How to Choose a Vision Measurement System: 8 Specs That Actually Matter

Quick Summary

Compare measurement architecture, stated error formula, field of view, travel and load, repeatability, lighting, throughput and software scope. Use your drawing and representative parts to agree acceptance criteria before choosing a configuration.

Reading guide and reference links updated: 10 November 2026.

Examples and selection ranges are indicative, not acceptance guarantees. Confirm model-specific specifications, sensor access, fixturing, environmental conditions and measurement uncertainty. Published guidance does not certify an individual machine.

Check sources and model specifications·Discuss your application

Most buyers start a vision measurement system search the same way: they open three datasheets, find the number with “μm” after it, and pick the smallest one. That number is usually the least useful figure on the page.

Here is the problem. A quoted “±2 μm” is almost always a best-case, short-range, single-point figure — not the error you will see across a 200 mm part, on your material, with your operator, on a Tuesday afternoon. What actually decides whether a vision measuring machine will hold your tolerances, hit your takt time, and still be in spec five years from now is a different set of numbers — and the most important one of them isn’t printed on any datasheet at all. If you are working out how to choose a vision measurement system, that is where to start.

This vision measurement system buying guide walks through the eight specifications that do the deciding, in the order you should decide them. It ends with a model-matching table and a copy-paste RFQ checklist you can send to any supplier — including us.

Start With Your Part, Not the Brochure

Before you compare a single VMM specification, answer four questions. Every spec decision downstream is derived from these, and getting them wrong is the most common reason a machine gets bought twice.

  1. What is the tightest tolerance you must verify?​ Not the typical one — the worst one on the drawing.
  2. What is the material and surface finish?​ Opaque and matte, polished, transparent, or soft and easily deformed? This decides lighting, and lighting decides whether you get an edge at all.
  3. Are the critical features flat 2D geometry, or do you need true 3D form?​ Height, depth, flatness, and coplanarity pull you toward a different architecture than a pure outline.
  4. How many parts per day, and how many part numbers?​ Volume drives automation; part-number count drives changeover time — and changeover, not stage speed, is what kills most high-mix cells.

Write these down. You will use all four in the eight specs below.

Spec 1 — Measurement Architecture: Traveling VMM or One-Shot Flash?

This is the decision that everything else hangs on, and it appears on no specification sheet, because it isn’t a number. It’s a choice between two fundamentally different ways of measuring.

A traveling video measuring machine​ moves a lens or a stage from feature to feature, capturing an edge at each position. Its accuracy is expressed as a length-dependent formula: the error grows with the distance travelled.

A one-shot flash measuring instrument​ images the entire part — or an entire batch of parts — in a single exposure through a wide-field dual-telecentric lens, then lets software find every feature in that one image.

The consequences are not subtle. Compare two real specifications:

  • A high-precision traveling VMM rated1.2 + L/200 μm​ delivers 2.2 μm at 200 mm — and 1.45 μm at 50 mm. The number depends on how big your part is.
  • A flash instrument rated2.0 μm in an 82 × 55 mm field​ delivers 2.0 μm whether you measure 5 dimensions or 100, and whether the part is 5 mm or 50 mm. The length term only appears when you stitch multiple fields together (on a mobile-platform flash unit, stitching is 3.5 + L/200 μm).

That is the whole argument for flash measurement, and it is a metrological one, not a marketing one:if your part fits inside one field of view, you remove the mechanical travel, the fixturing, the refocusing, and the operator technique from your error budget all at once.

The trade-off is real, and you should know it before you buy: flash instruments are essentially 2D. A dedicated flash unit like our K100 has no Z measurement axis; even the dual-vision K433 carries its Z capability on the zoom-lens channel (EUZ 5 + L/200 μm), not the wide-field channel. If your program genuinely needs height, depth, or flatness, say so now and go to a traveling machine.

How to specify it:

Your situationRight architecture
Part fits one field, many dimensions per part, high volume, frequent changeoversOne-shot flash
Part larger than one field, but still mostly 2DMobile-platform flash (stitched)
Mixed: needs both fast large-area scanning and micron-level detailDual-vision (wide field + zoom on one machine)
True 3D form, deep features, hidden geometryTraveling VMM, optionally multisensor
Very large panels, shafts, or parts over ~1 mGantry or horizontal VMM

How we map them at NAVINERTIA:​ K100 for small precision parts in batches; K221 / K321 when the part outgrows a single field;K433 for PCB, double-sided stamped parts, and connectors; the S, E, and G Series when you need 3D or large format.

