“Field stability: ±0.01%.”
That looks like a specification.
It is not yet an acceptance criterion.
The first question should be:
Over what period of time?
A magnet system that remains within ±0.01% for 30 minutes may behave very differently over an 8-hour working day. A system that passes an 8-hour test may still show overnight drift during a 24-hour experiment.
The same ambiguity appears in magnet power supplies, Helmholtz coils, electromagnets, Hall measurement systems, and closed-loop magnetic field controllers.
For buyers, the important question is therefore not simply:
“Is the system stable?”
It is:
“Stable by which metric, over what time window, after what warm-up period, under what operating and environmental conditions?”
This article explains how to define a practical magnetic field stability test before Factory Acceptance Testing (FAT), final acceptance, or purchase order approval—and when 30 minutes, 8 hours, or 24 hours actually makes sense.
1. There Is No Universal “Correct” Stability Test Duration
A longer test is not automatically a better test.
The test duration should represent the way the equipment will actually be used.
A laboratory running 20-minute Hall measurements does not necessarily need the same acceptance test as a calibration laboratory operating a 3-axis Helmholtz coil continuously overnight.
Similarly, a 24-hour test may add little value if the customer’s normal experiment lasts only 30 minutes.
The correct stability window depends on:
- Normal experiment duration
- Required measurement uncertainty
- Magnetic field level
- Open-loop or closed-loop operation
- Magnet type
- Power supply stability
- Cooling architecture
- Field sensor type
- Laboratory temperature stability
- Whether the system will operate unattended overnight
The purpose of an acceptance test is not to collect the largest possible amount of data.
It is to demonstrate that the system remains within the required performance envelope for the intended experiment.
2. First Separate Warm-Up Time from Stability-Test Time
This distinction is extremely important.
A specification might say:
“Stability shall be tested for 8 hours.”
But does the 8-hour clock start immediately after the equipment is switched on?
Or after thermal equilibrium has been reached?
Those are very different tests.
Warm-Up Is Not the Same as Stability
Immediately after startup, several components may still be changing:
- Power electronics temperature
- Coil temperature
- Magnet yoke temperature
- Cooling-water temperature
- Hall probe temperature
- Mechanical dimensions
- Electronic reference voltage
- Current-sensing electronics
This initial transient may be completely normal.
If a product specification states that its stability performance applies after 30 or 60 minutes of warm-up, then measuring from cold startup tests something different from the stated long-term stability.
A professional acceptance specification should therefore define two separate parameters:
Warm-up period:
How long the system is allowed to reach its normal operating state.
Stability observation period:
How long performance is measured after the defined warm-up condition.
For example:
Warm-up: 60 minutes at the test operating point
Stability test: subsequent continuous 8-hour measurement
That is much clearer than:
Run the system for 9 hours.
3. What Does a 30-Minute Stability Test Actually Prove?
Thirty minutes is useful—but it answers a relatively narrow question.
It can show whether the system has obvious short-term instability after warm-up.
A 30-Minute Test Is Useful For
- Basic FAT screening
- Detecting oscillation
- Detecting poor current regulation
- Identifying rapid thermal drift
- Checking field-feedback behavior
- Comparing short-term repeatability
- Verifying a system intended for short experiments
It is particularly useful when normal measurements themselves are relatively short.
What It May Miss
A 30-minute test may not reveal:
- Slow power-supply reference drift
- Long-term Hall probe drift
- Chiller temperature cycles
- Laboratory HVAC cycles
- Progressive magnet heating
- Multi-hour mechanical movement
- Overnight ambient-temperature change
Therefore, a 30-minute test should not automatically be described as a long-term stability test.
It is better understood as a short-term or post-warm-up stability verification.
4. Why 8 Hours Is a Very Useful Engineering Window
For many research magnet systems, an 8-hour stability test is a strong practical compromise.
It roughly represents a normal laboratory working day and is long enough to expose thermal and environmental effects that a short test may miss.
This is not merely a theoretical convention.
Published magnet-power-supply testing for the ATF2 accelerator, for example, explicitly distinguished a 10-minute short-term stability test from an 8-hour long-term stability test. The reported current data showed ppm-level short- and long-term behavior, while the researchers also recorded temperature during the 8-hour measurement.
