Color Mark Sensor

What is a Color Mark Sensor

A color mark sensor is a diffuse photoelectric sensor optimized to detect a small printed mark, often called a registration mark, eye mark, or print mark, as it passes the sensing window at high web speed. The sensor illuminates a fixed spot on the moving surface, measures the intensity of the light reflected back to its receiver, and switches a discrete output the instant the reflected brightness crosses a taught threshold. That output edge is the timing pulse a packaging machine, labeler, or press uses to register a downstream operation such as a cut, fold, seal, or label index, often by referencing the mark event against the position feedback of a servo drive that moves the web or the tool.

The defining characteristic is that the sensor works on contrast, not on absolute color. As the mark enters the light spot, the reflected brightness changes relative to the background. The sensor only needs to see a reliable difference between two states, mark and no-mark, to produce a clean edge. To maximize that difference across many mark-and-background combinations, modern sensors carry red, green, and blue LEDs and select whichever color gives the strongest contrast for the taught pair. This is why the same device is described in datasheets as a contrast sensor or a registration mark sensor rather than a color measurement instrument.

It helps to separate three closely related device families that buyers often confuse. A color mark sensor (contrast sensor) converts the scene to a grayscale brightness and triggers on a brightness change. A true color sensor reflects white light, splits the return into red, green, and blue channels, and assigns a measured color value, so it can distinguish dark blue from black or sort parts by hue. A luminescence sensor detects invisible UV-excited markers. All three are photoelectric, but only the color mark sensor is built for the very high switching speed that print and packaging registration demands.

The working principle traces back to ordinary diffuse photoelectric sensing, which detects the presence of an object by the light it returns. The refinement that created the color mark category was the addition of selectable colored illumination plus a fast teach-in routine. In the 1990s and 2000s, single-color LED contrast sensors required the technician to choose red, green, or white light from a contrast chart for each mark. RGB models that auto-select the best color, such as the SICK KT family, the Banner R58E Expert, and the Omron E3S-DC, removed that manual step and became the print-industry standard. Keyence later pushed a parallel approach with full-spectrum white-light sensing that evaluates chromaticity rather than only brightness.

In application terms, the color mark sensor is a workhorse of the converting, printing, and packaging industries. It synchronizes form-fill-seal baggers to the printed graphic, triggers rotary and intermittent labelers, registers gravure and flexo presses, and confirms the presence of date codes laid down by a coding machine as well as tear notches. Its value is measured in registration accuracy at speed: a few hundred microseconds of timing error at high web velocity translates directly into misplaced cuts and scrapped product, which is why response time and jitter dominate the selection conversation.

Sensor Types and Classification

Buyers meet several overlapping device names on supplier sites: color mark sensor, contrast sensor, registration mark sensor, color sensor, true color sensor, and full-spectrum sensor. The distinctions are not marketing noise: they map to different detection physics, speeds, and use cases. Choosing the wrong family is the most common selection mistake, because a true color sensor sized for hue verification is usually far too slow for cut-off registration, and a contrast sensor cannot report which color it sees. The table below frames the four families a packaging or print engineer actually compares.

Color mark and contrast sensors are the same family under two names. They convert the illuminated spot to a single brightness value and trigger on the change as a mark passes. They are the fastest of the group because they evaluate one number, not a color computation. The SICK KTS reaches 50 kHz switching frequency, and the Banner R58E and Omron E3S-DC reach a 50 microsecond response, which is what high-speed flexible-packaging lines require. These sensors cannot tell you what color a mark is, only that the contrast crossed a threshold.

True color sensors reflect light off the target and assign a measured color value from the red, green, and blue returns, so they can be taught to separate similar hues and confirm that a part is the correct color. They suit incoming inspection, color sorting, and verifying that the right cap or label color is present. The trade-off is speed: color computation and tighter tolerance windows push response into the sub-millisecond to millisecond range, which is acceptable for a presence check but not for resolving short marks at press speed.

Full-spectrum white-light sensors, exemplified by the Keyence LR-W series, illuminate with a white LED and analyze the whole reflected spectrum, evaluating chromaticity (hue) rather than only brightness. This lets them flag changes in surface finish or subtle color shifts that a contrast sensor would miss, and they offer selectable response times such as 200 microseconds, 1 ms, 10 ms, 100 ms, or 500 ms to trade speed against stability. They overlap both the contrast and true-color roles depending on the chosen mode.

Luminescence sensors sit just outside the color mark family but solve a related problem: detecting invisible UV-fluorescent marks, adhesives, or security features that have no visible contrast. They are the fallback when a visible mark is not available or not allowed on the finished product. When the task grows beyond a single contrast point to reading whole printed graphics, verifying print quality, or decoding codes, the job shifts to a machine vision system or a dedicated industrial barcode scanner rather than a single-point sensor. For a buyer, the practical takeaway is to start from the question being asked. If the question is when did the mark pass, choose a color mark sensor. If the question is what color is it, choose a true color sensor. If the question is did the appearance change, consider a full-spectrum sensor.