Spec 2 — Accuracy: Demand the Formula, Not the Number

Every credible vision measuring machine quotes accuracy in the formMPE = a + L/K μm, whereais the fixed error of the axis andLis the measured length in millimetres. Insist on it, and insist that it is stated perISO 10360-7​ and tied to 20 °C. A bare “±2 μm” describes a best case, not a machine.

Then ask for all three axes, because they are not the same number. On the NAVINERTIA S Series:

SpecV323S / V433SV543SWhy you care
EUX/EUY, MPE1.2 + L/200 μm1.5 + L/200 μmLength measurement in X and Y
EUXY, MPE1.5 + L/200 μm1.8 + L/200 μmDiagonal / 2D features — always worse than single-axis
EUZ, MPE2.8 + L/150 μm2.8 + L/150 μmHeight — typically the weakest axis, and the one most buyers forget to ask for

Notice the Z axis. At 100 mm, an S Series machine is at 1.7 μm in XY but 3.5 μm in Z. If your drawing has a height callout, the XY number you were quoted is irrelevant to it.

There is a fourth figure that almost nobody requests and that often matters more:probing error — P2D, PV, and PZ. On the S Series these are 1.2 μm, 0.8 μm, and 0.5 μm respectively. This is the machine’s ability to actually resolve a small feature, independent of how long it is. A machine with excellent length error and poor probing error will fail you on small holes and tight radii.

How to specify it — the decision rule:

Size the instrument toone-quarter to one-tenth of your tightest tolerance. If your tightest callout is ±0.02 mm (20 μm), you need a machine that holds 2–5 μm on that feature. If it’s ±0.01 mm, you need 1–2.5 μm.

Mapping that rule to real machines:

Tightest toleranceRequired instrument errorNAVINERTIA options
±0.01 mm≤ 1–2.5 μmS Series; K100 with the optional 8 × 5 mm field (1 μm)
±0.02 mm≤ 2–5 μmK100 (2.0 μm in 82 × 55 mm); K433 wide-field channel (2.0 μm single field); E Series
±0.05 mm≤ 5–12 μmK433; K321 stitched; G Series
±0.1 mm≤ 10–25 μmMH / ME Series (2.5 + L/200 μm)

The trap:​ comparing machines at one length. A 1.2 + L/200 μm machine and a 2.0 μm single-field machine swap places depending on whether your part is 10 mm or 200 mm. Always evaluate atyourL.

Spec 3 — Field of View vs. Resolution (Not “Magnification”)

Magnification is the number buyers ask for and the one that tells you the least. What matters is thefield of view at a given accuracy​ — because field of view and resolution trade directly against each other. A wide field measures large features in one shot but resolves fine edges less sharply; high magnification does the reverse, and costs you travel time.

This is where a flash instrument’s spec sheet looks completely different from a VMM’s, and where it wins:

Field of view (mm)Single-field accuracyStitching accuracyNAVINERTIA models
82 × 55 × 1002.0 μm​—K100, K221, K321, K433 (Lens 2)
26 × 17 × 311.5 μm2.5 + L/200 μmK221, K321
20.6 × 13.7 × 24.81.5 μm—K100
8 × 5 × 10 (optional)1 μm​2.0 + L/200 μmK100, K221, K321
Lens 1 zoom channel: 1.67 × 1.24 → 10.71 × 82.0 + L/200 μm—K433 (Lens 1)

The K100 pairs a 20-megapixel industrial camera with a dual-telecentric lens that covers a field roughly50× that of a conventional video measuring machine​ — which is how it measures100+ dimensions in a single exposure. The K433 goes further and puts two optical systems on one machine: a 6.5:1 continuous zoom (0.7–4.5×) for micron-level detail, plus a 0.16× dual-telecentric wide field for fast large-area scanning. For a PCB or a double-sided stamping, that means you scan the whole board in one shot, then re-measure the critical features at zoom — without changing setups or re-fixturing.

At the other extreme, the S Series runs a 13.3:1 zoom from 1.2× to 16× optical (≈920× total), which is what you want when the feature itself is the hard part.

