A preparation machine displays 70°C. A probe in its reservoir shows a similar value. The first water entering a room-temperature bottle is cooler, and the temperature when the water meets the powder differs again. None of these observations can be replaced by a photograph of the display.
For an OEM buyer, temperature accuracy therefore begins with a defined question: what temperature must be achieved, where, during which part of the operation, and under which intended conditions? Without that definition, two competent teams can obtain different readings and disagree because they have measured different things.
This matters particularly when a product offers both a warm-water setting and a function described as suitable for preparing powdered infant formula. The convenience setting, the preparation condition and the final feeding temperature are not interchangeable requirements. A display marked 40°C or 70°C does not explain which requirement the machine actually fulfils.
The practical task is to connect the intended market guidance and product instructions to a measurable condition, then verify that condition on the actual water path. This article sets out an engineering approach for importers, private-label brands and product teams. The broader architecture and sourcing questions belong in our formula dispenser buyer guide.

Establish the intended preparation method for the destination market
The product team should first identify the formula type and intended use. Powdered formula, liquid concentrate and ready-to-feed formula do not present the same preparation task. A machine described simply as a “formula maker” may leave these distinctions unclear. State what the proposed product is intended to dispense, mix or warm before choosing temperature settings or test points.
For the United Kingdom, NHS guidance on making up formula states that water from a preparation machine should be at least 70°C when mixed with the powder, with cooling before feeding. The relevant engineering observation is therefore not merely a reservoir setpoint. It must address the water and process described by that guidance.
The CDC’s U.S. preparation guidance distinguishes circumstances. It generally directs users to follow the formula container instructions with safe water, and identifies extra precautions for infants younger than two months, born prematurely or with weakened immune systems. These distinctions should remain visible when a product team defines its intended market and use. They should not be flattened into one worldwide machine setting.
The CDC also states that formula does not need to be warmed before feeding. A warm-water convenience feature is consequently a separate design proposition, not proof that a universal feeding temperature must be 40°C. The specification should describe the supported function without converting an optional setting into a health requirement.
Keep a dated copy or reference for the guidance used in the product review and identify who resolves updates. A consumer-care instruction can change independently of the machine design. For example, UK guidance published in November 2026 addresses supply disruptions and travel. Such contexts need their own consideration rather than an automatic extension of an ordinary household test.
Where formula instructions, appliance instructions or a proposed market statement appear inconsistent, resolve the conflict with the appropriate responsible specialists and authoritative guidance. Do not change the wording simply to match the available machine setting. The intended use must be coherent before engineering can demonstrate that the product meets it.
Define the temperature that the result is meant to represent
A temperature measurement needs a clearly identified measurand: the quantity intended to be measured. For this product, that can be the temperature of water at an identified point and time, not an abstract property called “machine accuracy.” Different measurands can all be relevant while answering different questions.
The reservoir temperature can help explain control behaviour. An internal sensor reading can explain the machine’s decision to heat or dispense. The outlet stream can show what leaves the nozzle. Water in a receiving vessel can show the condition after transfer. The temperature at the relevant powder-contact stage concerns the actual preparation sequence. Identify these separately on the test drawing.
Time must be part of the definition. The first portion of water, the middle of a dispense and the last portion may differ. A final mixed temperature can conceal a transient earlier in the operation. Conversely, one brief outlet reading may not describe the condition that the product requirement actually concerns. The measurement plan should capture the relevant event, not whichever event is easiest to photograph.
Specify how the receiving container is prepared and positioned. Its material, mass, initial temperature and geometry can affect the transfer. If the supported use allows a range of bottles, the technical team should determine suitable representative and challenging configurations. A pre-warmed laboratory vessel should not silently stand in for every bottle used in the market.
Draw the water route before adding instruments. Show storage, heating, any mixing or bypass path, the nozzle and the receiving vessel. Mark where the powder is introduced in the supported sequence. This can reveal that the proposed temperature check sits upstream of the event on which the claim depends.

Distinguish an accurate sensor from an accurate delivered result
A sensor may agree closely with a reference at its own location while the delivered water misses the intended condition. The nozzle, unheated tubing, retained water and receiving vessel can create a different thermal history. Calibration of one sensor does not remove those parts of the system.
Sensor location can also affect how quickly a controller recognises a change. A sensor near a heater may respond differently from water moving through the outlet. A sensor mounted against a wall may not follow a changing stream in the same way as a suitable immersed reference. The product team needs to understand what the control variable represents before comparing it with a separate probe.
The metrology vocabulary distinguishes calibration from adjustment. Calibration establishes a relationship between indicated values and reference values under specified conditions. Adjustment changes the instrument or system. A supplier’s statement that the product was “calibrated” should therefore be clarified: what was compared, what was changed if anything, and what verification followed?
