices and avoiding mixed battery sets.

Do not use an unapproved adapter or improvised contact spacer. Correct operation should rely on the intended electrical supply.

Once the warning clears, repeat normal startup and check each scale with the reference. A cleared message is a starting point for verification, not proof that every measurement is now correct.

A controlled comparison procedure

Use an appropriate stable reference within both instruments’ capacities. Follow any formal procedure required for your application; this sequence is a practical screening approach.

1. Identify both models and their relevant specifications.

2. Confirm the reference value and suitability.

3. Prepare each scale as instructed.

4. Select the same supported unit and normal mode.

5. Establish the intended empty-platform reference.

6. Place the reference gently near the center.

7. Record the settled value using the documented ready criterion.

8. Remove the reference and observe the return toward zero.

9. Repeat placements under unchanged conditions.

10. Compare each series with the accepted reference.

METTLER TOLEDO’s routine testing resources describe distinct checks for sensitivity, repeatability, and other performance characteristics.

Record original results before making adjustments so you retain evidence of the initial disagreement.

Worked example: Two consistent scales that disagree

Imagine a suitable reference with an accepted mass of 100.00 grams. Two hypothetical instruments produce the following screening results.

Placement	Scale A	Scale B
1	100.02 g	100.18 g
2	100.03 g	100.17 g
3	100.02 g	100.18 g
4	100.03 g	100.19 g
5	100.02 g	100.18 g

Scale A spans 0.01 gram; Scale B spans 0.02 gram in this example. Both groups are relatively tight compared with the separation between them.

Their readings also differ from the stated reference. These observations support further assessment of indicated error at this load.

They do not establish a universal pass or fail. Reference uncertainty, model performance, procedure, and application tolerance still govern the decision.

Worked example: One scale gives scattered readings

Now imagine Scale A remains consistent while Scale B produces changing values over repeated placements.

The first priority is to investigate Scale B’s environment, support, container contact, sample stability, and operating state. A fixed adjustment may not address the cause of the spread.

Do not take the average of scattered values and assume the problem is solved. The individual measurements may remain unsuitable for the task.

Repeat the comparison after correcting an identified disturbance and preserve both sets of results. A clear improvement helps connect the action with the behavior.

If suitable conditions and dependable power do not resolve inconsistency, seek support rather than repeatedly changing calibration.

Check relevant loads rather than only one point

Agreement at a single reference mass does not establish agreement across the complete range. Your task may involve several different load levels.

Use appropriate references and defined test points when the application requires broader verification. Keep every load within capacity, including containers and accessories.

Do not stack objects with assumed masses and treat the sum as a precise standard. The quality of the reference remains part of the comparison.

A difference that changes across load levels deserves further assessment. Do not assume one matching point proves the entire measurement response.

For formal work, follow a procedure specifying test points, conditions, tolerances, and documentation. A simple consumer comparison cannot establish all performance characteristics.

Do not adjust one scale merely to match the other

Adjustment should follow an appropriate reference and procedure. Making two unknown instruments agree can leave both wrong.

Keep the original observations and confirm which model supports user adjustment. Use the specified reference mass and exact control sequence where adjustment is appropriate.

Distinguish the product’s adjustment routine from a formal calibration service. The terminology in everyday manuals can differ from the full measurement process.

After any adjustment, verify the resulting performance independently as required. A success message does not replace checking the measurements.

If a device fails to adjust or still disagrees with suitable references, contact support rather than repeating the same sequence indefinitely.

Inspect care and handling history

Consider whether either scale was dropped, overloaded, stored under pressure, exposed to moisture, or moved recently. History can help explain a change in performance.

Inspect accessible areas for debris and platform interference with the instrument switched off. Remove only accessories identified as user removable.

Use the AWScales maintenance guidance alongside the model instructions. Do not pry up the platform or open the main housing.

A clean exterior does not rule out damage, and an impact history does not prove a specific internal fault. Record observations without overdiagnosing.

If the disagreement began after a damaging event, seek assessment before relying on the instrument for measurements that matter.

Decide whether the difference matters to your task

A numerical difference must be evaluated against an appropriate requirement. Avoid declaring every last-digit change unacceptable or every small percentage acceptable.

Define the allowed measurement difference through the process, manufacturer information, or applicable procedure. Consider reference uncertainty and decision rules where required.

Do not choose the tolerance after seeing the results solely to make them pass. That defeats the purpose of checking performance.

Keep rounding appropriate to the measurement and reporting requirements. Extra digits from a calculator do not create extra knowledge about the scale.

For demanding applications, obtain qualified advice about suitability rather than translating an informal comparison into a formal assurance claim.

Keep a comparison record you can use later

Record both model numbers, capacities, readability, selected units, power sources, locations, and preparation. Identify the reference and its accepted value.

Keep every trial, any warning, and observations about settling or zero return. Note container use and measurement order.

Document corrective actions and subsequent results separately. “Moved from flexible bench to stable table; spread decreased” is more informative than “fixed.”

The AWScales Knowledge Center provides additional operating resources for a consistent routine.

Comparable records over time can reveal a new change or recurring setup issue. Unrelated measurements with different household objects are much harder to interpret.

When to ask for help or choose different equipment

Seek help if one instrument remains inconsistent under suitable conditions, a documented adjustment fails, or reference differences exceed the relevant criteria.

Prepare the comparison record and mention any fall, overload, spill, or recent configuration change. Use the American Weigh Scales contact page for model-specific assistance.

If the scales serve different applications, review specifications before deciding that either is defective. A suitable instrument must match the load and required performance.

For portable small-load tasks, compare the AWScales pocket scale range with your actual capacity, platform, and measurement needs.

Do not choose a replacement solely because its display has more digits. Appropriate capability and verified performance matter more than appearance.

Frequently asked questions

Which scale is correct when two disagree?

You cannot determine that from the disagreement alone. Check both against a suitable accepted reference under controlled conditions.

Is the newer scale automatically more accurate?

No. Age alone does not establish current performance. Setup, suitability, handling, and reference checks are more useful evidence.

Can rounding explain different weights?

Sometimes. Different display increments can produce different-looking results, but compare the actual difference and specifications before assuming rounding explains everything.

Should I average the two scale readings?

Averaging two unknown results does not establish the accepted mass. Investigate each instrument independently instead.

Why do my scales agree with a weight but disagree with a sample?

Check sample changes, container contact, tare, temperature, handling, and suitability for that load. One reference point does not reproduce every task.

What is the practical answer to Guide 363?

Align the measurement conditions, verify each scale independently, and compare with an adequate reference. Agreement between instruments is useful information, but it is not proof of accuracy.

Resolve the comparison through a reference

When two digital scales show different weights, begin with units, display increments, zero, tare, setup, and sample behavior. Then assess each instrument using a suitable reference and repeated placements.

Keep the evidence before adjusting anything. The goal is a result that meets the task’s requirements, supported by a controlled comparison rather than a preference for one screen.