Frequency in hertz, period in microseconds and digits can describe the same wire reading, but they are not interchangeable unless you know the definition used by the manufacturer. First preserve the raw value. Then apply exactly the conversion, factor, zero and sign that belong to the specific sensor. Acceptance ends only when a second person reproduces the result from several samples and gets the same engineering unit.
In brief
- Hz says how many cycles the wire makes in one second, and period says how long one cycle lasts. The relation is
f = 1 / T.Digitsis not an SI unit. It is a scaled form of frequency or period defined in the manual of a given readout system.- In the current GEOKON piezometer manual, the example definition is
digits = Hz² / 1000. Do not transfer that rule to other hardware without confirmation.- Calibration factor, zero reading, sign and temperature correction solve different problems. One number does not replace the others.
- The most expensive error usually does not look like a failure. The chart is smooth, only the result has the wrong scale or the opposite sign.
Four numbers, one wire and four chances to get it wrong
A field reader shows 2828 Hz. An export from the logger contains 354 µs. An old spreadsheet uses a value of 8000 digits, and the report for the investor shows pressure in kilopascals. All four numbers can refer to the same observation. Each one, however, plays a different role.
Frequency is the result of measuring the vibration of a tensioned wire. Period is the inverse of frequency. Digits, sometimes translated as readout digits or arbitrary units, are an intermediate format. Engineering value appears only after the calibration relation, reference point and required corrections are applied. The physical principle itself and the limits of the technology are developed in the introduction to vibrating wire sensors.
This split matters financially. When a company changes logger, platform or integrator, frequency still describes the phenomenon in the sensor. The digits format may belong to a specific family of readers. If the archive contains only the final result, it is hard to verify whether the old software used the correct factor. Migration then starts to resemble spreadsheet archaeology.
| Level | Example notation | What it describes | Most common error |
|---|---|---|---|
| signal | 2828 Hz | number of cycles per second | measuring an unrelated interference component |
| signal | 354 µs | time of one cycle | treating microseconds as seconds |
| intermediate format | 8000 digits | value calculated according to the manufacturer’s definition | assuming every brand calculates it the same way |
| result | 146.2 kPa | quantity used in the design | wrong factor, zero, sign or unit |
In practice, it is worth storing at least the reading as close to the source as possible, temperature, measurement time, quality status and the identifier of the calculation rule used. The single column 146.2 is not enough, even if the heading says kPa.
Frequency and period: simple relation, tricky units
The physical relation is short:
f = 1 / T
If f is to be in hertz, the period T must be given in seconds. Vibrating wire readers often show period in microseconds. Then, before inversion, you must account for one million microseconds in a second:
f [Hz] = 1 000 000 / T [µs]
For a period of 354 µs, the result is about 2824.9 Hz. The difference from the rounded example of 2828 Hz comes from the limited number of digits in the period. If the system stores only 354 instead of a more precise measurement, part of the resolution is already lost in the first step.
That is a minor issue for a large change, but it can matter for small increments. Later displaying eight decimal places will not fix it. Digits added after resolution is lost are cosmetic.
Do not confuse resolution with accuracy
If the screen changes every 0.1 Hz, that means display resolution. It does not mean the entire measuring chain has an error of 0.1 Hz. The result is also affected by the reader’s timebase accuracy, excitation method, band selection, signal quality, sensor repeatability and final-value calibration.
The GK-404 readout specification separates display resolution from timebase accuracy. That is a good pattern for technical requirements: the client should ask for both pieces of information, not one impressively small number.
Inversion changes the direction of the scale
When frequency rises, period falls. If the algorithm compares the current period with the zero period in the same order as frequency, the sign of the result may flip. That is why it is not enough to say that the system "supports microseconds". You must check the full formula and the positive direction for the specific sensor.
What digits are and why the name guarantees nothing
In the current GEOKON 4500 Series piezometer manual, the reader calculates digits according to:
digits = f² / 1000
where f is frequency in Hz. The same manual gives an equivalent form based on period and an example: about 8000 digits corresponds to a period of about 354 µs and a frequency of about 2828 Hz. By squaring the frequency, the intermediate value is linearly related to the quantity that changes the wire tension in a typical transducer.
This is a manufacturer example, not an industry standard. Another reader may:
- use a different scale constant,
- store the square of the inverse period in a different order of magnitude,
- return a unit already multiplied by the sensor type factor,
- round the result before transmission,
- call
digitsa number with a different definition.
A field name is not documentation. A field named vw, raw, reading or digits needs a definition, input unit, formula and source software version. During migration, ask for a sample raw message and a parallel reading from a certified field reader. Comparing CSV headers alone will not reveal a scale mismatch.
Quick consistency test for three representations
Pick one record and recalculate it independently:
- Use the period in microseconds to calculate Hz.
- Use Hz to calculate digits according to the manual for the specific reader.
- Compare the result with the value sent by the logger.