That example illustrates an important principle:
Stability has no useful meaning unless time is part of the definition.
An 8-Hour Test Can Reveal
- Slow thermal equilibrium effects
- Power-electronics drift
- Cooling-system cycling
- Laboratory temperature effects
- Gradual field drift
- Sensor offset changes
- Long-duration control-loop behavior
For many laboratory electromagnets, Helmholtz coils, and precision power supplies, 8 hours is therefore a defensible long-term acceptance window.
But it still should not be adopted blindly.
5. When Does a 24-Hour Test Make Sense?
A 24-hour test becomes valuable when the intended experiment crosses day/night operating conditions or routinely runs unattended.
Typical Applications
- Overnight magnetic measurements
- Long-term sensor calibration
- Magnetic aging studies
- Cryogenic experiments
- Automated material characterization
- Continuous field exposure
- Environmental testing
- High-precision reference-field generation
A 24-hour test may reveal effects that are almost invisible over 8 hours.
For example:
- Day/night HVAC changes
- Laboratory ambient-temperature cycles
- Cooling-water temperature changes
- Slow electronic drift
- Long-term Hall sensor offset variation
But 24 Hours Is Not Automatically “More Scientific”
If laboratory temperature changes significantly during the test, a 24-hour result may primarily characterize the room rather than the magnet system.
That may still be useful—but only if that is what the customer wants to test.
The acceptance specification should make this distinction clear.
6. Short-Term Noise and Long-Term Drift Are Different Problems
One of the biggest specification mistakes is combining them into one number.
Imagine two systems.
System A
The measured field fluctuates rapidly by ±0.005%, but its average does not move over eight hours.
System B
The field is extremely quiet over a few seconds, but gradually drifts by 0.05% over eight hours.
Which one is “more stable”?
There is no answer until the experiment is defined.
A fast measurement may prefer System B.
A long calibration experiment may prefer System A.
This is why professional stability specifications often separate:
- Noise
- Short-term stability
- Long-term drift
- Repeatability
Do not force all four into one percentage.
7. Peak-to-Peak Stability Is Not the Same as Drift Rate
Suppose a field is recorded for eight hours.
Several different numbers can be calculated from exactly the same dataset.
Peak-to-Peak Variation
The difference between the highest and lowest values.
This is useful when the requirement is:
“The field shall never leave this tolerance band.”
Deviation from Mean
The largest positive or negative deviation relative to the average field.
This is useful for defining a stability band around a nominal value.
Standard Deviation or RMS Variation
This describes statistical fluctuation.
It can be useful for noise-like behavior but may hide a slow directional drift.
Linear Drift Rate
A fitted slope such as:
ppm/hour
or:
%/hour
This is particularly useful when the field gradually moves in one direction.
These metrics answer different questions.
An acceptance document should state which one applies.
8. Never Write “±50 ppm Stability” Without Explaining the Calculation
A procurement specification such as:
Stability: ≤ ±50 ppm
still leaves multiple interpretations.
Is that:
- ±50 ppm from the initial value?
- ±50 ppm from the mean?
- 50 ppm peak-to-peak?
- 50 ppm RMS?
- 50 ppm per hour?
- 50 ppm over eight hours?
Those are materially different requirements.
Better Specification
For example:
After a 60-minute warm-up, the measured magnetic field shall remain within ±50 ppm of the mean value during an 8-hour continuous test at the specified operating point.
Or:
Long-term drift shall not exceed 10 ppm/hour, determined from the linear fit of the measured field during the 8-hour observation period.
Now the supplier knows what must actually pass.
9. Current Stability Is Not Automatically Magnetic Field Stability
This distinction matters especially in procurement.
A magnet power supply manufacturer may specify excellent output-current stability.
That does not automatically prove that the magnetic field at the sample is equally stable.
The complete chain is:
Current source → magnet → magnetic circuit → environment → field sensor → measurement system
Any part can influence the recorded result.
