How to Compare Your Test Result to Manufacturer Specification

How to turn a measurement from your analyzer into a defensible pass or fail decision

Testing medical equipment is not just about getting a number.

Published August 16, 2026 · Revised September 5, 2026

Back to Biomed Basics

What This Page Explains

This page covers:

The Simple Version

Use a repeatable process: find the correct specification for the exact model and function, match it to the test point and required conditions, put both values in the same units, calculate the lower and upper limits, and then compare the unrounded result. Document the source, setup, limits, and measurement so another technician can follow the same path.

Do the limit calculation before you see the result whenever possible. If a 500 mL delivery has a tolerance of ±5% of the setting, calculate 25 mL and write down the acceptable range of 475–525 mL first. A measured 474 mL is then a failure under that rule; it does not become a pass because it is “only one milliliter off.”

Worked Bench Example

Suppose an infusion pump is set to 100 mL/hr and the service manual allows ±5% of the programmed rate at that test point. Five percent of 100 mL/hr is 5 mL/hr, so the acceptance window is 95–105 mL/hr. If the analyzer reports 96.8 mL/hr, the result passes. Record the actual 96.8 mL/hr result instead of writing only “within spec,” because the number helps the next technician see drift over time.

Now suppose the same analyzer reports 94.9 mL/hr. Before blaming or calibrating the pump, check the specified test duration, priming method, tubing set, fluid, head height, analyzer zero, and whether the manual permits an average of multiple runs. Those checks do not change the limit. They determine whether you performed a valid test. If the setup is correct and 94.9 mL/hr remains outside the 95–105 mL/hr window, document the failure and follow the manufacturer's troubleshooting or adjustment procedure.

Start With the Correct Specification

Before performing the test, find the requirement that actually applies.

Check:

Do not use a limit from:

unless the manufacturer says it applies.

Match the Specification to the Test

Suppose you are testing an infusion pump at:

100 mL/hr.

You need the manufacturer's flow-delivery specification.

Not:

Use the specification for the exact function you are testing.

Read the Whole Specification

Example:

Flow accuracy ±5% from 1 to 999 mL/hr under specified test conditions.

Do not record only:

±5%.

Also understand:

The tolerance may only apply under certain conditions.

Fixed Tolerance

A fixed tolerance gives the same allowable error regardless of the test value.

Example:

Pressure accuracy:

±3 mmHg.

Reference:

100 mmHg.

Lower limit:

97 mmHg.

Upper limit:

103 mmHg.

Measured:

101 mmHg.

Pass.

Fixed Tolerance Formula

For:

Reference ± Fixed Tolerance

calculate:

Lower Limit = Reference - Tolerance

Upper Limit = Reference + Tolerance

Example:

Reference:

200 mmHg.

Tolerance:

±3 mmHg.

Acceptable:

197 to 203 mmHg.

Percentage Tolerance

A percentage tolerance changes with the reference value.

Example:

Defibrillator energy tolerance:

±15%.

Set:

200 J.

Calculate:

200 × 0.15 = 30 J.

Acceptable:

170 to 230 J.

Measured:

198 J.

Pass.

Percentage Formula

Basic calculation:

Allowed Error = Reference × Tolerance Percentage

Then:

Lower Limit = Reference - Allowed Error

Upper Limit = Reference + Allowed Error

Example: Infusion Pump

Set rate:

100 mL/hr.

Tolerance:

±5%.

Allowed error:

100 × 0.05 = 5 mL/hr.

Acceptable:

95 to 105 mL/hr.

Analyzer:

97 mL/hr.

Pass.

Same Percentage, Different Test Point

Now test:

10 mL/hr.

Same tolerance:

±5%.

Allowed error:

10 × 0.05 = 0.5 mL/hr.

Acceptable:

9.5 to 10.5 mL/hr.

The allowed absolute error changed because the test point changed.

Percentage of Reading vs Full Scale

Be careful.

These are different.

Suppose the range is:

0 to 300 mmHg.

Specification:

±2% of reading.

At 100 mmHg:

Allowed error:

2 mmHg.

But if the specification says:

±2% of full scale

then:

300 × 0.02 = 6 mmHg.

Acceptable at 100 mmHg:

94 to 106 mmHg.

Read the exact wording.

Combined Specifications

Some specifications use both a percentage and a fixed amount.

Example:

±2% of reading + 1 mmHg.

At:

100 mmHg.

Two percent:

2 mmHg.

Add:

1 mmHg.

Total allowed error:

3 mmHg.

Acceptable:

97 to 103 mmHg.

Follow the manufacturer's formula exactly.

Minimum Limits

Some tests specify only a minimum.

Example:

Battery runtime:

Minimum 90 minutes.

Measured:

94 minutes.

Pass.

Measured:

89 minutes.

Fail.

There is no plus/minus calculation.

Maximum Limits

Example:

Leak rate:

Maximum 50 mL/min.

Measured:

42 mL/min.

Pass.

Measured:

56 mL/min.