Light Source Technologies

The illumination source is the single most important design choice inside a color mark sensor, because contrast between a mark and its background depends entirely on which wavelength strikes the surface. The same mark can show strong contrast under one color and almost none under another. Three illumination strategies dominate: selectable RGB LEDs, fixed single-color LEDs, and broadband white LEDs. The table below compares them on the metrics that drive selection.

RGB LED illumination is the modern default for registration. The sensor flashes red, green, and blue in turn during teach-in, measures the contrast each color produces between mark and background, and locks the wavelength that gives the largest difference. Banner publishes representative RGB wavelengths of about 636 nm red, 525 nm green, and 472 nm blue for the R58E. Because red, green, and blue can be combined to approximate almost any color, an RGB sensor handles nearly any mark-and-background pair without a manual color-chart lookup. SICK describes the same behavior on the KTS and KTX: the optimum emitted light is selected automatically on RGB variants.

The reason this matters is physical. A green mark printed on a white background returns strongly under red light and weakly under green light, so red illumination yields high contrast and green yields almost none. A red mark on a yellow background behaves the opposite way. With only a single fixed color, the technician must consult a contrast chart and order the correct variant for each job, and a later artwork change can force a different sensor. RGB auto-selection collapses that decision into a one-button teach.

Single-color LED sensors remain in service for cost-sensitive lines that run one fixed mark-and-background pair. A red, green, or white LED is chosen once from a contrast table and never changes. They are inexpensive and perfectly reliable for a stable job, but they lack the flexibility to follow artwork changes and they offer poorer contrast on awkward color combinations. For a job shop that switches products frequently, the saving rarely justifies the lost flexibility.

White LED full-spectrum sensing takes a different path. Instead of switching among discrete colors, it illuminates with a broadband white LED and analyzes the full reflected spectrum, deriving chromaticity rather than only brightness. The Keyence LR-W series uses this approach to flag changes in surface finish and subtle color shifts that are hard to see by eye. It is less about a single fast trigger and more about stable detection of appearance change, with selectable response times trading speed against noise immunity. For pure high-speed registration on a clear mark, an RGB contrast sensor is usually faster and simpler; for distinguishing nearly identical finishes, full-spectrum sensing earns its place. The same dependence of contrast on wavelength is why a separate vision light source is engineered around a chosen color when an inspection camera, rather than an integrated contrast sensor, handles the print check.

Substrate, Optics, and Mounting

Once the light source is chosen, the substrate and optics decide whether the sensor sees a clean mark or a noisy one. Color mark sensors are diffuse devices that depend on light scattering back from a matte surface. Glossy, transparent, and metallized webs all break that assumption and need specific handling. The two parameters a buyer must respect are the focus distance, the working gap at which the spot is sharpest, and the spot geometry, the size and orientation of the illuminated rectangle on the surface.

Focus distance and mounting gap are tightly specified. The Banner R58E focuses at 10 mm from the lens with a tolerance of plus or minus 3 mm, and reliable sensing requires mounting within that window. The SICK KTS family is built around a 13 mm sensing distance with a tolerance of about plus or minus 5 mm. Web flutter that moves the surface in and out of this band reduces contrast, so a stable web path, an idler roller, or a backing plate under the sensing point is part of a good installation rather than an afterthought.

Spot geometry sets the smallest mark the sensor can resolve. Contrast sensors use a focused rectangular spot, not a round dot, so the short dimension can define minimum mark length while the long dimension averages out print noise across the mark. The SICK KTS spot is about 0.9 mm by 3.8 mm and the Banner R58E image is about 1.2 mm by 3.8 mm at focus. Critically, the long axis is oriented across the direction of web travel, so the small dimension along travel is what limits how short a mark can be. As a rule of thumb, keep mark length in the travel direction at least 1.5 to 2 times the spot dimension along travel, or the sensor averages mark and background and loses the edge.

Reflective and transparent substrates are the classic problem cases. Shiny foil and metallized film produce a specular hotspot that can saturate the receiver and hide the mark. Clear film reflects very little light, so there is barely any signal to work with. The standard remedy for both is to skew the sensor about 15 degrees from perpendicular, which redirects the mirror reflection away from the receiver on glossy webs and improves the usable signal on clear film. The table below summarizes practical substrate handling.

Enclosure and connection details belong to this stage too. Print and packaging environments are dusty and sometimes washed down, so an IP67 housing, exemplified by the SICK KTS and Banner R58E, and the IP65/IP67 rating of the Keyence LR-W, plus an M12 connector and a robust die-cast or metal body, are the baseline. Verifying ingress protection against IEC 60529 and confirming the operating temperature window, for example the LR-W rated for minus 20 to plus 50 degrees Celsius, prevents field failures that have nothing to do with optics.