Rule of thumb:​ if your smallest critical feature is below about 0.05 mm, you need a single-field accuracy at or under 2 μmanda zoom channel on the same machine. That combination is the K433.

Spec 4 — Measuring Range, Z Travel, and the Load Nobody Reads

Travel​ should be sized aslargest part + 30–50 mm of fixturing allowance. The jaws, the clamp edges, and the lens approach all eat travel. Model numbers are usually a travel code: V433 = 400 × 300 mm, K433 = 400 × 300 mm.

Z travel is a separate decision.​ The S, E, and K433 all carry 300 mm of Z; the mobile-platform flash units (K221 / K321) carry only 75 mm. Z governs the tallest boss you can focus through and the deepest pocket you can reach.

Then there is the load figure almost no buyer reads.​ Most spec sheets list one number — maximum load. Serious ones list two, because they are not the same thing:

ModelMax. loadGuaranteed-accuracy load​
V323S /V433S/ V543S30 kg5 kg​
V323E/ V433E / V543E30 kg10 kg​
K43330 kg10 kg​
K100 /K221 / K3215 kg5 kg
G Series60 kg30 kg

“It fits on the stage” and “it measures accurately on the stage” are two different statements.​ Always ask which load the quoted accuracy is guaranteed at.

Where each range lands:

  • Over ~1 m in any axis → G Series gantry (V562G at 550 × 650 mm through V8102G at 850 × 1050 mm, 500 mm/s XY, 60 kg load), built for battery electrodes, bipolar plates, LCD panels, and large PCB panels.
  • Shafts, rods, lead screws, cutting tools, fibre, glass rod → W Series horizontal (V2515W) or the K341W horizontal dual-vision, which handles parts up to 400 mm long and Ø200 mm.
  • Everything between → S / E / K433 at 300–500 mm X.

Spec 5 — Resolution, Repeatability, and the Operator Variable

Three words get used interchangeably and shouldn’t be:

  • Resolution​ is the smallest increment the machine reports. Every NAVINERTIA VMM and flash instrument runs a 0.1 μm linear scale with a 0.1 μm minimum display unit.
  • Accuracy​ is how close the report is to truth (Spec 2).
  • Repeatability​ is the spread when you measure the same feature repeatedly.

0.1 μm display resolution does not mean 0.1 μm accuracy.​ Resolution is the ruler’s finest mark; accuracy is whether the ruler is the right length.

Here is the part most buying guides skip: in production, the biggest repeatability problem is usually not the machine — it’s the fixture, the part placement, the focus, and the operator’s edge-finding habit. This is precisely where a one-shot architecture changes the equation. Flash measurement removes the fixture entirely: parts are placed inany arrangement, and the software recognises the features and matches the image automatically. Within one exposure, 100+ dimensions are captured with no stage travel, no refocus, and no per-feature operator judgement.

The hardware still matters. Across the NAVINERTIA range: three-axis precision guideways, three-axis servo motors, fully closed-loop CNC control, 0.1 μm linear scales, acceleration to 1000 mm/s², precision linear-motor stages on the K221 / K321, and a granite main structure verified by finite element analysis on every VMM. Granite isn’t decoration — it’s what keeps the geometry stable over a ten-year service life.

And on the human side: iVision is built so a front-line operator is productive after1–2 hours of training. When the result depends less on who is standing at the machine, your gauge R&R improves.

Spec 6 — Illumination: The Spec That Decides Whether You Get an Edge at All

You can have perfect optics and a perfect scale and still measure nothing, because edge detection is a lighting problem before it is an optics problem.

Ask for four sources, and ask whether each isprogrammable:

SourceWhat it doesNAVINERTIA implementation
Telecentric parallel backlightClean silhouette and true outlineLED telecentric parallel cold light, 256-level
Programmable surface ring lightSurface features, contrast on varied finishes5-ring / 8-zone programmable LED, 256-level (S, E, G, K433 Lens 1)
Coaxial lightReaches into deep holes and blind features256-level; standard on K433 Lens 1, optional on MH / ME / K100 / K221 / K321
Height-adjustable lightControls shadowing on stepped partsStandard on K100, K221, K321

If your product mix includes polished, transparent, or rubber parts on the same machine, programmable segmented lighting is a hard requirement — not an option.​ A fixed ring light produces specular highlights that shift with tiny angular changes; the edge detector follows the highlight instead of the edge, and you ship a false pass.

