Industry Information
How to Read a Vision Inspection Machine Specification Sheet
Repeatability is not accuracy, throughput always has conditions, and camera counts vary by model - how to read each line of a vision inspection machine spec sheet.
The Spec Sheet Is a List of Conditions
Every vision inspection machine has a specification sheet, and most purchasing mistakes start with reading it the wrong way - treating each number as a standalone promise. The sheet is actually a list of capabilities with conditions attached: throughput that depends on part size and camera count, repeatability that describes measurement rather than detection, camera counts that vary by model for structural reasons. This guide walks through the typical lines of a spec sheet, using the Shengyi New Material (Huzhou) Co., Ltd. brochure as the worked example, so you can read any supplier's sheet - ours or a competitor's - and know which questions to ask under each number.
Repeatability vs Accuracy: What ±0.01 mm Actually Says
Spec sheets list a repeatability figure per model - in our brochure, ±0.01 mm on the single-table, dual-table, food and pharma and compression machines, ±0.02 mm on the tensile series and the double-layer machine, ±0.05 mm on the wire and tube machine. Three things to know when reading it:
• Repeatability describes how consistently the machine measures the same feature across parts - it is not an accuracy claim against a certified standard, and it is not a detection-rate promise.
• Different models carry different figures because their measurement tasks differ. Ask each supplier what task drives the figure - for our wire and tube machine, the ±0.05 mm figure sits on a continuous-line machine measured in metres per minute.
• The number that matters to you is the one compared against your part tolerance: if the tolerance is tight relative to the figure, raise it at quotation, along with how the figure was measured.
Throughput: Always Read the Parenthesis
Throughput lines are ranges with printed conditions: 120-840 pcs/min on the standard model "depending on product size, actual feeding and/or number of cameras"; 500-2000 pcs/min on the double-layer machine under the same style of conditions; 4-6 pcs/min on the ring machines with automatic feeding noted. The parenthesis is the specification. When you compare vision inspection machines, compare the conditions, not the top numbers: the top of the range assumes small parts, more cameras and (on some models) automatic feeding. Your realistic planning number is the one confirmed on your samples at your part size - which is why the published process on this site ends every rate discussion with "confirmed at quotation".
Camera Count: Why It Ranges from 1 to 32
Camera counts in our line run from 1-12 on the food and pharma machine to 3-32 on the dual-table machine, and the range is structural, not marketing: cameras buy coverage and speed. A single-camera configuration checks fewer features per pass; more cameras capture more faces and dimensions in the same rotation - which is also why camera count drives both the throughput range and the price. When reading any sheet, ask: which faces and features of my part are covered in one pass, and does thickness need an additional camera (on several of our models it does)? Full logic: how many cameras does your machine need (/posts/vision-inspection-machine-camera-count).
Size Bands: Read This Line First
Every machine lists an inspection size range - OD 6-80 mm on the standard model, 6-120 mm on the dual-table, 8-35 mm on the compression machine, 100-400 mm and 300-1000 mm on the tensile machines. This is the least negotiable line on the sheet: a part outside the band usually cannot be configured onto the machine at all, while bands the brochure marks "customizable" are confirmation items. Read size bands first, everything else second - the full map across our line is in choosing an O-ring inspection machine by size (/posts/o-ring-inspection-machine-by-size).
Reading Other Suppliers' Wordings
Sheets use different words for related ideas - translate before comparing: one supplier's "accuracy" may be what another calls repeatability; throughput may appear as pcs/hr rather than pcs/min; "inspection points" may stand in for camera count. When a sheet prints a detection rate, ask for its basis: which defect definition, how many parts, measured how, and whether it will be demonstrated on your samples. Conditions and demonstrations are what make any number comparable.
The Lines That Are Missing
What a spec sheet does not contain is as telling as what it does. The turntable-machine sheets in our line do not publish detection rates (the one brochure figure we do quote - the flying camera robot's - is attributed as the brochure states it and treated as a design target, with detection on your parts confirmed by sample test); ours does not list lead times or prices either, because they are configuration-dependent and confirmed at quotation. A sheet full of unconditioned superlatives - one detection rate for all parts, one speed with no conditions - deserves more questions than one with honest parentheses. The same logic applies to reading between suppliers: compare which conditions each one prints, because the conditions are where the engineering honesty lives.
How to Compare Vision Inspection Machines on Paper
Build a comparison grid with one row per spec line - repeatability, throughput and its condition, camera count and coverage, size band, materials, data output (what the machine logs and exports; the audit value of records is covered in vision inspection data traceability (/posts/vision-inspection-data-traceability)) - and fill it per supplier from their published sheets. Then take the grid to the sample test: the figures get you to a short list, and the sample run decides it. Sample arrangement terms are confirmed when you inquire; what to prepare is in the quote preparation checklist (/posts/vision-inspection-machine-quote-checklist), and how the numbers relate to price in vision inspection machine cost: 7 price factors (/posts/vision-inspection-machine-cost-7-price-factors).
Frequently Asked Questions
Is repeatability the same as accuracy?
No. Repeatability (±0.01 mm in our brochure) says the machine measures the same feature consistently; accuracy against a certified standard and defect detection performance are different questions - the latter is confirmed on your samples.
Which throughput number should I plan around?
Not the top of the range. Plan around the rate confirmed on your parts at your size and camera count during the sample test - the ranges on the sheet exist because those factors move the number.
Why do camera counts differ so much between models?
Coverage and structure: a dual-table machine with 3-32 cameras flips parts past cameras for both faces; a compact counting unit needs one. More cameras mean more features per pass - and more cost - so the count follows the part. The full logic is in the camera count guide (/posts/vision-inspection-machine-camera-count).
What if my part is outside the listed size band?
Check whether the band is marked customizable - several models are - and send the dimensions for confirmation. Bands decide fit before price ever comes up.
Next Step
Send us your part, its size band and the spec lines you are comparing, and we will fill in our column with each figure's printed condition. The sample test then turns the shortlist into evidence. Contact the team (/contact-us) - Monica (+86 138 5726 5705) or Arya (+86 173 2430 9735), or email monica@aivisioncheck.com.