- Repeat for a reading near the lower, middle and upper part of the range.
One matching sample can hide a factor or rounding error. Three points show whether the relation has the right shape. If the calibration document contains a polynomial, you also need points near the ends of the range.
From digits to engineering unit
The simplest relation for many transducers has the form:
result = (current reading - zero reading) × factor
In the GEOKON piezometer manual, this is an example of converting a difference in digits to pressure. For higher accuracy, the manufacturer also describes the use of a second-degree polynomial. A strain gauge, crack meter, load cell and piezometer do not necessarily use the same subtraction order or factor unit.
A complete calculation sheet should include:
- the identifier of the specific unit,
- the current and zero values in the same representation,
- the linear factor or the full set of polynomial coefficients,
- the unit of each coefficient,
- the definition of positive sign,
- temperature, atmospheric pressure or level correction, if the method requires it,
- the range over which the relation applies,
- the data source, document number and revision.
The 4200 Series strain gauge manual shows another important detail: for some sensors, the theoretical factor supplied with the gauge is adjusted by a batch factor because clamp mechanics affect the theoretical coefficient. That is a warning against copying a constant from a catalogue. The card for a given sensor or batch takes precedence over a number found in an article.
Acceptance example: the piezometer gives a smooth but wrong curve
Illustrative example. An integrator takes over 48 vibrating wire piezometers on a hydraulic structure. The old system exports three columns: time, reading and temperature. Reading values are around 7000-9000. The new platform treats them as Hz, squares them and multiplies by a factor. Huge pressures appear on the dashboard, so someone manually divides the result by a thousand. The chart looks reasonable.
It is still the wrong model. The 7000-9000 values were already digits, not frequency. The manual divisor accidentally restored the scale in the middle of the range, but it did not guarantee the correctness of the relation or the unit.
The team fixes the process in five steps. First, it connects a portable reader to two accessible points and records Hz, period and digits in parallel. Then it finds the calibration cards for the specific sensors. The third step is an independent calculation of three points per sensor: initial, current and close to the highest observed value. Next it compares the sign with the response to a known water level change. Finally it runs the old and new paths in parallel, without switching alarms.
The test result reveals two errors. Six points used a factor from another pressure range. Two more have the subtraction order reversed. Not one error caused a jagged chart. That is why the check "does the trend look natural" is too weak.
Minimum sample protocol
| Field | Illustrative value | Evidence |
|---|---|---|
| period | 353.61 µs | portable reading |
| frequency | 2827.95 Hz | portable reading and conversion |
| digits | 7996.99 | definition from the reader manual |
| zero | 7650.20 digits | accepted initial record |
| difference | 346.79 digits | independent calculation |
| factor | from the unit card | document scan |
| result | according to the factor unit | control sheet |
| positive direction | pressure increase | controlled event or instruction |
The numbers serve only to show the procedure. They are not coefficients for project use.
Error matrix and business cost
| Error | What appears on the chart | Operational consequence | Test that detects it |
|---|---|---|---|
| µs treated as s | scale out of range or zero | no usable result | dimensional check |
| digits treated as Hz | smooth trend with wrong curvature | wrong threshold | three-point recalculation |
| factor from another unit | plausible shape, wrong amplitude | wrong reserve assessment | match with serial card |
| wrong sign | reversed trend | wrong process interpretation | known-direction test |
| silent new zero reading | jump to zero | loss of cumulative movement | reference history |
| temperature omitted | daily or seasonal cycle | false alarms | chart together with temperature |
| rounding too early | steps, loss of small changes | worse analytical sensitivity | compare raw period |
| linear formula instead of polynomial | growing error at the ends | error at high load | points near range limits |
The cost does not end with correcting the formula. If an alarm operated on a wrong value, the state history and reports must be reviewed. If the data supported contractual decisions, the team should document the scope of impact. Automatic recalculation of the entire history without a plan may only replace one invisible error with another. A separate procedure is described in the article on recalculating history after a calibration change.
What to include in the technical specification and SAT
The statement "the system converts vibrating wire sensors" is not verifiable. An acceptance requirement should specify input, output and evidence.
Data and documents before configuration
- [ ] List of sensors with type, range and unit identifier.
- [ ] Calibration cards in a readable version, not rewritten from memory.
- [ ] Definition of the source format: Hz, period, digits or engineering result.
- [ ] Unit and accuracy of measurement time.
- [ ] Initial reading with date, conditions and approval, made according to the explicit procedure for choosing a zero measurement.
- [ ] Sign rule and physical orientation.
- [ ] Required corrections and auxiliary data.
Technical test
- [ ] Recreate the period-Hz relation for at least three samples.
- [ ] Recreate the digits definition from the manual of the correct reader.
- [ ] Check values below, within and above the typical operating range.
- [ ] Compare with an independent field reading.
- [ ] Test the absence of temperature or another required input.
- [ ] Test retransmission of the same record.