In an Air-Core Helmholtz Coil
Field is primarily determined by:
- Coil geometry
- Number of turns
- Current
- Position
With a well-regulated constant-current supply, coil resistance increasing with temperature does not directly force the current to decrease—as it would in a simple voltage-driven circuit.
However, temperature can still matter through:
- Current-source drift
- Mechanical thermal expansion
- Sensor temperature sensitivity
- Changes in surrounding magnetic conditions
In an Electromagnet
Additional effects may include:
- Magnetic hysteresis
- Pole and yoke temperature
- Magnetic material behavior
- Pole-gap geometry
- Residual magnetization
So a power-supply stability certificate and a system field-stability test are not interchangeable.
10. Hall Probe Stability Can Become the Measurement Limit
There is another problem:
How do you know whether the magnet drifted—or the field sensor drifted?
Hall probes are extremely useful for magnetic field measurement, but precision measurements must consider the long-term behavior of the measurement chain itself.
Research associated with CERN accelerator magnetic-field measurement explicitly notes that the long-term stability of Hall probe offset and gain can become problematic in demanding applications and may require recalibration.
This creates a fundamental acceptance problem.
If a 24-hour field test shows 100 ppm of change, the observed result could contain contributions from:
- Magnet
- Power supply
- Hall probe
- Hall electronics
- Probe temperature
- Probe position
- Ambient field
A stability test therefore needs an adequate reference measurement system.
Otherwise, the test may be measuring the tester.
11. Record Temperature During Every Serious Stability Test
A field-versus-time graph is useful.
A field-versus-time graph together with temperature data is much more useful.
At minimum, consider recording:
- Laboratory ambient temperature
- Coil or magnet temperature
- Cooling-water inlet temperature
- Power-supply temperature, if available
- Hall probe temperature, if available
The ATF2 magnet-power-supply study mentioned earlier recorded temperature together with current during its 8-hour long-term stability test and related the observed drift to the temperature change.
That is good engineering practice.
Without temperature data, a slow drift may be visible but difficult to explain.
12. Cooling Conditions Must Be Frozen for Acceptance
A water-cooled electromagnet may behave differently depending on:
- Coolant flow
- Inlet temperature
- Chiller control mode
- Chiller hysteresis
- Ambient temperature
Therefore, the acceptance test should define the cooling condition.
For example:
Chiller setpoint: 22 °C
Required flow: according to manufacturer specification
Test performed after coolant temperature has stabilized
For an air-cooled magnet or power supply, laboratory airflow and ambient temperature may instead be relevant.
The objective is reproducibility.
The supplier and buyer should be able to repeat essentially the same test under comparable conditions.
13. Field Setpoint Must Be Specified
A system may not have identical stability at every operating point.
Testing only at 5% of rated current may not tell the buyer much about operation near maximum field.
Similarly, testing only at maximum current may not represent the low-field conditions that matter most in calibration.
Possible Test Points
Depending on the application:
- Low field
- Mid-range field
- Maximum normal operating field
- Both positive and negative polarity
- Zero or near-zero field
The correct number of points depends on the project.
A single stability test does not necessarily characterize the entire operating range.
14. Electromagnet Pole Gap Must Be Recorded
For electromagnets, stability data without pole-gap information is incomplete.
The test record should identify:
- Pole gap
- Pole geometry
- Field measurement position
- Probe orientation
- Magnet current
A result such as:
0.8 T stable within 0.01%
becomes far more reproducible when written as:
0.8 T at 25 mm pole gap, measured at the magnetic center, after defined warm-up and under specified cooling conditions.
This level of detail matters during FAT and later site verification.
15. Helmholtz Coil Tests Should Define the Axis
For a 3-axis Helmholtz coil, “field stability” raises another question:
Which field?
Possible tests include:
- X axis only
- Y axis only
- Z axis only
- All three axes independently
- Combined vector field
The power electronics, winding resistance, geometry, and thermal environment may differ slightly among axes.
For precision calibration systems, it can be useful to test each axis separately.
If simultaneous vector-field generation is part of normal operation, at least one representative multi-axis condition may also deserve verification.
16. Do Not Confuse Stability with Repeatability
A system can be stable and still have poor repeatability.
