One of the most complicated aspects to consider when laser-cutting is the delicate balance between translational speed and laser power. We want to cut as fast as possible, but are there times when it is best to slow down? The short answer is: “definitely yes”.
We are going to share with our maker community the effects that speed and power have on cut width, kerf angle and engraving depth. This will be the first edition of an ongoing series of articles, with help from Mako.
So what exactly are we doing?
As any “laser maker” knows, the two most important parameters when calibrating a laser are translation speed and laser power. Other parameters such as resolution, engraving direction, and frequency will also have an impact on your work, but none as much as the big two. We will determine to what extent these two parameters have any predictable behavior on the quality of cut. More precisely, when we cut a piece of material we notice that the kerf, or the thickness of the cut, depends on the overall energy that we are focusing onto a point on a work surface.
For a first approximation in 2D, we have defined this Energy level as the Power of the Laser x the Laser Focal Surface divided by the Translation Speed.
For the sake of our tests:
Our experiment was simple: using “AutoDesk Fusion 360” we drew a series of 5mm by 5mm squares and cut 10 examples of each using different Energy settings. Each individual square was measured with a micrometer for two different characteristics:
1. Average Kerf width: The average kerf width (a.k.w.) is the width of the laser beam that cuts into the manufactured part. This measurement is important as it must be considered into the design for making tight fitting junctions while assembling multiple parts.
Experimentally we determined the a.k.w. by measuring the top and bottom surfaces of our samples in the X-direction and Y-direction. The formula used for average kerf width is:
2. Average Kerf angle: The angle of the kerf channel (a.k.a.) is the angle of the v-shaped burn channel in relation to the laser beam. Experimentally we determined the a.k.a. by measuring the dimensional difference between the top surface and the bottom surfaces for each sample. The formula used for the kerf angle is:
A second test was done where we engraved (rasterization) the 5mm X 5mm squares’ surface while varying the Energy levels and tested to see how much material was removed. This test required us to precisely measure the thickness of the material before and after each test with a high precision micrometer.
What we show is a clear relationship between the thickness of the cut, the angle of the cut and cut depth and the laser energy. As we increase the Energy, the conic laser beam will burn the material from the work-face towards the base in a cylindrical volume and therefore create parallel walls between the kerf. However, as the energy will be much higher, this will burn more material and therefore enlarge the overall tolerance within the cut channel.
The Kerf tolerance seems to be fairly predictable and is related to the Energy level that we focus onto the work surface. This is very convenient as it allows us to adjust our cut off-sets based upon the level of tolerance that we seek!
Within certain limits we can predict the tolerance of the Kerf by applying a simple linear equation. However, we must be careful because, as we can see, the more energy we focus onto the cut area, the less accurate our cut becomes. When we start to heavily melt the material within the localized region of the cut, the liquid nature of the cut zone becomes less predictable.
We see that the tolerance is directly related to the power and speed. Equally, the size, focal length of the laser lens, and the users ability to precisely position the Z-height above the work surface play an important role in the accuracy of this measurement. This will be explained in more detail in the “future developments” section of this article.
Within certain limits we can predict the angle of the Kerf by applying a simple linear equation. However, we must be careful because, as we can see, the more energy we focus onto the cut area, the less accurate our cut becomes.
The tests done on the kerf angle are far less precise as the sample group is much smaller, and there is not enough data to verify that there is a linear relation between energy and kerf angle. The following calculations express a linear solution for the data set that we measured.
The Kerf angle is equally predictable and is related to the energy level that we focus onto the work surface. This is perhaps generally less useful when machining parts, but there exist some very specific cases in which it is very useful.
Example 1 (gears):
Making Gears: When laser cutting spur gears, it is important to maintain precise off-set (to minimize backlash and ensure tooth contact). For spur gears, it is also important to consider the tooth profile. Since the laser cannot actually cut perpendicular to the surface, all laser cut spur gears are in fact “conical gears”. The tooth profile on the front surface will not be the same as the profile on the back. This phenomenon may be totally negligible for most practical applications, but it is certainly something best understood so that considerations can be made during the design phase.
Example 2 (Microfluidics):
“Microfluidics deals with the behaviour, precise control and manipulation of fluids that are geometrically constrained to a small, typically sub-millimeter, scale at which capillary penetration governs mass transport. It is a multidisciplinary field at the intersection of engineering, physics, chemistry, biochemistry, nanotechnology, and biotechnology, with practical applications in the design of systems in which low volumes of fluids are processed to achieve multiplexing, automation, and high-throughput screening. Microfluidics emerged in the beginning of the 1980s and is used in the development of inkjet printheads, DNA chips, lab-on-a-chip technology, micro-propulsion, and micro-thermal technologies.” – Wikipedia
Due to the relative scale and consistency required to manufacture a Microfluidic system, it is imperative to take into consideration the capillary geometry as it will have an impact on the performance of the system.