Verification of the complete dispensing function is another task. It uses the defined product configuration and operating conditions to determine whether the output requirement is met. A calibration certificate can support that work, but it does not establish the entire product result. The same distinction prevents a factory from treating a component supplier’s sensor tolerance as the tolerance of every completed appliance.
NIST’s traceability guidance emphasises that traceability belongs to a measurement result and involves a documented calibration chain with uncertainty contributions. Merely owning a calibrated instrument is insufficient. For a buyer, the useful question is whether the complete result and its uncertainty are suitable for the decision, including the method used on the product.
Select a reference system that can observe the event
The reference probe should be appropriate to the water path, temperature range and response required. A slow probe may smooth a short transient. A large probe can disturb a small water stream or receiving volume. The measurement arrangement needs to capture the intended event without materially changing the process it is meant to assess.
Follow the probe manufacturer’s requirements for the sensing location and immersion. A long metal stem does not mean that its whole length measures the same way. Insufficient immersion or contact with a vessel wall can produce a result influenced by another temperature. The fixture should locate the sensing element reproducibly without obstructing the supported dispensing operation.
The instrument’s display resolution is not its total measurement uncertainty. A display showing tenths of a degree may still have larger uncertainty from calibration, response, placement, sampling and environmental effects. Choose the system against the decision requirement rather than the number of decimal places on its screen.
Use calibration information relevant to the temperatures and complete channel involved. The probe, readout and logger configuration should be identifiable. If a correction is applied, preserve its basis and use it consistently. An ice-point check can be useful for an appropriate purpose, but agreement near one temperature does not by itself establish performance across the operating range of a formula preparation machine.
Logging should connect temperature with the dispensing sequence. A time record, dispense signal or another suitable event marker can help identify first flow, subsequent flow and completion. The logging interval and response should be selected for the dynamics. A long interval that misses a short cold portion is unsuitable if that portion matters to the stated requirement.
Arrange the work safely with the intact appliance and suitable fixtures. Hot water can create a scald risk, and water should not reach unsuitable electrical connections. Specialist internal measurements or modifications require competent personnel. Purchasing staff can review the protocol and results without improvising an exposed electrical test beside a receiving bottle.
Design the sample programme around actual operating boundaries
Cold start and repeated use can produce different results. In a cold start, the water path and vessel may begin at room temperature. After several dispenses, those components may retain heat. A test performed only after warming the system can miss the first-use condition that customers encounter each day.
Small and large volumes can challenge the design differently. A short dispense can be dominated by retained water or a brief transient; a long dispense can reveal changes in heating capacity or control. The relevant minimum and maximum must come from the supported product range. Do not select arbitrary volumes that the machine is not intended to deliver.
Reservoir level and water starting temperature may also affect the result. The engineering team should select conditions that represent the claimed operating range and any relevant challenging cases. A statement that the machine works “in every climate” is not established by one room-temperature run. The supported boundaries should be explicit and verified appropriately.
The regional electrical configuration belongs in the programme. The heater, control and supply arrangement can differ between models that share an enclosure. Record the exact version, permitted supply range and conditions of evaluation. Evidence from one configuration should not be extended to another without a documented technical basis.
Normal maintenance conditions deserve consideration. Deposits and the supported cleaning or descaling process may influence performance or the water path. Define an appropriate product-specific programme with the responsible team. Do not add an unsupported lifespan, universal descaling interval or improvised chemical recipe simply because maintenance needs a paragraph in the manual.
| Test dimension | Examples to define for the actual product | Why it matters |
|---|---|---|
| Start state | Cold start and defined repeated-use state | The path may retain different amounts of heat |
| Dispense size | Supported small and large volumes | Transients and heating capacity can differ |
| Water condition | Specified start temperature and reservoir level | The control system faces different inputs |
| Receiving vessel | Relevant material, geometry and initial condition | Transfer can change the observed temperature |
| Configuration | Model, firmware, heater and regional supply | Similar enclosures may contain different systems |
| Maintenance state | Defined new and relevant maintained condition | The supported use includes more than one ideal run |
The table is a planning aid, not a universal certification matrix. Select combinations that answer the product’s risks and requirements, and explain any grouping or representative choice. This is more defensible than running every possible combination without a rationale or selecting only the easiest demonstration case.
Preserve the first flow and the whole dispensing sequence
An outlet trace should remain connected to the event it describes. Identify when water first appears, how flow develops, and what happens before dispensing stops. If several temperature locations are recorded, align their times appropriately and account for transport and instrument response. A delayed probe response should not be mistaken for a delayed machine response without investigation.
Do not report only the highest value reached. A peak shows that the system reached that value somewhere in the observed sequence; it does not show that the relevant portion met the required condition. An average has a different limitation: it can conceal both low and high portions. Choose the acceptance quantity before selecting a summary statistic.