- [ ] Test an incorrect unit and an out-of-range value.
Engineering acceptance
- [ ] A second person calculates the sample without help from the configuration author.
- [ ] The result has the expected direction for a known stimulus.
- [ ] The units of all terms match.
- [ ] The alarm threshold is defined in the same quantity as the result.
- [ ] The document states who approved the model and from when it applies.
- [ ] It is defined what remains after a platform change or contract end.
This checklist shortens acceptance. Without it, the post-failure discussion starts with the question whether the number 8000 was frequency, readout digits or already pressure.
How this looks in Inclify
Inclify accepts automated measurements over HTTP/JSON with time saved in UTC. A logger or adapter can send a value close to the source, for example frequency, period or digits, together with temperature. The platform does not impose one digits definition for all manufacturers.
In a project, you can build explicit equations that convert input channels into engineering quantities and use a saved reference. The editor checks syntax, result type and cycles between equations. An unknown identifier may represent an external measurement input, so the availability of the value must be checked with a separate test. The editor also does not confirm that the factor comes from the correct card or that the sign matches the physics of the structure. An engineer must approve that.
Input data can be shown next to the result, and the selected range can be exported to CSV. A project configuration change leaves the values before and after in the log, but the log is not an automatic calibration certificate and does not recreate the full calculation state by itself. For exact device-to-server exchanges, a raw frame log is available for the configured period, by default seven days.
Inclify does not have a universal wizard that recognizes the sensor brand and automatically selects the formula from the certificate. A correct model requires an agreed data contract, sensor documents and a control sample. A broader way of describing the result is presented in the card on the transition from raw value to engineering unit.
When configuration must be stopped
Do not publish the result when the calibration card is missing, the meaning of the input field is unknown, and the sign was set by matching the expected trend. Those are three red flags. In that state you may collect and secure raw data, but you should not call the result a verified engineering value.
Also stop the migration if the old system provides only post-conversion numbers, without configuration versioning. First run a parallel period and independent field readings. Sometimes the historical raw value cannot be reconstructed. Then the limitation must be stated explicitly instead of creating false precision.
USACE in manual EM 1110-2-1908 treats the measurement system as a combination of transducer, acquisition and link, and describes vibrating wire devices as one of the methods for measuring pressure, strain and stress. That is a useful reminder: the result does not belong only to the sensor. It is formed along the entire chain.
FAQ
Are Hz and digits the same value?
No. Hz is a unit of frequency. Digits is a conventional value calculated according to the definition of the readout system. In the GEOKON piezometer manual, digits is the square of frequency divided by 1000, but you should not assume that definition for every manufacturer. Keep the formula description together with the device or software version.
Is it better to store frequency, period or the final result?
The safest approach is to keep a representation as close to the measurement as possible, plus the engineering result with the calculation rule. Frequency or exact period allows the conversion to be checked later. The final result alone is not enough to audit the factor, sign and zero reading. The retention scope must be set in the contract and data design.
Why does a sensor have an individual calibration factor?
The actual response of the transducer depends on its geometry, materials, installation and production process. The manufacturer determines the relation for the unit or a specific batch and records it on the card. A catalogue factor may serve as an approximate type description, but it should not replace the document supplied with the sensor.
Does a linear formula always suffice?
No. Over part of the range the relation may be linear enough, but the manual of a given sensor may provide a polynomial for better fit. The choice criterion comes from the required uncertainty, operating range and manufacturer documentation. Test points near the boundaries as well, because that is where the difference between a linear and a polynomial relation usually grows.
Can a new zero reading fix a wrong factor?
No. Zero sets the reference point, while the factor defines the scale of change. A wrong scale can be temporarily hidden at one point, but the error will grow as the deviation from zero increases. Changing the reference also does not fix a wrong unit, sign or loss of resolution.
Who should approve the conversion?
The integrator is responsible for correct mapping and implementation, and the engineer of the relevant discipline for the physical sense, factors, signs and units. For critical measurements, a second person should reproduce several samples from the source documents. A green application status confirms that the rule runs, not that the metrology is correct.
Sources and further reading
- GEOKON, Model 4500 Series Vibrating Wire Piezometer, Instruction Manual, definition of digits, pressure conversion and period/frequency example.
- GEOKON, Model GK-404 Vibrating Wire Readout, Hz, µs, digits and µε modes and timebase parameters.
- GEOKON, Model 4200 Series Vibrating Wire Strain Gauges, Data Reduction, theoretical factors and batch factor.
- USACE, EM 1110-2-1908: Instrumentation of Embankment Dams and Levees, measurement chain architecture and description of vibrating wire devices.
- JCGM, International Vocabulary of Metrology, measurement model, definition of the mathematical relation between quantities in a measurement.
Do you have the calibration card and a sample frame, but are not sure whether the system calculates Hz, period and digits in the right order? Book a technical sample review. First we will reproduce a few numbers. Only then is it worth building the dashboard and alarms.