Two real cases from our own application work make the point better than any parameter:

  • Rubber seal rings.​ A contact probe compresses rubber and reads the deformation, not the part. Flash measurement with a combined ring + coaxial source raises contrast on the rubber and captures OD, ID, cross-sectional diameter, thickness, and inner/outer concentricity without touching it.
  • Oil seals.​ Multi-angle illumination reveals the lip and chamfer that a single top light hides, so ID, OD, height, lip, chamfer, eccentricity, and concentricity all come out of one cycle.

Spec 7 — Throughput and Automation

Three tiers, and the right one follows from question 4 at the top of this article:

TierFitNAVINERTIA model
Manual / semi-automaticLow volume, high mix, first-article and sample workME Series (manual) / MH Series (manual XY, motorised Z with autofocus)
Fully automatic CNCHigh volume, same part number, unattended runsS, E, G Series
One-touch flashSmall parts, many dimensions, high volume, frequent changeoversK Series​

The numbers that matter:

  • K100:​ multiple parts placed in any arrangement, several parts from the same batch measured in one shot, 100+ dimensions per field.
  • Dedicated O-ring module:​ no programming at all. One touch measures ID, OD, and wall thickness inabout one second — over 30× faster than manual measurement. For a sealing-part supplier running hundreds of parts a shift, that single line item is usually the entire business case.
  • K433:​ fully automatic three-axis programmable measurement, plus a 300 mm Z travel in an open frame designed for collaborative-robot integration.

If you are running high mix, the metric that matters is not stage speed — it’schangeover time. Can the software import your CAD file, let you click the features, and generate the routine? Or does every new part number mean a scripting session? iVision supports CAD-based quick measurement, automatic barcode scanning, and map navigation precisely so that switching part numbers stays cheap.

For automated cells, NAVINERTIA also supplies an SDK-based control application that handles collaborative-robot loading and unloading, conveyor control, program invocation, light-source adjustment, and stage movement, with a remote server that supervises multiple instruments and exposes both report-query and live monitoring interfaces.

Spec 8 — Software, GD&T, and Where the Data Goes

Hardware sets the ceiling. Software sets the floor​ — it decides your daily throughput and whether the result is the same on Monday as it is on Friday.

iVISION is developed in-house by Navinertia. Licensing, included modules and upgrade scope are confirmed in the quotation. Check the following when evaluating measurement software:

  • Element coverage.​ Points, lines, circles, arcs, rectangles, ellipses, keyway slots, open and closed curves, planes, spheres, cylinders, cones, seal grooves, threads, and mesh screens — plus constructed elements (centre points, extreme points, parallel and tangent lines, inscribed and circumscribed circles).
  • GD&T and reporting.​ Import CAD and compare to the drawing; export to Excel, PDF, Word, or TXT with configurable formatting so pass/fail is readable at a glance.
  • Automation hooks.​ Autofocus, automatic edge finding, automatic illumination, staged with platform movement for batch runs.
  • Sensor fusion.​ If you later add a contact star probe, a 3D line-scan laser, a point or line spectral confocal sensor, an areal confocal 3D microscope, or a triangulation laser, the software must drive them in one coordinate system.
  • Temperature compensation.​ iVision supports both temperature-mode and ratio-mode compensation, so accuracy can be corrected quickly when the room isn’t at 20 °C. Very few vendors offer this; on a real shop floor it matters more than the last micron on the spec sheet.
  • Data destination.​ Measurement results should flow straight into your SPC, MES, or ERP — no manual transcription. NAVINERTIA also supplies a full data layer: first-article / sampling / patrol / full-inspection capture, online inspection submission, distributed SaaS or on-premises storage, automatic SPC control charts, and an IQC/PQC/OQC quality dashboard.

The trap:​ ask whether the licence includesallmeasurement functions, and whether CAD comparison, GD&T, or SPC reporting are paid add-ons. Ask whether upgrades are free for the life of the machine.