Suppose an electromagnet remains extremely stable for eight hours after reaching 500 mT.
Then you return to zero and set 500 mT again tomorrow.
If the field becomes 501 mT, the system passed the stability test but may have failed a repeatability requirement.
Stability Asks
Does the output remain constant while the operating condition is held?
Repeatability Asks
Does the system return to the same output when the same condition is reproduced?
For electromagnets, magnetic hysteresis makes this distinction particularly important.
A proper acceptance plan may therefore require both tests.
17. Define the Magnetic Approach History for Electromagnets
The measured field of an iron-core electromagnet can depend on how the setpoint was reached.
For example:
- Ramp upward from zero
- Ramp downward from a higher field
- Reverse polarity
- Perform a defined magnetization cycle first
These paths may not give identical results because of hysteresis.
Therefore, if a high-precision stability measurement begins at 1 T, the test procedure should specify how 1 T is reached.
Example
Before the stability observation begins, the magnet shall follow the defined preconditioning sequence and approach the final field setpoint from the same direction.
Otherwise, two tests may begin from slightly different magnetic states.
18. 30 Minutes, 8 Hours, or 24 Hours? A Practical Decision Rule
There is no universal industry rule, but a useful procurement framework is:
Choose Approximately 30 Minutes When
- Normal measurement runs are short
- The objective is FAT screening
- Short-term regulation is the main concern
- The system is standard rather than highly customized
- Longer-term stability is not critical to the experiment
Choose Approximately 8 Hours When
- Experiments typically run for much of a working day
- Long-term calibration stability matters
- You need to observe thermal equilibrium and drift
- The system will operate continuously for several hours
- The equipment is a precision power supply, electromagnet, or Helmholtz calibration system
Choose 24 Hours or Longer When
- The actual experiment is overnight or multi-day
- The system acts as a reference field source
- Environmental day/night cycles matter
- Very slow drift is scientifically important
- Unattended continuous operation is part of the requirement
The key rule is:
The stability test should cover the timescale on which unacceptable drift would damage the experiment.
19. Should Every Customer Demand a 24-Hour FAT?
No.
That can easily become specification inflation.
A longer FAT consumes:
- Factory time
- Test-equipment time
- Engineering supervision
- Data processing
- Production capacity
If a buyer’s experiment lasts 20 minutes, demanding 24-hour stability without a technical reason may increase project cost while adding little useful risk reduction.
Procurement specifications should distinguish between:
scientifically necessary
and:
nice to have
The strongest technical specification is not the one with the largest number of requirements.
It is the one whose requirements correspond directly to the experiment.
20. Sampling Interval Matters Too
Suppose two laboratories both perform an 8-hour test.
Laboratory A records one value every hour.
Laboratory B records one value every second.
These are not equivalent datasets.
Too Slow
Important transients may be missed.
Too Fast
A long-term stability dataset can become dominated by high-frequency noise that should really be characterized separately as ripple or noise.
For many stability tests, sampling every few seconds or tens of seconds may provide adequate trend information, depending on system dynamics.
But there is no universal optimum.
If high-frequency ripple matters, measure it separately using instrumentation and bandwidth appropriate to that requirement.
Do not expect an eight-hour trend plot to replace a ripple measurement.
21. Do Not Hide Warm-Up Drift by Choosing the Wrong Start Point
There are two legitimate ways to test a magnet system.
Test A — From Cold Start
Question:
How long does the system take to become usable?
This is a warm-up or settling test.
Test B — After Stabilization
Question:
Once ready, how stable is it?
This is the conventional stability test.
Both are useful.
They should simply not be confused.
For laboratories that need immediate operation after startup, cold-start behavior may itself be a procurement requirement.
For laboratories that routinely warm equipment for one hour before precision measurements, stabilized performance is more relevant.
22. A Closed-Loop System Requires Two Stability Questions
Closed-loop magnetic field control can compensate for slow drift by measuring the actual field and adjusting magnet current.
That improves one type of stability.
But it also creates a second measurement chain.
The acceptance plan should ask:
Control Stability
How closely does the measured field stay at the commanded field?