One must take into consideration the cross section of the capillary, of which the two principal dimensions will be depth and taper (see figure). The proposed calculations for taper angle may allow for the development of more precise microfluidic systems.
The cut depth is very predictable and is related to the energy level that we focus onto the work surface. This is interesting as it allows us to adjust our engraving depth as well as our cutting parameters for varying thickness of PMMA.
The tests were all done at relatively low energy levels as we wanted to reduce the impact of smoke. Tests at higher power are less precise as we have a high concentration of smoke that stays within the channel, as well as the refraction of the laser against the non-parallel kerf walls. The results of these tests still put us within 10% of our required energy levels for cutting material up to 6mm.
This information is useful when doing grayscale engraving for images, 3D reliefs and lithophany jobs. Using software such as Photoshop, CorelDraw, or any image processing software, we can transform any image to a black and white image. From which we can apply a rasterization work-flow using our laser.
Generally, we adjust the image to create a 120dpi-300dpi 16-bit grayscale image. Increasing the resolution or color depth requires extremely slow engraving speeds to extract any improvement in quality. Once the image is uploaded into our laser processing software (RdWorks V8 at Mako), we calibrate our laser to engrave the black regions (RGB: 255,255,255) at our maximum depth and the white regions (RGB: 0,0,0) to our minimum required depth (usually 0.1mm).
This technique allows us to laser engrave images into plastic or to have precise pocket depths for mechanical parts. The depth and scope of this technique and all of the important parameters is outside the scope of this article.
There are a number of shortcomings to this model.
In future articles we would like to determine 2 additional criteria:
- The incertitude of the laser placement in Z-axis;
- The X,Y acceleration rates for our laser (and a defined method to measure this) so that consistent energy transfers can be calculated;
- A more developed model that considers laser intensity and optical absorption, refraction and reflection during the burning of the material;
- Heat dissipation in the work material and its effects on kerf width.
First, our first treatment does not consider the imprecision of placing the laser over the working surface. We provided merely an approximation where we considered the laser diameter to be 250 microns and the intensity of the Laser to be constant across the Beam. As we can see in the figures below, the intensity and focal diameter are direct consequences of the user’s ability to properly focus the beam. In future articles we would like to use ray-tracing optical simulators to have a better model that will allow us to determine the gradient of the beam intensity in function of our Z-height.
Secondly, the lasers translation speed is not constant. As we decrease the overall length of a line segment, the effects of the acceleration/deceleration of the laser carriage becomes more apparent. This means that for a small 5mm line, it may be impossible to test with translation speeds of over a few mm/s as the machine will never have the time to reach the designated speed. There are companies that have designed this acceleration/deceleration into their time estimation and it will be necessary to determine the “critical speed” for given line lengths, over which the calculated energy transmission can be considered false.
Lastly, we would like to develop a more complete model where we can take into consideration the absorption of 1016nm light by given materials, their ability to refract the light as well as their ability to transmit this energy as heat. The current test shows that there is a certain correlation to what we crudely called “energy transmission” and its effects on machining. It would be far more elegant and useful to have a formula to apply, defined by real physical characteristics, that would give us the effects on machining.
We hope this article gives existing Fablabs and makerspaces a new perspective on the finesse that we can attain with the Co2 laser engraving technology. Equally, for new labs, this article can help find your first set of “rough” parameters to get your laser cutting efficiently and safely. We will continue to test on different materials and we hope to develop a community of makers that would like to contribute to this project.
We need to improve the physical model that we have developed so that we can take into account the physical properties of materials. In the long term, the goal would be to correlate the effects of laser cutting to properties (optical refraction & absorption, thermal conductivity, energy of vaporization, etc.) made available on the datasheets of new materials. With these means, we may be able to predict the parameters for materials and determine their feasibility before having to purchase or test.
Our final goal would be to develop a small app/widget where a user could input their laser’s nominal power and material choice for the app to output the required settings. This would allow new makers to become operational faster, and workshops to be more profitable.
This series of experiments was interesting and illuminating for us at Mako. We hope the reader has found it equally as interesting and helpful. We are open to all discussion in regard to working to build a community to refine and perfect our project. We wish you luck with your lasing and keep a look out for our next installment: Effets mécaniques de la calibration laser sur MDF 3mm.