The Launch of Króna

For the past years I have learned to use various design tools, which I have applied on large building- and infrastructure projects. About a year ago, I started using these tools to create lamps and sculptures that could be CNC produced. This experiment has provided me with a creative outlet, which has been very satisfying for me.

Since january I have experimented a lot and created many shapes. Very few of them have made it across the barrier between the digital and the physical. However, some of them have. Many of these shapes now adorn our livingroom, but also the homes of a few of the people that have shown my experiments the most interest.

The amount of interest this hobby has been shown by friends and colleagues has been amazing. The interest has been such that I have decided to establish a small webshop where some of these projects can be purchased. I have refined them and made them easier to assemble, and turned them into better products.

I call this project Króna, which is a play on the icelandic word ljósakróna (En: Chandelier).

In the same way my first shapes were experiments, this is also an experiment. I have learned a lot about setting up a webshop, product photography and how to promote your product, making this experiment already partly successful.

I have had plenty of support from interested friends and colleagues, received lots of good advice, inspiration and help. I know that I have also been a bit demanding at times, as I have really liked talking about this journey. Thanks for everything!


Exponential development in surveying technology

Land surveying is a profession that relies heavily on technology. Being a technology-reliant profession it has gone through multiple periods of disruptive development. The rate of innovation was slow to begin with but has gradually been accelerating. It took us a really long time to go from ropes and sticks to optics and precision angular measurement using mechanical instruments and a much shorter time to go from the mechanical instruments to electric and digitial ones. With digitalization the rate of development changed and has accelerated even further.

For a long time the main form of surveying involved starting from known points, measuring the angles between them to determine the position of the instrument. This is the method used in theodolites and total stations. Then in the 80s the first commercial GPS instruments started entering the scene - indroducing a brand new mode of surveying. Since then many new modes of surverying have been introduced, as well as combinations of existing ones. Aerial Lidar systems mounted on commercial airplanes or spacecraft have made it possible to scan huge areas with high relative accuracy, fixed position 3D scanners have made it possible to capture complex scenes quickly, capturing millions of points from each station. 

A point cloud in Pix4D

For the past 10 years even more new modes of surverying have been added to the suite of tools available to surveyors. These tools are becoming increasingly less expensive and easier to use - to a degree that some have tried to rebrand the field as reality capture. Due to the falling price of technology, more and more consumer grade products have been appearing. An example of this is 123D Catch, which uses photogrammetry to capture 3D geometry through a smartphone. The flood of affordable drones has also left its mark with pro-sumer products such as Agisoft Photoscan or Pix4D that make mapping and the generation of 3D point clouds possible for consumers. This is an area that has been exclusive to surveyors and people with access to aircraft and expensive sensors.

For some years now, depth camera sensors such as the Kinect, have been available to consumers. These sensors combine a normal camera with a depth sensor which allows for creating a living point cloud. While the kinect was initially meant as an input method for controlling video games, many have shown that this can also be used for mapping. This is made possible by SLAM (Simultaneous-localization and mapping). SLAM allows for both building a point cloud of the sensors surroundings, but also keeps track of the sensors position and orientation. This technology is currently under rapid development as this is a key ingredient for making it possible for automated cars and robots to understand their environment. Recently a smartphone with a depth camera has been released by Lenovo and Google. SLAM is also being used with other modes of data collection such as lidar for rapid mapping of large buildings. Here an example from danish engineering consultants Cowi. 

Even more recently, methods for doing the same with regular cameras, have been invented. The video below shows how a point cloud is built using nothing but input from a camera. The point cloud is built in real time. 

The amount of data most certainly has gone exponential in surveying. We have gone from sparse measurements, collected one by one by a surveyor to tens of thousands of points being collected each second. The price of equipment is falling. Just recently researchers have developed a Lidar sensor on a chip that they state will be available to consumers within a few years for as low as $10. These sensors currently lie in the range of hundreds or thousands of dollars.

A Lidar chip
As the devices used for surveying or reality capture fall in price it is safe to assume that more and more people will gain access to them. Many of the tasks that are currently in the domain of surveyors will fall out of their domain. The amount of field work for data capture will most certainly be reduced for surveyors and the amount of data processing will increase. In many cases the surveyors will process data captured by others. Similarily the amount of setting out will also be reduced as the industry moves towards machine guidance and control systems.

One thing is certain, there will be no status quo. 

Generation of a laser cuttable geodesic sphere in Grasshopper

Since making my first lamp, I have continued to experiment with parametric geometry and laser cutting. I was inspired by the late Einar Þorsteinn's geodesic domes and Ólafur Elíasson's model room when I saw it at Louisiana last year. Following, I worked out a Grasshopper script that can generate a geodesic sphere, and expanded it so that the rules for generating a geodesic sphere can be applied to more or less, any convex shape. This became the basis for a script that generates a parametric, laser cuttable geodesic sphere.

To generate a geodesic sphere you can start with an icosahedron that contains a sphere. Each triangle face in the icosahedron is subdivided into identical triangular shapes and the triangles are then projected onto the sphere. This results in a network of lines that are called geodesic curves.

The sphere can be further subdivided.