Reference Stability
How stable and accurate is the field sensor used by the feedback system?
A closed loop cannot correct drift that it cannot distinguish from sensor drift.
If the Hall probe slowly changes sensitivity, the controller may alter magnet current to compensate for an error that exists in the sensor rather than in the field.
For demanding calibration systems, sensor specification and calibration strategy therefore belong inside the stability discussion.
23. Field Stability and Current Stability Should Sometimes Be Tested Separately
For a complete system, it may be useful to record both.
Power-Supply Test
Measure output current using an appropriate precision current reference.
This isolates the performance of the electrical source.
System Test
Measure magnetic field at the defined measurement point.
This evaluates the full system.
If current remains extremely stable but field drifts, attention moves toward:
- Magnet temperature
- Mechanics
- Sensor
- Environment
If both current and field drift together, the power supply becomes a stronger suspect.
This diagnostic separation makes FAT data far more useful than a single field trace.
24. Acceptance Limits Should Match Measurement Uncertainty
There is little value in specifying:
Field stability ≤10 ppm
if the measurement system cannot reliably distinguish changes at that level.
The acceptance method must have sufficient:
- Resolution
- Repeatability
- Stability
- Calibration quality
- Environmental control
Ideally, measurement uncertainty should be comfortably smaller than the acceptance tolerance.
Otherwise, a pass/fail decision may depend more on the test equipment than on the product.
This is especially important when customers move from percentage-level requirements into tens of ppm.
25. Do Not Use the First Data Point as a Perfect Reference
Another common method is:
Take the first field reading and calculate all later deviation from that value.
This is simple, but it makes the entire result dependent on one measurement.
That first point may contain:
- Sensor noise
- Small transient
- Data-acquisition error
- Temperature settling
Depending on the acceptance objective, it may be better to evaluate:
- Mean value
- Defined baseline average
- Maximum-minimum range
- Fitted drift slope
The calculation method should be chosen before the data is collected.
Do not select the statistic afterward because it makes the result look better.
26. A Good Stability Report Needs More Than One Graph
A professional stability-test report should make the operating conditions reproducible.
Useful recorded information includes:
- Equipment model and serial number
- Magnet or coil configuration
- Power supply
- Field sensor and serial number
- Calibration status
- Field/current setpoint
- Electromagnet pole gap, if applicable
- Helmholtz coil axis, if applicable
- Warm-up duration
- Test start and end time
- Sampling interval
- Cooling conditions
- Ambient temperature
- Magnet or coil temperature, if available
- Raw field or current data
- Mean
- Maximum
- Minimum
- Peak-to-peak variation
- Drift slope, if applicable
- Acceptance limit
- Pass/fail conclusion
This transforms a screenshot into traceable acceptance evidence.
27. Example: A 30-Minute FAT Requirement
The following is an illustrative format, not a universal performance recommendation:
Operating point: 100 mT
Warm-up: 30 minutes at target field
Observation period: 30 minutes
Measurement position: magnetic center
Sampling interval: 5 seconds
Metric: maximum deviation from mean
Acceptance criterion: as agreed in final technical specification
Environmental record: ambient temperature and cooling condition recorded
Notice what is missing:
There is no arbitrary stability percentage.
That number should come from the actual experiment and agreed product performance.
28. Example: An 8-Hour Helmholtz Coil Acceptance Requirement
Another illustrative structure:
System: 3-axis Helmholtz coil
Test axis: X
Operating point: agreed calibration field
Warm-up: 60 minutes
Observation period: 8 continuous hours
DUT position: center of specified homogeneous volume
Measurement: independently calibrated field sensor
Recorded parameters: field, coil current, ambient temperature, coil temperature if available
Metrics: peak-to-peak variation and fitted long-term drift
Acceptance criteria: values defined in the approved technical specification
The same test can then be repeated for Y and Z if required.
29. Example: A 24-Hour Electromagnet Test
For a system intended for overnight operation:
Magnet condition: defined pole gap and magnetic preconditioning sequence
Target field: representative normal operating field
Warm-up: until defined thermal condition or specified warm-up time is achieved
Observation period: 24 hours
Cooling: fixed chiller settings and recorded flow/temperature
Data: magnet current, field, ambient temperature, cooling temperature
Metrics: maximum deviation, drift slope, and any interruption events
Acceptance: according to agreed system-level stability requirement
A 24-hour test should also document unusual events.
