How Should a Mechanical Smart Water Meter Sensor Be Validated? From Low-Speed Movement to Continuous High-Speed Testing

1. Before Testing, Start With the Mechanical Structure

When a new mechanical meter arrives, the first step is usually not to connect an oscilloscope.

Start by understanding the meter itself.

At minimum, it is useful to clarify:

• which component actually moves with the mechanical metering mechanism;
• how much volume corresponds to one mechanical rotation;
• the approximate rotor speed at normal flow;
• the possible maximum rotor speed at maximum flow;
• whether reverse-flow recognition is required;
• where the conductive target can be integrated;
• where the PCB sensing structure can be installed.

If these points are not clear, a clean pulse waveform alone does not necessarily mean the complete meter architecture is correct.

In a mechanical smart meter, the mechanical movement itself is the source of the electronic metering information.

Mechanical structure, conductive target and PCB sensing position should be evaluated together before electrical validation begins.

2. Low-Speed Testing Can Be More Revealing Than Normal-Speed Testing

At normal speed, many sensing solutions can appear to work well.

The more interesting behavior often appears when the rotor slows down.

At low speed, mechanical movement may not be perfectly uniform.

The rotor may:

• approach the sensing position slowly;
• pause briefly;
• oscillate slightly;
• then continue forward.

The real question is:

Does one real mechanical metering event produce only one valid electronic metering event?

If the sensing response is too weak, real movement may be missed.

If the system is overly sensitive around a threshold position, small mechanical oscillations may be interpreted incorrectly.

So the purpose of low-speed testing is not simply to prove that the sensor is “sensitive.”

The more useful question is:

Does the metering decision remain clean and stable during slow and non-uniform mechanical movement?

3. The Sensing Threshold Is Worth Testing Separately

Laboratory testing often uses continuous rotation.

Real pipe conditions may be less predictable.

Pressure variation, small vibration or stopped flow can leave the mechanical mechanism close to the sensing threshold.

That condition is worth checking separately.

A simple validation sequence can be:

1. move the conductive target slowly toward the sensing area;
2. pause near the threshold position;
3. apply a small forward/reverse oscillation;
4. continue the movement forward.

Then observe:

• whether duplicate pulses appear;
• whether false counts occur;
• whether forward/reverse recognition becomes unstable;
• whether the Host MCU count still matches the actual mechanical movement.

This test looks simple, but it can reveal interface behavior that may not appear during a basic constant-speed rotation test.

4. If Reverse Flow Matters, Test It From the Beginning

If the final product requires:

• reverse-flow detection;
• bidirectional metering;
• reverse-flow alarm;

forward and reverse operation should be included in the sensing validation from the beginning.

A straightforward test sequence is:

Key points to observe include:

• whether the forward pulse remains continuous;
• whether the reverse pulse remains continuous;
• whether direction changes cause missed or duplicate counts;
• whether the Host MCU accumulation logic remains correct.

With WF8002 Metering Chip Based Non-Magnetic Inductive Sensing, forward and reverse information is part of the metering layer itself, rather than something the communication layer has to infer later.

Real forward/reverse pulse waveform capture during WF8002-based metering validation.

5. Then Move to the Maximum Rotor Speed

Low-speed validation answers one question:

Can slow and complex mechanical movement still be interpreted cleanly?

High-speed validation answers another:

Can the sensing system and the complete pulse chain keep up?

Maximum rotor speed can vary significantly between meter projects, so one universal speed value should not be applied to every meter.

For some high-resolution reference designs, if the project requires 10 rps and defines:

1 rotation = 1 valid pulse

then the complete chain must continuously support:

The important point is not reaching 10 rps momentarily.

The real objective is:

Maintain complete and stable counting during continuous operation at the target maximum rotor speed.

This is why sensing speed, pulse timing and Host MCU capture strategy should be validated together.

When the project defines 1 rotation = 1 pulse, 10 rps corresponds to 10 valid metering events per second.

6. Looking Only at the Sensor Pulse Is Not Enough

Many validation sessions stop at the oscilloscope.

The sensor produces 100 pulses.

Everything looks correct.

But in the final product, the oscilloscope is not doing the counting.

The Host MCU is.

So the complete chain should also be checked:

One direct validation method is to rotate the mechanical side for a known number of cycles, for example 1000 rotations.

Then record:

• sensor output pulse count;
• Host MCU received count;
• forward/reverse count;
• mechanical reference count.

The final comparison is simple:

Do these counts remain consistent with the actual mechanical movement?

This is closer to the final product than looking at the pulse waveform alone.

7. Keep Metering Running While Communication Is Active

In a real smart meter, the metering layer does not operate alone in a perfectly quiet electrical environment.

The same product may also include:

• Wireless M-Bus;
• LoRaWAN;
• NB-IoT;
• Cat.1.

So one practical complete-meter test is:

Keep mechanical metering running while the RF subsystem is transmitting normally.

Then observe:

• whether pulse output remains stable;
• whether the Host MCU misses any counts;
• whether the supply shows significant disturbance;
• whether metering data remains consistent with the RF-idle condition.

This does not assume that RF activity will cause a problem.

It simply reflects the fact that both subsystems will operate together in the final product.

8. Lower Supply Voltage Is Also Worth Testing

Battery-powered products introduce another real-world condition:

A new battery and a battery after years of service do not operate under exactly the same conditions.

Metering validation should therefore not always remain at nominal supply voltage.

Useful conditions can include:

• normal supply voltage;
• lower operating voltage;
• lower voltage + maximum rotor speed;
• lower voltage + RF activity.

The objective is not to create the longest possible test list.

The important question is:

Does the metering chain remain stable when several realistic operating conditions occur at the same time?

9. Sensing Signal Strength Can Be a Useful Engineering Metric

If the project can provide sensing signal-strength information, that data can be useful beyond simply confirming whether detection occurs.

It can help observe:

• changes in PCB-to-target installation;
• mechanical tolerance;
• assembly consistency;
• variation between different samples.

During engineering development, a reference range can first be established.

During pilot production, multiple samples can then be compared against that range.

This provides more information than a simple:

PASS / FAIL

result alone.

10. A Practical Way to Structure the Validation

This layered approach makes troubleshooting easier when an unexpected result appears.

Instead of treating the entire meter as one black box, the problem can be traced back to the mechanical, metering, interface or system layer.

11. How Does This Relate to ISO 4064 and OIML R49?

Formal water meter metrological verification is governed by established standards and test procedures.

ISO 4064-2:2024 covers water meter type evaluation and initial verification test methods, including complete water meters as well as relevant metering units such as measurement transducers and calculators.

ISO 4064-2:2024 — Water meters — Part 2: Test methods ->

OIML R49-2:2024 also provides water meter test methods covering metrological and technical evaluation.

OIML R49-2:2024 — Water meters for cold potable water and hot water — Part 2: Test methods ->

One Final Thought: Do Not Stop at “Does It Work?”

Validation of a metering sensing solution should not stop at:

“Does it work?”

A more useful set of questions is:

Under which mechanical conditions does it work? What happens at low speed? At high speed? During direction changes? And what does the Host MCU finally receive?

In a mechanical smart meter, stability does not belong to one electronic component alone.

It belongs to the complete chain:

Once this chain is clear and stable, communication and AMR / AMI systems have a much more reliable starting point.

Related Technical Resources

Product line: Inductive metering sensors — WF8002-based, non-magnetic, anti-magnetic/RF interference, RoHS 2.0, samples and datasheets available.