A collection method can create its own averaging. Measuring only the final contents of a large receiving vessel may be useful for one requirement while missing the temperature of earlier contact with powder. Conversely, placing a probe in a narrow stream can disturb the stream and may not reproduce the actual preparation. The protocol should explain which event each method can support.
Some architectures use a small hot portion followed by additional water. Such a sequence needs assessment as the actual supported preparation process. A single temperature reading from the hot portion cannot establish the adequacy of the complete sequence or its microbial outcome. The responsible specialists must determine the relevant thermal and other validation, including the interaction with the intended powder and mixing procedure.
This is a central boundary for purchasing. Temperature measurements can substantiate defined thermal conditions. They should not be presented as a complete microbiological validation when no such evaluation exists. A product team should obtain the appropriate evidence for the preparation claim rather than infer it from a well-controlled sensor trace alone.
Read a cold-start trace as a sequence
Consider an illustrative outlet trace that begins near 60°C and rises above 70°C later in the dispense, while a subsequent run begins above 70°C. These invented profiles demonstrate a measurement question, not the behaviour of a particular product. Reporting only the final value would remove the most obvious difference between the runs.
The investigation should preserve how much water was delivered during each relevant part and how the outlet observation relates to the supported preparation event. Time alone does not identify that relationship if flow changes. A product with an intentionally staged sequence may require a different interpretation from a continuous dispense, but the sequence must be described and evaluated explicitly.
The two traces also do not identify the cause by themselves. They give the engineer a reproducible condition to investigate, such as the initial thermal state of the path or a control transition. Further measurements can distinguish competing explanations. Increasing a displayed setpoint without that investigation might leave the transient unchanged or create a different problem later in the run.
The resulting report should state the requirement being checked, which observed part of the sequence addresses it and what remains unresolved. This lets the buyer distinguish a successful corrective action from a second demonstration performed under an easier starting condition. The original trace remains part of the evidence because it explains why the revised programme includes that condition.

Agree how uncertainty affects a boundary decision
A minimum requirement and a target with a symmetric tolerance are different things. If a project requires a defined temperature to be at least 70°C, writing “70°C plus or minus 1°C” permits values below that minimum. The supplier and buyer should not use those two specifications interchangeably. The first step is to state the requirement correctly.
Measurement uncertainty then affects how confidently a result can support a statement of conformity. It should include relevant contributions from the complete measurement, not only a probe’s catalogue accuracy. The responsible laboratory or technical team should evaluate it and agree the decision rule appropriate to the requirement and risk.
For an illustrative decision only, assume a measured result of 70.4°C and an expanded uncertainty of 0.8°C. Suppose the project has agreed a conservative lower-bound acceptance rule for a 70°C minimum. Subtracting 0.8 from 70.4 gives 69.6°C, so that result does not demonstrate acceptance under the stated rule. These are invented inputs, not test results or a universal legal decision rule.
Another agreed decision rule can treat risk differently. The important requirement is to define and document it rather than selecting a rule after seeing a marginal result. ILAC-G8 guidance on decision rules explains guard bands and the relationship between specification limits and acceptance limits. The actual project needs an appropriate application of those principles.
Do not respond to uncertainty merely by increasing a setpoint without further evaluation. A change can affect other parts of the product and preparation sequence, including handling and cooling requirements. The engineering task is to achieve the defined output with suitable control and evidence, while maintaining all other relevant requirements.

Diagnose differences by location and time
Suppose the reservoir and internal sensor agree, but a cold-start outlet trace begins below the required product condition. That pattern suggests investigating the downstream path and the control sequence. It does not prove which component is responsible. Retained water, thermal losses, sensor timing and flow behaviour are possible lines of inquiry that need controlled evaluation.
If the outlet condition appears adequate but the observed receiving condition differs, examine the vessel, transfer, probe arrangement and timing. Check whether the measurements describe the same portion of water. A temperature difference between two locations may be physically real, partly influenced by the measurement method, or both.
If readings differ between laboratories, compare the protocol before debating instrument brands. Review the measurand, probe position, receiving vessel, start state, volume, event timing, calibration corrections and decision rule. Two laboratories can both measure competently yet answer different questions because one measures the nozzle stream and the other measures a final mixed volume.
Reproduce the relevant observation with identified samples and a controlled arrangement. Change one meaningful variable at a time where practical, while preserving suitable operating conditions. A well-described failure is often more useful than repeated successful demonstrations under a different condition. Keep the failed trace because it defines what the corrective action needs to resolve.
After a modification, repeat the affected verification and assess related requirements. A firmware change may alter heating duration or dispensing order. A nozzle or tube change can alter retained volume or heat transfer. A supplier’s statement that “only software changed” is not a reason to assume the delivered thermal behaviour is unchanged.