Key Specification Parameters

Reading the datasheet is the core skill of color mark sensor selection. A spec sheet may list twenty parameters, but only a handful decide whether the sensor will register reliably at line speed. The table below compares the headline numbers of four widely deployed sensors so the parameter names become concrete, then each parameter is decoded underneath. All values are taken from the published manufacturer datasheets.

Switching frequency is the maximum rate of mark events the sensor can resolve, quoted in hertz or kilohertz. It sets the ceiling on how fast marks can arrive. The SICK KTS is rated at 50 kHz, which is far beyond most line requirements and gives ample headroom. To size it, divide web speed by the spacing between marks to get the event rate, then keep at least a 2x margin so that print variation and web speed surges do not push the sensor past its limit.

Response time and jitter together govern registration accuracy. Response time is the delay between a mark crossing the threshold and the output switching, for example 10 microseconds on the SICK KTS and 50 microseconds on the Banner R58E and Omron E3S-DC. Jitter, sometimes published as repeatability, is the cycle-to-cycle variation in that delay, such as the 5 microsecond jitter of the KTS and roughly 15 microsecond repeatability of the R58E. At high web speed it is jitter, not raw response time, that limits how tightly a cut or seal can be placed, because the constant part of the response can be calibrated out while the random part cannot.

Sensing distance and tolerance define the working gap and how much web flutter the sensor tolerates before contrast collapses. A 13 mm distance with plus or minus 5 mm tolerance, as on the KTS, is more forgiving of an unsupported web than a tight focus, while the Banner 10 mm plus or minus 3 mm focus demands a steadier path. Spot size, covered in Chapter 4, sets the smallest resolvable mark and must be matched to mark geometry and orientation.

Teach-in capability is a parameter buyers underweight. The available modes (1-point, 2-point, static, dynamic, and auto) decide how easily the line can be set up and re-taught after an artwork change. SICK and Banner expose a Quality of Teach indicator that reports the contrast margin, so an operator can confirm before production that the mark-and-background pair gives enough separation. A high teach quality is the field predictor of stable running.

Output, supply, and protection close the list. Most sensors accept 10 to 30 V DC and provide a fast push-pull or bipolar NPN/PNP discrete output with selectable Light-ON or Dark-ON operation, so the same part fits PNP-input or NPN-input controllers such as a PLC high-speed input or counter card. IP67 ingress protection per IEC 60529 and an M12 connector are the packaging-line baseline, and an increasing number of models add IO-Link for remote teach, live threshold readout, and diagnostics over the same two wires.

Selection Decision Factors

To turn the preceding five chapters into a specific model, follow the decision sequence below. Most registration problems trace not to a single wrong number but to deciding one level before the level above it is settled, for example fixing on a part number before confirming whether the job even needs a contrast sensor or a true color sensor. These eight steps double as a fixed RFQ template.

1. Detection task: Decide first whether the question is when did the mark pass (color mark / contrast sensor), what color is it (true color sensor), or did the appearance change (full-spectrum sensor). Getting this wrong wastes the entire selection.
2. Mark and background pair: Identify the mark color and the substrate color and finish. If the pair will vary across jobs, an RGB auto-select sensor is mandatory; a fixed single-color sensor only suits one stable pair.
3. Speed requirement: Compute the event rate from web speed divided by mark spacing, then require a switching frequency with 2x margin and a response time and jitter that hold registration tolerance at full line speed.
4. Mark geometry and spot size: Confirm the spot short dimension is smaller than the mark length in the travel direction, target a 1.5 to 2x ratio, and orient the long spot axis across the web.
5. Substrate handling: For glossy, metallized, or clear webs, plan a 15 degree skew and a stable web path or backing plate, and verify teach margin at production speed, not just when stopped.
6. Mounting and distance: Match the sensor focus distance and tolerance to the achievable mechanical gap and the expected web flutter, choosing a more forgiving tolerance for unsupported webs.
7. Electrical and protection: Specify 10 to 30 V DC, the correct PNP or NPN polarity for the controller, Light-ON or Dark-ON operation, IP67 or higher per IEC 60529, M12 connection, and IO-Link if remote teach and diagnostics are wanted.
8. Teach and serviceability: Prefer sensors with dynamic teach and a Quality of Teach readout for fast changeover, and confirm local stock, spare availability, and documentation before committing to a high-throughput line.

One dimension teams routinely overlook is serviceability and changeover speed. On a line that runs many SKUs, the time to re-teach the sensor after an artwork change is a real production cost, so dynamic teach, a clear teach-quality indicator, and accessible push-buttons or remote teach matter as much as the headline accuracy. SICK, Banner, Omron, and Keyence all maintain local distribution, calibration support, and documentation in major manufacturing regions, which makes them dependable choices for production lines that must keep registering for years. For non-critical, single-SKU lines, a lower-cost RGB color mark sensor can be adequate, provided you verify switching frequency, jitter, and teach quality against the same checklist before committing.