Knowledge
HM Series Online Image Measuring Heads: Five Models, Ranges and Selection
The POMEAS HM series is a family of online projection image measuring heads. Each head projects collimated green LED light past the product, images the resulting shadow on a CMOS receiver, and measures the silhouette — outer diameter, width, height, corner radii and other geometry — while the product is moving, without stopping the line. Five heads cover measurement fields from 6 mm to 120 mm. This page gives the published figures for each, explains the optical design that makes the family stable, and sets out the selection logic.
The Five Heads
| Model | Measuring range | Repeatability | Position measurement accuracy | Positioning in the family |
|---|---|---|---|---|
| HM-1006 | φ6 mm | ±0.03 µm | ±0.2 µm | High-accuracy model |
| HM-1015 | φ15 mm | ±0.05 µm | ±0.3 µm | High-accuracy model |
| HM-1040 | φ40 mm | ±0.08 µm | ±1.0 µm | Standard model |
| HM-1065 | φ65 mm | ±0.1 µm | ±1.2 µm | Wide-field model |
| HM-1120 | φ120 mm | ±0.2 µm | ±2.5 µm | Extra-wide-field model |
Read the pattern before you read the numbers: field of view and achievable accuracy move in opposite directions. The smallest head gives the tightest figures because the same receiver resolves a smaller field more finely. Selecting a head is therefore a question of what is the largest feature you must see, and what tolerance must you hold on it — not of buying the best number on the sheet.
Why the Repeatability Figures Are So Small
The ±0.03 µm repeatability published for the HM-1006 is 30 nanometres. That is a manufacturer's specification-table figure and the test conditions — environment, sample count, artefact, integration time — are not stated in the material we hold. It should be read as an instrument capability under defined laboratory conditions, not as an achievable production tolerance. On a real line, achievable tolerance is set by product presentation, vibration, temperature and contamination as much as by the head. Any figure quoted here should be confirmed against your own artefact before it enters a control plan. This is the same discipline described in calibration and measurement stability.
The Optical Design That Does the Work
Three elements of the design produce the stability:
- Telecentric optics on both sides. A telecentric system is arranged so that the accepted light is parallel to the optical axis, which means the apparent size of an object does not change as it moves toward or away from the lens. The HM family puts telecentric optics on the transmitter and the receiver. A partly telecentric arrangement — telecentric on one side only — still shows size change with positional shift, because some rays arrive at an angle.
- High depth of field. Combined with the double telecentric design, this is what allows the datasheet to claim that a workpiece offset within the depth of field still measures accurately. In practice it buys tolerance to pass-line wander, which is the most common source of inline measurement error.
- Near-zero distortion plus full-field correction. Low distortion design, corrected across the whole field, is what lets a single calibration hold accuracy at the edges of the view and not just at centre.
The comparison between a telecentric and a conventional lens — and when each is justified — is set out in measuring shaft outer diameter with a telecentric lens.
35 µs Exposure: Measuring Without Stopping the Line
The family uses an exposure of 35 microseconds. At that exposure, a product moving through the field is effectively frozen, so the shadow is captured without motion blur and the line does not need to pause or index. This is the single biggest operational difference from a bench measuring system: the part is measured in transit. The trade-off is that a very short exposure needs sufficient illumination, which is why the transmitter uses a dedicated green LED source rather than ambient or general lighting.
What One Shot Returns
The manufacturer states that a single capture can output up to 100 items — outer diameter, width, height, corner radii and other drawing-indicated dimensions — and that the same capture can also assess defect-type features such as foreign matter, burrs and chips. Two caution flags. First, "up to 100 items" is a software capability statement about how many dimensions can be evaluated from one silhouette; it is not a statement that any 100 tolerances can be held to the same accuracy. Second, defect detection from silhouette works for defects that change the outline — a burr protruding from the edge — and not for defects that only change surface appearance inside the outline. The capability boundary is discussed further in what one-shot multi-dimension measurement really returns.
Applications Named by the Manufacturer
The datasheet names carbon film resistors, heat sinks, drills and cylindrical parts. These share a profile that suits silhouette measurement: a consistent cross-section presented to the beam, dimensions that are defined by the outline, and production rates high enough that stopping to measure is not acceptable.
Selecting a Head: The Sequence
- Fix the largest dimension that must appear in one field. That sets the model — do not size down and stitch.
- Check the tolerance against that model's position measurement accuracy, not its repeatability. The accuracy figure is the one that governs agreement with true size.
- Confirm the pass-line stability you can actually hold. If the product wanders more than the depth of field allows, no head will save the reading.
- Decide whether the job is 100% inline at line rate or sampled at a bench. For the comparison, see inline versus offline optical measurement.
What the HM Family Does Not Do
- It measures the silhouette. Internal features, bore diameter and wall thickness are outside it.
- It is a profile method. True 3D form, flatness across a surface and step heights measured normal to the view need a different principle — see online gauge, confocal and vision compared.
- It does not replace a roundness instrument for lobing or out-of-round tolerances.
FAQ
Which HM model do I need for a 30 mm part?
The HM-1040, with a φ40 mm field, is the smallest head that contains a 30 mm object in one view. Its published figures are ±0.08 µm repeatability and ±1.0 µm position measurement accuracy. Do not select the HM-1015 (φ15 mm) and attempt to cover the part in two captures — stitching fields reintroduces the positioning error you were trying to avoid.
Why is the repeatability better on the smaller heads?
The receiver resolves a fixed number of pixels. Spread that over a 6 mm field and each dimension is sampled more finely than the same receiver spread over 120 mm. Smaller field, finer resolution, tighter repeatability — which is why the selection starts from the largest feature, not from the best headline number.
Is ±0.03 µm achievable in production?
It is a manufacturer's specification-table figure for the HM-1006, published without test conditions in the material we hold. Treat it as instrument capability under defined conditions. Production tolerance depends on presentation, vibration, temperature and window cleanliness as well.
Can the product keep moving during measurement?
Yes. The 35 µs exposure freezes the product, so the head measures in transit and the line does not need to stop or index for the capture.
How is this different from the LM-1030 diameter gauge?
Both use transmitted light and a shadow. The LM-1030 is a dedicated diameter gauge for 0.3 to 30 mm stock at 2400 samples per second. The HM heads are image measuring heads that return a full set of geometry — diameter, width, height, radii — across fields from 6 mm to 120 mm. The principle-level comparison is in inline diameter measurement methods.
Does it detect surface defects?
Only those that alter the silhouette — burrs, chips, foreign matter on the edge. A scratch or stain entirely inside the outline does not change the shadow and will not be reported by this method.
Can it measure runout or swing of a rotating part?
The manufacturer lists swing during rotation and workpiece position coordinates among the measurable items, so yes — provided the swing manifests as a change in the shadow position, which it does for a part rotating in the beam.
Next Step
Tell us the largest dimension you must capture in one field, the tolerance you must hold, the line speed and the material. We can confirm which HM head fits, or whether the job is better served by a dedicated diameter gauge such as the LM-1030 or by an offline one-click system such as IMAGE 3.
Subscribe to Our Newsletter
Simply enter your email to receive the latest news and insights from Pomeas. Stay connected with Pomeas and be the first to discover new innovations in optical excellence.
