Power control enables clean marking

Steve Hastings

A scanned carbon dioxide (CO₂) laser beam used for marking is switched on at the point of movement and can mark a distance of more than 2.5 mm in one millisecond. In contrast, visual resolution of a mark's flaws is as fine as 0.1 mm. Many attempts have been made to introduce scanning of CO₂ lasers into industries where visible acceptability of the processed article is a key requirement, such as the nonmetallic print, packaging, and component converting areas. Introduction into these industries has been hampered by poor marking performance.

Such effects can result in missing portions of a mark, where no energy has been generated by the laser from the modulation signal input (this is common over the first 1 to 2 mm of movement, but occasionally affects a complete marking vector movement); pulse pile-up as a result of overprocessing in the marking vector movement acceleration, deceleration, and sharp direction changes between adjoining movements; and power fluctuations between process repeats.

The standard method of CO₂ scanning sends a modulation signal from the system controller card to the laser radio-frequency (RF) amplifier to directly correspond with marking movements of the beam spot at the target plane. The modulation signal is switched off when a jump between two marking vectors takes place.

Pulse pile-up or overprocessing is a direct result of applying a single level of modulation input to the RF amplifier. When marking two adjoining lines with the first in the X direction and the second in the Y, the corner in which the two lines meet will show heavier processing than the rest of the lines, even in perfect conditions (and not accounting for any laser strike irregularities). This effect is directly related to X and Y galvanometer-motor deceleration and acceleration.

In industrial label processing, pulse pile-up can pierce the face material, the adhesive, and the carrier. This is visibly unacceptable, and on automatic material transports is a cause of catastrophic web tear and failure to remove the waste material around the labels. When this occurs, the machinery must be stopped and the transport reloaded, losing valuable processing time. The modulation stream can be controlled to reduce the laser energy in conjunction with X-Y galvanometer acceleration and deceleration, but this solution cannot account for the lowest 10% to 12% of movement speeds in which the lasing process cannot remain active and the CO₂ laser gas action dies. This solution does not improve laser power or wavelength instabilities, which, though they are separate effects, have the same result at the material-processing level.

Most CO₂ laser manufacturers specify their lasers to approximately ±7% power output stability. Tests on label stocks show that the difference in laser power between the desired "kiss cutting" and over-scoring the silicon face of the carrier can be a mere 4%. Wavelength instability commonly increases the problem. In a range of polypropylenes, for example, the absorption of light at 10.6 µm is 75% to 85%. At 10.2 and 11.0 µm, absorption can drop to 15%. Many CO₂ laser manufacturers specify their lasers between 9.6- and 11.2-µm wavelength stability. No commercially available RF-excited CO₂ laser constantly emits at 10.6 µm. Switching the laser input modulation stream in time to marks and jumps is the prime cause of instabilities, especially when each stream is only a few milliseconds duration. For example, an RF-excited CO₂ laser switched in conjunction with scanning requirements can have ±20% power output instabilities.

Bringing output under control
We began experiments on the potential of CO₂ laser energy as a processing medium for nonmetallic materials in 1988, when many modern techniques were unknown. Radio-frequency-excited CO₂ lasers were uncommon; the reasons for visibly unacceptable and irregular processing were as yet unproven. By 1996 it was determined that these problems were caused by the difficulties in the physics of trying to accurately control a gas laser. Such problems are a direct result of the actions of the ultraviolet (UV) starter lamp as it excites the gas within the cavity.

One way to improve striking is to increase the number and size of UV starter lamps, but this solution would compromise cavity integrity. There is an alternative, however, that fulfills all necessary criteria. Termed a power control device (PCD), the apparatus can accept a 10-mm-diameter 450-W beam and control transmission anywhere within a 0% to 99.8% range with a 600-µs response. Soon to be developed is a version handling a 15-mm-diameter 1000-W beam with an estimated transmission of 0% to 99.8% in less than 1 ms. In essence, the techno-logy controls the exact required laser energy reaching the target in direct relationship to targeting velocities. It works on the basic principle that a CO₂ laser is most stable running with constant input control and cooling. In this state, it is possible to hold the laser power output to within ±3%—and, with further research into coolant capacities, materials, and feedback loops, to within ±1%.

