Blog · November 9, 2026

Guaranteed Operating Window for PU Metering Equipment

A guaranteed operating window is a written set of material, ratio, output and utility conditions under which a polyurethane metering system will hold a stated accuracy, plus the method that proves it. It is not the accuracy figure on the brochure. Two machines can both print ±0.5% and behave completely differently on the same line, because one number describes a pump at a single nominal point while the window describes what happens at the edges. Most commissioning disputes sit in that gap, and the gap closes on paper before the purchase order, not on the shop floor afterwards.

What a Guaranteed Operating Window Actually Covers

The window has five axes, and a loose value on any one of them can invalidate the other four.

Material comes first: polymer family and grade, filler type and loading, viscosity at the pump inlet, and the temperature at which that viscosity was measured. Ratio comes second: the A:B range by weight, the accuracy limit at each end of that range, and whether the figure applies across the whole range or only at one nominal mix. Output comes third: the continuous flow per component and the turndown between the lowest stable flow and the highest continuous flow. Accuracy definition comes fourth: what is measured, over how many shots, with which instrument, and whether the number is a mean or a worst case. The utility envelope comes fifth: inlet pressure band, discharge pressure, material temperature control tolerance, ambient range and the services the machine needs.

The word guaranteed carries the weight. In a usable specification it means a written commitment with a method attached, a defined sample, a pass limit and a stated consequence when the limit is missed. Without those four elements, the window is a description rather than a guarantee.

Maximum ratings and continuous ratings are different quantities and belong on separate lines of a data sheet. A machine rated for a large peak shot may hold ratio only in the upper half of its flow range, and that difference stays invisible until a low-output product enters the schedule.

a11

Why Two Machines With the Same Accuracy Claim Behave Differently

Metering pumps move volume. The process cares about mass. Polyol density falls as temperature rises, so a pump holding a fixed volumetric ratio across a temperature swing delivers a different mass ratio at the two ends of that swing. Nothing inside the machine has changed, and the parts produced are not the same.

The second mechanism is slip. In a gear pump, some material passes back across the gears instead of leaving through the discharge, and the proportion lost grows as discharge pressure rises and pump speed falls. At the top of the flow range that loss is a small share of total output. At the bottom it can be a large share of a small number, which is why ratio accuracy is hardest to hold at minimum output and why one accuracy figure rarely describes a whole turndown.

The third mechanism is the test definition itself. An accuracy claim measured on unfilled polyol at controlled temperature, at mid-range output, over a short run, is a real measurement of a narrow condition. It says little about filled material, cold material, or the first shots after a colour change.

Metering principle decides how much of this behaviour you inherit. <Polyurethane Foam Machines: Types, Principles & Applications> covers how the main machine types differ in working principle and where each is applied.

Haifeng Polyurethane Machinery’s LJJ series, for example, states a metering accuracy of ±0.5% on its metering pumps and flow control system. That figure is a legitimate starting point for a specification, and it becomes a guaranteed operating window only when the material, ratio, output and temperature conditions around it are written down and tested.

The Six Conditions Your Operating Window Has to State

Condition Write this into the specification What fails when it stays vague How it is verified
Material state Polymer family and grade, filler type and loading, viscosity band with the temperature it refers to, moisture limit Pump fill drops at low output with a colder or more heavily filled batch, so ratio error appears in one product only Corner test with the production grade at the lowest material temperature
Ratio range A:B by weight, minimum and maximum ratio, accuracy limit at each end Ratio holds near the middle of the range and drifts at the ratio that uses the smallest component flow Mass-ratio measurement at minimum, nominal and maximum ratio
Output and turndown Continuous flow per component in g/min, lowest stable flow, highest continuous flow, peak shot stated separately A machine rated on its peak shot cannot hold ratio on a low-output product Timed shots at minimum, middle and maximum continuous output
Accuracy definition What is measured, instrument, sample count, and whether the figure is a mean or a maximum single-shot deviation A mean figure passes acceptance while single-shot excursions generate scrap Every shot recorded and reported, not only the average
Utility envelope Inlet pressure band, discharge pressure limit, material temperature tolerance, ambient range, cooling water and air demand Inlet pressure sags during a long draw, the pump cavitates, and output falls with the ratio Test at the lowest permitted inlet pressure and the highest discharge pressure
Change control Notification threshold for material or component changes, re-validation trigger, document set delivered A resin or filler change invalidates a passing factory test with nobody noticing Written clause plus a documented change record

a5

How to Test at the Corners Instead of the Middle

Corner-point acceptance means testing combinations of extremes rather than the comfortable middle. Four corners cover most of the risk: minimum output with the highest-viscosity production batch at the lowest material temperature, minimum output at the highest ratio, maximum continuous output at the lowest ratio, and maximum output at the highest material temperature.