Fail.

Again, know which direction matters.

Range Specifications

Sometimes the manufacturer gives an acceptable range directly.

Example:

Supply voltage:

22.8 to 25.2 VDC.

Measured:

23.9 VDC.

Pass.

No calculation required.

Nominal Values Need Tolerances

Suppose a schematic labels a rail:

24 V.

That is often a nominal value.

Do not automatically fail it because you measure:

23.6 V.

Find the actual allowed tolerance.

Units Must Match

Before comparing two numbers, make sure the units match.

Examples:

Never compare raw numbers from different units.

Example: Flow Conversion

Analyzer displays:

0.1 L/min.

Specification:

100 mL/min.

Since:

1 L = 1000 mL.

0.1 L/min = 100 mL/min.

Same result.

Different unit.

Test Conditions Must Match

A specification may only apply under defined conditions.

Examples:

If you change the conditions, the comparison may no longer be valid.

Example: Battery Runtime

Manufacturer says:

Minimum 90 minutes under standard operating load.

You test:

Screen off, Wi-Fi disabled, modules removed.

Result:

95 minutes.

That may not be comparable to the published requirement.

Match the required test configuration.

Test Point Matters

Suppose the manual specifies:

24 VDC at Connector J5.

You measure:

23 VDC at a board farther downstream.

The result may include voltage drop in wiring.

Do not compare it directly unless that is the specified point.

Reference Value vs Device Setting

Sometimes your reference is the value you intentionally apply.

Example:

Patient simulator outputs:

120 bpm.

Monitor displays:

119 bpm.

Reference:

120 bpm.

Measured:

119 bpm.

Now compare to the monitor's heart-rate accuracy specification.

Other Times the Analyzer Is Measuring Output

Example:

Defibrillator set:

200 J.

Analyzer measures:

195 J.

The set value is the target.

The analyzer gives the actual delivered result.

Compare according to manufacturer tolerance.

Accuracy of the Analyzer Matters

Your analyzer is not perfect.

It has its own:

If the device tolerance is tighter than your analyzer can reliably measure, your pass/fail decision may not be defensible.

Example

Device requirement:

±1 mmHg.

Analyzer accuracy:

±3 mmHg.

That analyzer is not very useful for proving the device meets a ±1 mmHg specification.

The reference must be suitable for the test.

Resolution Matters Too

Suppose the pass range is:

9.95 to 10.05 V.

Meter resolution:

0.1 V.

It displays:

10.0 V.

You do not have enough detail to confidently determine how close the actual value is to the limit.

Test Equipment Calibration

Before trusting the analyzer, make sure it is within required calibration status.

An expired or unverified analyzer can undermine the entire result.

Apply the Limit Before Looking at the Result

A good habit is to calculate the acceptable range first.

Why?

Because otherwise it is easy to see:

104.8

and think:

That looks close.

Instead determine:

Pass range:

95 to 105.

Then compare.

That reduces bias.

Borderline Results

Suppose the limit is:

95 to 105.

Measurement:

105.0.

That may pass if the limit is inclusive.

Now:

105.1.

That may fail.

Do not casually stretch the requirement because the result is close.

Do Not Round Too Early

Example:

Actual:

105.46.

Limit:

105.4.

If you round to:

105

before comparison, you may incorrectly pass it.

Keep sufficient precision until the decision is made.

But Do Not Add Fake Precision

If the analyzer reports:

98.2

do not document:

98.23714.

Use the resolution the equipment actually provides.

Repeated Measurements

Some procedures require multiple test runs.

Example:

Three infusion measurements.

Do not automatically choose the best one.

Follow the manufacturer procedure for:

Do Not Average Failures Away

Results:

98 99 110

Average:

102.3.

If one point violates the required acceptance criteria, the average does not necessarily make the device pass.

Use the defined test method.

Multi-Point Testing

Many devices must be checked at more than one point.

Example:

Pressure:

50 mmHg 100 mmHg 200 mmHg

A device may pass at 100 and fail at 200.

One passing point does not prove the whole range.

Why Multiple Points Matter

Testing different points can reveal:

These patterns can help troubleshoot the failure.

Offset Example

Reference → Device

50 → 55 100 → 105 200 → 205

Constant:

+5.

That suggests a consistent offset.

Gain Example

Reference → Device

50 → 55 100 → 110 200 → 220

Error increases proportionally.

That suggests a scaling or gain issue.

Nonlinear Example

Reference → Device

50 → 50 100 → 101 200 → 230

The high end behaves differently.

A single mid-range test would have missed it.

Passing One Point Does Not Prove Calibration

A device can be perfect at one point and wrong elsewhere.

Use all required test points.

Manufacturer Test Sequence

Some procedures require testing in a defined order.

Example:

Follow the sequence when specified.

Some systems behave differently depending on hysteresis or previous state.

Hysteresis

A device may give slightly different readings depending on whether the value is approached from above or below.

If the manufacturer specifies ascending and descending tests, do both.