Keep temperature and water quantity as separate requirements
A correct temperature does not demonstrate correct dosing. The intended preparation also depends on the supported water amount and, where relevant, powder delivery and compatibility. A machine that dispenses the desired temperature but a different quantity has not established that it performs the complete preparation function described to the customer.
Define what the water-volume statement represents and how it is verified. The method should account for the relevant measurement conditions and uncertainty. Where a mass-based method is used, apply an appropriate density relationship rather than assuming that every gram reading is exactly the same number of millilitres under all conditions.
For automatic powder delivery, identify the formula products and configurations that are actually supported. Powder properties and the machine’s dosing arrangement need relevant validation. A temperature test cannot demonstrate universal compatibility with all formulas. Keep the approved compatibility information and any restrictions aligned with the instructions and customer support material.
The buyer should never solve a machine discrepancy by suggesting informal changes to the formula recipe. Resolve the appliance’s performance and instructions through the responsible product process and authoritative guidance. The claimed capability should match the verified preparation function, including its limits, rather than relying on customers to compensate for an unexplained output.
Make the approval record usable during production and service
The approved file should identify the complete model, regional configuration, relevant hardware and firmware versions, supported modes and the measurement programme. Include the actual results and limitations, not only a statement that the machine is accurate. A reviewer should be able to determine which conditions were evaluated and which claim they support.
Distinguish design verification from production checks. A comprehensive development programme can establish behaviour across conditions that cannot all be repeated on every production unit. The manufacturer should define suitable production controls and explain their relationship to the verified design. Do not claim that every unit receives every development test unless that is the actual controlled process.
A production check should still be capable of detecting the relevant variation. The appropriate variables may include sensor response, heater or flow behaviour, assembly differences and software configuration. The engineering and quality teams should choose the checks from the product’s design and risks. A convenient room-temperature display check may not address the failure identified during development.
Maintain a clear change-notification scope. Changes to the heater, sensor, thermal interface, pump, water path, nozzle, firmware or regional electrical configuration may affect the delivered result. The assessment should identify which requirements need review and which evidence can remain applicable. This makes change control proportionate without silently extending old results to a new system.
For service enquiries, collect the model and version, supported mode, volume, starting condition and description of the observed issue. A consumer’s unverified thermometer reading is a useful report to investigate, not automatic proof of either compliance or non-compliance. The service process should provide appropriate product guidance and a competent assessment without asking customers to conduct unsafe experiments.
Write instructions around the verified function
The meaning of a ready signal should be unambiguous. It might refer to a reservoir condition, completion of a heating stage or completion of a supported preparation sequence. The display and instructions should not allow the customer to infer a broader function than the product has established. This is a product-communication issue as well as an engineering one.
Use different terms for water heating, mixing, preparation and optional warming where those functions differ. A single temperature icon can be convenient on a control panel, but the surrounding instructions must explain what it represents. Translation should preserve that distinction across languages rather than compressing every state into “ready to feed.”
Keep the supported formula types, containers, maintenance conditions and operating boundaries accessible. Instructions should describe what the verified product can do and the relevant restrictions. Avoid statements such as “perfect temperature every time” when the actual evidence concerns a defined range, sequence and decision rule.
For an OEM project, connect those consumer statements to the technical requirement matrix. The public instructions do not need the full laboratory procedure, but the company should be able to trace a meaningful claim to its supporting requirement and evidence. This connection helps prevent marketing or translation changes from expanding the product’s intended use without review.
For a VITALBUB development discussion, send the destination markets, intended formula type, water and powder functions, supported volumes and proposed preparation wording. Include any existing prototype traces or supplier specifications. The OEM enquiry contact can then address the specific water path, measurement questions and documentation available for the proposed configuration.
Questions that clarify a temperature specification
Does a 70°C display demonstrate the required temperature at powder contact
No. The display may represent a setpoint or an internal measurement elsewhere in the system. Define the relevant point and event, then verify the delivered process with an appropriate method. A display photograph alone cannot establish the thermal condition at a different location or time.
Is 40°C a universal safe formula preparation setting
No universal preparation requirement follows from that number. A warm-water setting and the conditions for preparing powdered formula are different matters. Define the intended market, formula type and supported use using the relevant authoritative guidance and product instructions before selecting or promoting the setting.
Can a probe certificate settle a disagreement about a machine
It supports the measurement system but does not establish the complete product result. Review what was measured, placement, dynamics, corrections, uncertainty and operating conditions. The disagreement may concern different measurands or methods even when both probes have suitable calibration information.
Should the buyer specify one universal tolerance for all operating states
Only if a properly developed requirement justifies that approach. Different quantities and events may need different acceptance criteria, and a minimum should not be replaced by a symmetric tolerance that permits values below it. Agree the requirements, measurement method and decision rule as connected parts of the product specification.
We’re the VITALBUB Team—engineers, parents, and baby care advocates on a mission to deliver safe, smart appliances that make parenting easier for families worldwide.