The sequence is the same at every corner.

  1. Warm up with a fixed number of shots and discard the output, so the first measurement does not include a cold mixing head or an unprimed line.
  2. Fix the material temperature at the corner value and record it at the pump inlet rather than at the tank gauge.
  3. Collect each component separately over a defined number of shots, or use a calibrated gravimetric method that reports both components.
  4. Weigh, compute the mass ratio per shot, and keep every shot in the record. A mean hides the single-shot excursion that creates scrap.
  5. Convert total output into component flows and check that each pump sits inside its stated range. For total output Q at ratio A:B = a:b, the A-side flow is Q × a/(a+b) and the B-side flow is Q × b/(a+b). At 120 g/s and 100:40 that is 85.7 g/s of A and 34.3 g/s of B. The arithmetic is trivial; the point is that a comfortable total output can still push the small component below its stable minimum.
  6. Repeat the first corner at the end of the run, when the machine is at operating temperature, to separate start-up behaviour from steady-state behaviour.

Repeatability and reproducibility are different measurements, and ISO 5725-2 defines both terms precisely [1]. A machine can be highly repeatable shot to shot within one afternoon and still fail to reproduce its ratio the following week when the material batch or the ambient temperature changes. The acceptance document should name which of the two it is checking.

a10

What an Acceptance Test Record Proves and What It Does Not

Record Method context it belongs to What it can legitimately support What must not be inferred
Factory acceptance record at nominal output One material, one ratio, controlled temperature, short run The pump and control loop hold ratio at that point Performance with filled material or at minimum output
Published accuracy figure, for example ±0.5% The manufacturer’s defined conditions, usually mid-range output and a single material A pump-level capability claim under those conditions The same figure across the full turndown or across polymer families
Material TDS viscosity value One temperature and one shear condition A reference point for the viscosity band Viscosity at the pump inlet after a day in an unheated tank
Calibration certificate for the balance or flow meter Traceability of the instrument [2] The reading is traceable to a reference standard That sampling, shot collection or operator technique was correct

These records belong to the parties that produced them. They are evidence about specific conditions, not a chemistry-wide or machine-wide guarantee, and a record copied from a similar project transfers only as far as the material and utilities match.

a19

Turning the Window Into Acceptance and Change-Control Language

An operating window becomes enforceable when it appears in the purchase contract as measurable acceptance criteria rather than a capability statement.

Reject this line: “Metering accuracy ±0.5%.” Replace it with a statement that names material, ratio, output, temperature and sampling: “At 100:40 by weight, with production grade X conditioned to the agreed temperature at the pump inlet, the mass ratio of every shot between minimum and maximum continuous output shall stay within the agreed limit across the agreed number of consecutive shots, measured by the agreed method.” The tolerances inside that sentence are yours to negotiate. The structure is what makes them testable.

Hold points that belong in the contract

Three hold points cover most of the risk. Factory acceptance runs at nominal output and at the corners, with the material you ship. Site acceptance repeats the corner tests after installation, because inlet pressure, ambient temperature and utility quality at your plant are part of the window. Production verification after the first agreed run length confirms that the machine still holds the limit once real cycle patterns, interruptions and colour changes are in the schedule. Each hold point needs a named witness, a record format and a pass rule written before the test starts.

Change control that survives a resin switch

A material change is the most common way a passing window becomes invalid. Resin lots vary in viscosity and reactivity, fillers vary in loading and particle size, and neither change announces itself to the machine. ISO 9001:2015 requires planned review and control of changes affecting production [3], and the same discipline belongs in the equipment contract: a notification threshold, a re-validation trigger and a defined document set. Proving the window afterwards depends on instrument traceability, which ISO/IEC 17025:2017 addresses through requirements for calibration laboratories and metrological traceability of results [2].

a12

How a Guaranteed Window Fails After Commissioning

Windows fail slowly, and the failure usually appears at minimum output first.