Environmental Conditions

Accuracy specifications may depend on:

If the device and analyzer are outside those ranges, the result may not be valid.

Warm-Up Time

Some devices or analyzers require warm-up.

If the specification applies after:

15 minutes

and you test after:

30 seconds,

you may get misleading results.

Accessory Effects

The test may depend on:

A leaking test hose can make a good device fail.

Verify the test setup before adjusting the equipment.

Zero the Analyzer When Required

Some analyzers need:

before measurement.

Skipping that step can create an artificial error.

Device Zeroing Matters Too

Example:

Pressure channel not zeroed before applying test pressure.

Result appears wrong.

The problem may be setup rather than device accuracy.

Correct the Setup Before Adjusting the Device

If a test fails, do not immediately calibrate.

First verify:

You want to know the device failed, not your test.

Failed Specification Is a Symptom

Suppose:

Reference:

100 mmHg.

Device:

110 mmHg.

Allowed:

97 to 103.

Fail.

You now know:

The device is out of specification at that test point.

You do not yet know why.

Possible causes include:

The failed result becomes the next troubleshooting clue.

Do Not Calibrate Blindly

If the device fails because of:

calibration is not the correct repair.

Find the cause first.

After Adjustment, Test Again

If calibration is required:

Do not only retest the point you adjusted.

Complete the manufacturer's required post-calibration verification.

Adjustment at one point may affect others.

Real-World Example: Defibrillator

Set:

200 J.

Manufacturer tolerance:

±15%.

Allowed:

170 to 230 J.

Analyzer:

194 J.

Pass.

You do not need exactly 200 J.

Real-World Example: Infusion Pump

Set:

50 mL/hr.

Tolerance:

±5%.

Allowed error:

2.5 mL/hr.

Acceptable:

47.5 to 52.5 mL/hr.

Analyzer:

46.8 mL/hr.

Fail.

Now verify setup and troubleshoot.

Real-World Example: Pressure

Reference:

100 mmHg.

Accuracy:

±3 mmHg.

Monitor:

103 mmHg.

Pass.

Monitor:

104 mmHg.

Fail.

The difference between those two results matters because the specification defines it.

Real-World Example: Power Supply

Nominal:

24 VDC.

Tolerance:

±5%.

Allowed:

22.8 to 25.2 V.

Measured:

22.5 V.

Fail.

Do not simply write:

About 24 volts.

Real-World Example: Battery Runtime

Requirement:

Minimum 60 minutes.

Measured:

58 minutes.

Fail.

It is tempting to say:

Basically an hour.

But a defined minimum is a defined minimum.

Real-World Example: Ventilator Volume

Set:

500 mL.

Tolerance:

±10%.

Allowed:

450 to 550 mL.

Analyzer:

482 mL.

Pass.

No adjustment needed.

Common Mistakes

Comparing Without Finding the Actual Specification

Know the limit first.

Using a Specification From the Wrong Model

Verify documentation.

Forgetting Percentage Calculations

Calculate the range.

Mixing Units

Convert first.

Ignoring Test Conditions

Match the procedure.

Rounding a Failure Into a Pass

Keep enough precision.

Assuming One Test Point Proves the Entire Range

Use all required points.

Calibrating Before Checking the Test Setup

Make sure the failure is real.

Ignoring Analyzer Accuracy

Your reference matters too.

A Useful Comparison Framework

For every measurement, write:

Target / Reference

What should the value be?

Specification

What error is allowed?

Acceptable Range

What values pass?

Measured Result

What did the analyzer show?

Decision

Pass or fail?

That structure removes a lot of ambiguity.

Another Useful Question

Ask:

If another technician saw only my recorded numbers, could they independently reach the same pass/fail conclusion?

That is a good test of your documentation.

What Did You Actually Prove?

Suppose:

Set:

100 mL/hr.

Allowed:

95 to 105.

Measured:

99 mL/hr.

You proved:

The device's measured delivery was within the manufacturer's specified acceptance range at that test point under the conditions tested.

You did not prove:

That is why required multi-point verification matters.

Final Thoughts for Biomeds

A test result by itself does not tell you whether a device is good.

You need the specification.

Start with the correct manufacturer requirement.

Match it to the exact function and test point.

Calculate the acceptable range.

Make sure your units and test conditions match.

Then compare the actual measurement.

Do not decide based on:

Looks close.

Do not chase perfect numbers either.

A device does not usually need to produce the exact target value.

It needs to meet its required performance specification.

When you can clearly show:

Target → Tolerance → Acceptable Range → Measured Result → Pass/Fail

your verification becomes objective, repeatable, and much easier to defend.

— Jake

Important Note

Manufacturer specifications, acceptance limits, test procedures, test points, and required analyzer capabilities vary by device, model, revision, and software version. Use current manufacturer documentation, approved procedures, appropriately maintained test equipment, and applicable facility requirements when making pass/fail or return-to-service decisions.

Related Biomed Basics

Back to Biomed Basics