If the laboratory air conditioning was turned off for four hours, that information matters.
30. What Buyers Should Put into the PO or Technical Specification
Do not write:
Field stability: good.
Do not write:
Field stability: 0.01%.
And preferably do not stop at:
Field stability: ±0.01% over 8 hours.
A much stronger requirement is:
After a defined warm-up period, at the specified magnetic field, pole gap or coil configuration, and under specified cooling and environmental conditions, the magnetic field shall remain within the agreed stability tolerance during an 8-hour continuous measurement at the defined measurement position. The calculation method and measurement instrumentation shall be stated in the FAT protocol.
Now the requirement can actually be tested.
31. How Cryomagtech Approaches Magnet Stability Requirements
Cryomagtech supplies magnetic field systems and related instrumentation including:
- Electromagnet systems
- Helmholtz coil systems
- 3-axis Helmholtz coil systems
- Precision magnet power supplies
- Magnetic field measurement solutions
- Hall-based field measurement configurations
- Customized field-control systems
👉 Product link placeholder: Cryomagtech Electromagnet, Helmholtz Coil & Magnet Power Supply Systems
For projects where stability is an important acceptance parameter, we recommend defining the measurement conditions before final technical confirmation.
That discussion may include:
- Required field
- Operating duration
- Warm-up time
- Continuous-duty requirement
- Open-loop or closed-loop control
- Stability metric
- Drift limit
- Field sensor
- Cooling conditions
- Environmental conditions
- FAT duration
This allows a phrase such as “high stability” to become an engineering requirement that both buyer and supplier can verify.
32. The Right Question Is Not “30 Minutes or 24 Hours?”
The better question is:
“What timescale must remain stable for my experiment to remain valid?”
If the experiment lasts 20 minutes, a well-designed 30-minute acceptance test may be meaningful.
If the experiment occupies a full working day, 8 hours becomes much more relevant.
If the laboratory runs unattended overnight, a 24-hour test may be justified.
But duration alone does not define a good stability test.
You still need:
- A defined warm-up period
- A defined operating point
- A defined measurement location
- A defined metric
- A defined environment
- A sufficiently stable measurement reference
- A numerical pass/fail criterion
Only then does magnetic field stability become a specification rather than a marketing phrase.
References
JACoW — ILC / ATF2 DC-Magnet Power Supplies
This accelerator magnet-power-supply study explicitly distinguishes short-term stability over 10 minutes from long-term stability over 8 hours. It also records temperature during the long-duration test, illustrating why observation time and environmental conditions belong in a meaningful stability definition.
CERN-Related Research — Development of a Real-Time Magnetic Field Measurement System for Synchrotron Control
This study discusses magnetic-field measurement for accelerator control and notes that long-term offset and gain stability of Hall probes can become problematic in high-precision applications. This supports an important acceptance principle: the measurement sensor itself must be considered when interpreting long-term field drift.
Key Takeaways
- Stability must always be defined together with a time interval.
- Warm-up time and stability-test duration are different specifications.
- A 30-minute test is useful for short-term performance and FAT screening.
- An 8-hour test is often a strong practical choice for working-day and long-term laboratory operation.
- A 24-hour test is most useful when the real experiment runs overnight or when very slow drift matters.
- Longer testing is not automatically better.
- Current stability and magnetic field stability are not identical.
- Hall probe drift can limit the apparent stability of the complete measurement.
- Peak-to-peak variation, standard deviation, and drift rate describe different behaviors.
- Electromagnet stability tests should define pole gap and magnetic history.
- 3-axis Helmholtz coil tests should identify which axis or vector condition is being tested.
- Ambient, cooling, magnet, and sensor temperatures should be recorded during demanding tests.
- Stability and repeatability should be specified separately.
- Measurement uncertainty must be adequate for the requested acceptance tolerance.
- A good FAT protocol defines the calculation method before the test begins.