Pump wear is the clearest case. Internal clearances open, back-flow across the gears increases, and the loss is proportionally larger at low speed than at high speed. The machine still reaches maximum output, so production notices nothing until a low-output product starts running out of ratio.

Inlet restriction behaves the same way on a shorter timescale. A loaded filter, a partly closed valve or a suction line too small for the viscosity drops the pressure at the pump inlet and the pump begins to cavitate. Output becomes uneven, ratio follows it, and the fault looks like a control problem rather than a plumbing one.

Temperature control drifts in its own way. A fouled heat exchanger or a sluggish thermoregulator moves material temperature by a few degrees, which changes viscosity at the pump and density at the mixing head at the same time. The two effects do not cancel; they compound in whichever direction the ratio is already leaning.

The practical countermeasure is a short corner test on a fixed schedule, run at minimum output with production material, with ratio deviation logged and trended rather than compared against a single limit. Trend direction tells you about wear; the absolute value tells you about the current batch. Keep inlet pressure, material temperature at the head and shot mass in the same record, so a deviation can be traced to one of the three instead of being argued about.

Maintenance schedules are part of the same envelope. <Polyurethane Foaming Machine: How It Works, Components & Maintenance Tips> covers the component set and the maintenance tasks that keep metering behaviour stable between acceptance tests.

Tell Us the Conditions You Need Guaranteed

Send the material family and grade, the filler type and loading, the viscosity band with the temperature it refers to, the A:B range by weight, the continuous output per component, the ambient and material temperature range, the inlet and discharge pressure limits, and the product tolerance the window has to protect. Haifeng Polyurethane Machinery will confirm which of those conditions can be guaranteed on a specific configuration and which need a trial before they enter a contract. Email info@haifeng-automation.cn or message WhatsApp 86 13566296633.

FAQ

How wide should a guaranteed operating window be?

Wide enough to cover the products you actually schedule, and no wider. A window spanning every material the machine could theoretically run is difficult to test and difficult to enforce, because each corner needs its own trial. Start from the production range: the lowest and highest viscosity you buy, the ratio extremes across your product list, and the output band between your slowest and fastest cycle. Anything outside that range is a new validation rather than a continuation of the original window.

Can a supplier guarantee ratio accuracy at minimum output?

Only when the minimum output is stated and tested with the production material. Ratio error grows at low flow because pump slip becomes a large share of a small delivery, and the effect is stronger with higher viscosity and higher discharge pressure. A supplier quoting one accuracy figure without an output range is describing a nominal point. Ask for the minimum stable flow, the material conditions attached to it, and a test record at that flow. Without the record, the minimum stays unproven.

Should acceptance testing use our production material?

Yes, at least for the corner tests. Test fluid shows that the machine runs and the control loop responds, but viscosity, filler content and deaeration behaviour decide whether the pumps fill properly at low output. Ship a production batch to the factory test, or plan a site acceptance test with your own material before the line is committed to a customer schedule. Where material is hazardous or hard to ship, agree a substitute with similar viscosity and filler loading and record the difference.

How many shots should an acceptance test cover?

Enough to show the distribution rather than a stable average. A short run hides the occasional excursion that creates scrap, so the count should follow from the rejection rule: if one shot outside the limit fails the test, the sample has to be large enough for that rule to mean something. A run spanning start-up, steady state and a material or colour change tells you more than a fixed count at one condition. Put the agreed number in the contract rather than a general recommendation.

References

[1] ISO 5725-2:2019, “Accuracy (trueness and precision) of measurement methods and results — Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method,” ISO, 2019.

[2] ISO/IEC 17025:2017, “General requirements for the competence of testing and calibration laboratories,” ISO/IEC, 2017.

[3] ISO 9001:2015, “Quality management systems — Requirements,” ISO, 2015.

If you’re interested, check out these related articles:

How Does a Double Density PU Shoe Machine Work
Cost of Setting Up a PU Safety Shoe Factory
What Is a Polyurethane DIP Machine for Safety Shoes
Rain Boot Factory Investment Cost
How to Reduce Defects in PU Shoe Production

Related Posts