Two Tables Worth Saving

A. How to read VMM specifications

ModelEUX/EUY MPEEUXY MPEEUZ MPEReal XY error at L = 200 mm
V323S / V433S1.2 + L/2001.5 + L/2002.8 + L/1502.2 μm
V543S1.5 + L/2001.8 + L/2002.8 + L/1502.5 μm
V323E / V433E1.8 + L/2002.0 + L/2005 + L/2002.8 μm
V543E2.0 + L/2002.2 + L/2005 + L/2003.0 μm
K433 Lens 1 (zoom)2.0 + L/2002.2 + L/2005 + L/2003.0 μm
K433 Lens 2 (wide field)​2.0 (constant)​2.0 (constant)​—2.0 μm — independent of part size​
K100 (82 × 55 mm field)​2.0 (constant)​——2.0 μm​
V562G / V682G2.2 + L/2002.5 + L/2005 + L/1503.2 μm
MH / ME Series2.5 + L/200——3.5 μm

Read the bottom two rows again. On the flash channel the error is a constant; on the traveling channel it grows with your part.That is the economic value of fitting the part into one field of view.

B. Match your part to a machine

Your part and volumeSpecs that actually decide itNAVINERTIA model
Small precision parts in batches — O-rings, gaskets, watch cases, radar housings, simple injection-moulded partsSingle-field coverage, no fixture, no programming, dedicated O-ring moduleK100​
Same, but the part exceeds one field of viewStitching accuracy on a mobile platformK221​ (200 × 200 × 75 mm) /K321​ (300 × 200 × 75 mm)
PCB, double-sided stamped parts, connectorsWide-field fast scan + micron-level detail re-check on one machineK433​
Shafts, rods, lead screws, cutting tools, optical fibre, glass rodHorizontal optical axis, radial + end-face in one setupW Series V2515W /K341W​
Large flat parts — battery electrodes, bipolar plates, LCD panels, PCB panels, large glass or plasticGantry open space, 500 mm/s, 60 kg loadG Series V562G / V682G / V8102G
Tightest tolerances and high magnification — optical communications, semiconductors1.5 μm class, 13.3:1 zoom to ≈920×S Series V323S / V433S / V543S
General industrial precision parts, balanced budget2.0 μm, fully automaticE Series V323E / V433E / V543E
High mix, low frequency, mostly 2DEase of use and costMH Series (semi-automatic) / ME Series (manual)
Solder-ball and lead coplanarityLine-laser 3D plus visionL Series V400LS
Wafer thickness and flatnessDual opposed point-spectral confocalC Series W350C

The Three-Step Decision Tree

  1. Architecture first.​ Does the part fit one field of view? If yes, and it has many dimensions and runs in volume → flash. If you need Z-axis form or the part is very large → traveling.
  2. Accuracy second.​ Take one-quarter to one-tenth of your tightest tolerance, then pick the class from Table A.
  3. Automation last.​ Daily volume and changeover frequency decide manual, CNC, or one-touch.

Do not reverse this order. Most bad purchases come from fixing the budget and the travel first and thinking about tolerance last.

Your RFQ Checklist — Copy and Send

Paste this into an email to any supplier (including NAVINERTIA). A vendor who answers all ten in writing is a vendor worth shortlisting.

  1. Please state MPE perISO 10360-7​ forEUX/EUY, EUXY, and EUZ​ as formulas, and state the test temperature.
  2. What are the probing errorsP2D, PV, and PZ?
  3. For flash-type instruments: statesingle-field accuracy​ andstitching accuracy​ separately, each with its field of view.
  4. What is themaximum load, and separately theguaranteed-accuracy load?
  5. Camera resolution, lens type (is it telecentric?), zoom range, andfield of view at each magnification.
  6. Which light sources are included, and which areprogrammable​ (number of zones, brightness levels)? Is coaxial light standard or optional?
  7. Z travel and Z-axis accuracy. Can a touch probe, laser, or confocal sensor be added later?
  8. Measured cycle timeon our sample parts​ — not a theoretical figure.
  9. Does the software licence include CAD comparison, GD&T, and SPC? Are upgrades chargeable?
  10. What will the acceptance report contain, and what are your annual recalibration and local service response terms?

Two Things That Never Appear on a Spec Sheet

The room is part of the measurement system.​ Every NAVINERTIA instrument is specified at20 ± 2 °C, 30–80% humidity, vibration below 0.002 g under 15 Hz, away from vibration sources, with a protective earth of≤ 4 Ω. An uncontrolled floor will erase a sub-micron specification faster than any machine defect. Budget for the room, the granite, and the isolation — and if you can’t hold the temperature, use software that compensates for it.