The laser-strike problem
Tests were made using two popular RF-excited CO₂ lasers to learn more about the problem of laser strike (UV starter lamp activation). In galvanometer scanning, when a marking vector is required, the laser is switched on with a modulation input stream and then switched off when the marking vector is complete and a jump (nonmarking) reposition is needed. The tests were aimed at clarifying exactly what creates instabilities, missing pulsing, and high pulsing just after the laser modulation signal input has started. The tests were carried out between the main usable scanning ranges of 10 to 50 kHz and duty cycles of 10% to 50%. These limits were arrived at as follows: at 10 kHz and 10% duty cycle, pulse separation is visible in a 70-mm2 field above 0.35-m/sec scan speeds (100-mm focal-length lens, 10.0-mm-diameter input beam, and 170-µm spot size), while quasi-continuous wave is effective above 50 kHz and 50% duty cycle.

Two examples clearly illustrate laser strike problems (see Fig. 1). After modulation is triggered, the gas has difficulty lighting from the UV starter. In the first example, at 25 kHz and 10% duty cycle, the first 24 optical pulses do not exist (0- to 960-µs time span). The optical pulse energy rises to a reasonably high output between the 25th and 46th optical pulses (960- to 1840-µs time span). The example demonstrates a common cause of missing portions of marks. The second example shows the laser strike problem at 50 kHz and 30% duty cycle (readings for 10% and 20% duty-cycle laser strikes were so poor as to be useless). Here, the first optical pulse only appears after 180 µs, or on the ninth input pulse. Relative stability is achieved by the 25th input pulse, or after 500 µs. This information can be translated into galvanometer-scanning-related marking times. Given a simple 200-mm2 box being marked within a 210-mm2 field (f-theta scanning, 300-mm focal length), at a 100-mm/s marking speed the box will be processed in 8 s; at 1000 mm/s it will take 800 ms; and at 2000 mm/s the marking time will be 400 ms (not accounting for X-Y mirror galvanometer-motor acceleration, deceleration, or jump, mark, and adjoining mark vector delays).

Benefits of stability
Power stability data from a 200-W RF-excited CO₂ laser using PCD show that the optical output is well within the ±7% manufacturers specification (see Fig. 2). Allied to the better than ±3% power stabilities demonstrated in these results, the wavelength stability is now also within ±1 laser transition line, thereby further improving uniform processing. With constant modulation, minor instabilities can be monitored and fluctuations compensated by a software feedback loop adjusting the offset of the PCD and smoothing output to the target to better than ±1% (see Fig. 3).

A stable laser and beam throughput controlled in direct relationship to target beam velocities allows arrays of scanheads to be used with beamsplitting from a single laser. Current optical coating technology generally limits the number of beam splits to eight; however, with the lasers now balanced, multiple lasers can be used instead, allowing the array to be infinite. In calibration, each split is measured after the PCD and each PCD's gain and offset finely tuned. Benefits of single-field array targeting are higher speeds, smaller spot sizes, larger targets, lower laser power requirements, and reduced waste material volume. The field can also be split into sections covered by one laser source.

The aeronautical, automotive, electronic, medical, packaging, and print industries demand faster, more accurate marking, larger field sizes, and—crucially—visibly acceptable digital processing. The power control device technology is effectively an ultrahigh-speed, multipositionable, intelligent power-attenuation switch acting upon the most stable output parameter of the CO₂ laser, providing visually acceptable marking in a simple manner.

STEVE HASTINGS is an applications manager at Raylase AG, Justus-von-Liebig Ring 9, D-82152 Krailling, Germany; e-mail: [email protected].