Total cost of ownership runs five years, not one.​ Purchase price, installation, environmental conditioning, annual reverification, calibration artefacts, software maintenance, and training. Consider the physical difference: a K100 is96 kg, 650 × 400 × 940 mm, and 100 W. A K433 is 600 kg and 1500 W; an S or E Series is 480–680 kg and 750 W. A cheaper machine that demands a better room, more operator skill, or more frequent recalibration usually costs more by year five.

FAQ

What is the difference between a vision measurement system and a CMM?

A CMM touches the part with a probe; a vision system measures optically and never contacts it. Vision is far faster on small, flat, delicate, or easily deformed parts, and contact is better for hidden geometry and deep internal features. Multisensor machines combine both in one coordinate system.

Is a flash (one-shot) measuring instrument just a cheaper, lower-grade VMM?

No — it’s a different error model. A traveling VMM’s error grows with measured length; a flash instrument’s single-field error is constant regardless of part size or how many dimensions you measure. Flash units are 2D and limited in Z, so they don’t replace a VMM — they beat one whenever the part fits one field and carries many dimensions.

Does higher magnification mean higher accuracy?

No. Magnification reveals detail; accuracy comes from the whole system — optics, scales, structure, lighting, alignment, and environment. And magnification trades against field of view: more magnification means a smaller field and more stage travel, which can actually cost you throughput.

How accurate does my machine need to be for a ±0.02 mm tolerance?

Target one-quarter to one-tenth of the tolerance — 2–5 μm. A 2.0 μm single-field flash instrument (K100, or the K433 wide-field channel) sits in that band for any part inside the field; on a traveling VMM you must also evaluate the length term at your actual part size.

Do I need 3D (Z) measurement, or is 2D enough?

If your drawing only calls out profile, position, diameter, and distances, 2D flash is faster and more repeatable. Add Z the moment you must verify height, depth, flatness, step, or coplanarity — the K433 covers both, using its zoom channel for Z (EUZ 5 + L/200 μm) and its wide-field channel for fast 2D.

Can O-rings really be measured without a fixture?

Yes. Flash measurement places parts in any arrangement and identifies features automatically. The NAVINERTIA O-ring module needs no programming — one touch returns ID, OD, and wall thickness in about one second, more than 30× faster than manual measurement. Non-contact measurement also removes the deformation error a contact probe introduces on rubber.

What does “guaranteed-accuracy load” mean, and why is it lower than max load?

Max load is what the stage can carry without damage. Guaranteed-accuracy load is the weight at which the published accuracy is actually assured — on the S Series, 30 kg versus 5 kg. Load the stage beyond the guaranteed figure and the quoted accuracy no longer applies.

What should an acceptance report contain?

Machine model and serial number, the standard and edition used, the tested configuration, the reference artefacts and their calibration, environmental conditions during testing, measured results against the stated MPE, the measurement uncertainty and conformity decision, the test date, and the identity of the testing organisation. A product brochure is not evidence that your installed machine passed anything.

Get Your Parts Measured Before You Buy

Specifications only get you to a shortlist. The only way to know whether a vision measurement system will holdyourtolerances onyourmaterial atyourtakt time is to measure your actual parts.

Send usyour drawing and two or three sample parts, and we will run them on the appropriate NAVINERTIA TECH instrument and return afree measurement evaluation report​ — cycle time, achieved accuracy, lighting setup, and the model we recommend, with no obligation.

If you are measuring O-rings, gaskets, watch cases, radar housings, or injection-moulded parts in batches, ask for a K100 evaluation. If you are measuring PCBs, double-sided stampings, or connectors, ask for a K433 evaluation.

→ Contact NAVINERTIA TECH to request your free measurement evaluation.

Or send your drawings, tolerances, and expected inspection volume, and we will come back with the model we recommend and a sample-part test plan — before you commit to anything.

NAVINERTIA TECH — 12 years in machine-vision metrology · 30+ patents and software copyrights · 3,000 m² manufacturing facility

Related reading:Optical inspection systems for precision parts·iVISION measurement software

Public References and Next Steps

For your own evaluation, agree the measured features and datums, fixture, program and sensor configuration, environmental conditions, calibration references, repetitions, uncertainty assessment, cycle-time boundary and acceptance decision rules. Customer drawings and actual measurement reports are not published here.

Request an Application Evaluation