This provides the network of curves that will be the basis of the sphere. Each curve will be extruded into a surface to become a member in the model. To assemble the members, a joint needs to be added where the members meet and slits generated where they meet. 

To visualize the end result, each member is given a thickness. 

The process works on other shapes as well, although the curves can only be regarded as geodesic when the process is applied to a sphere. 

While it proved easy to generate other shapes, I wanted to make a sphere. 

At this point it is trivial to generate the plans for laser cutting. As it turns out there are only 4 types of members in a geodesic sphere, which are easily nested manually by arraying. In order to prevent the laser cutter from doing overlapping cuts due to overlapping curves I must recommend the Topologizer plugin for Grasshopper. 

For this project I decided to go with for this project. They are Copenhagen based and offer a number of different materials, various thicknesses. You just upload a pdf, select a sheet size, press order and a few days later your project arrives by mail. 

So far two of these models have been produced. The one pictured below was produced for a friend who added a light source to it, with quite elegant results. 

Design and production of a parametric lamp in Grasshopper

Last Sunday, reinvigorated from extensive relaxation over the holidays, I wanted to create something. I decided upon creating a lamp as they are manageable in size and there are few constraints on their shape.

I have toyed with the idea of creating a lamp using a laser cutter before, but have never really gotten further than the design stage. This time would be different.

I decided upon using Rhino and Grasshopper and designing the lamp parametrically, so that its dimensions could be easily adjusted, depending on the dimensions of the laser cutter and materials available for production. I also wanted the design to be simple as this would be my first time using a laser cutter, but I still wanted something delicate that would be difficult to create by hand.


I started playing with a revolved shape and then cutting that shape with radials from the center to create slices of that shape that could be laser cut. The number of radials would be adjustable depending on the final shape and material thickness.

Overall shape
These radials need something to hold them together so added two central discs and I added slits to both the radial slices and discs.

Zero thickness geometry
At this stage the geometry can be arranged for laser cutting. However, the geometry has no thickness and, depending on the material thickness for production, it could prove to be impossible to assemble. A visualization is required to ensure that there are no internal clashes in the geometry. The planar surfaces can be extruded to give them thickness.

Geometry with thickness
No clashes in risk area


In order to produce the lamp with a laser cutter, the geometry needs to be laid out in 2D so that material is effectively utilized. This process of laying out the geometry is called nesting. In order to do effective nesting, the dimensions of the laser cutter are needed.

There are several options available for people that want to experiment with laser cutting in Denmark. There are commercial services like, where you can upload your design and they produce it for you, for a price (can be quite hefty, this lamp (before optimization) was around 700DKK). Students at DTU can take advantage of the facilities there. For others, non-students like me, there are two fablabs in Copenhagen. One in Nordvest and one in Valby.

Fablabs in Copenhagen, free usage
I chose the one in Valby. They have an Epilog laser cutter with an effective cutting area a bit over 600 x 400mm. It is possible to buy material on site. I bought two plates of 4mm thick MDF for 60DKK.

Once I knew the dimensions of the material I started the nesting process. I used a free nesting algorithm plugin for Grasshopper, Generation. It did not provide perfect results, but good enough so that I could improve them myself. During this process I realized that it could be beneficial to change the geometry of the slices so that the sides would be identical. That way the laser cutter would not have to do a double pass and the cutting time would be halved.

Michael Hviid, the fablab manager was super helpful in helping me picking the right format to use as input to the lamp and giving me instructions on how to use the printer.

Cutting the first batch
The printing process took about 45 minutes. I had not realized how much the process smells.

Assembly of pieces
Once the cutting was completed I assembled the pieces. As can be seen from the images, the edges are completely black and in some places there is a darkening of the sides. This gives the material a nuanced texture, indicative of that it has been laser cut.

I bought a light fixture and a vintage style light bulb to put into the lamp.

The final product

Those interested in understanding how the geometry is derived, or wanting to improve on where I left off can download the files below. I have included the Rhino file, Grasshopper script and the pdf output required for producing the lamp in a cutter that can handle at least 600x400mm.

The robot invasion is coming

3D printing, laser cutting and CNC milling are digital production methods that have become increasingly more widespread in the past years. These technologies have already seen some degree of implementation in construction. Laser cutting and CNC milling are already being used in lots of projects (see for example the beautiful Reindeer Centre Pavillion), although mostly for architectural details and less for production of structural components projects. 3D printing of building components has not yet been used on a large scale, although there are already some promising ventures underway such as the Chinese company that has produced 10 small houses with a concrete 3d printer and Skanska's plans for advanced geometry concrete production.

These production methods are already making the production of advanced geometry construction components cheaper. They are gradually reducing the amount of manpower needed per component and could ultimately far in the future remove the need for manpower altogether.

The Helsingør Motorway: Infraworks model

Digital production in earthworks

I happen to work in a field in which digital production has been implemented to a much larger degree than in building construction, namely Earthworks.

In Denmark, machine guidance and automated machine control has become the standard way of doing earthworks. Most work is carried out on the basis of a digital 3D model. The process may not be as fully automated as the one in a 3D printer or a laser cutter, there is still a middle man between the designer and the finished product. The machine operator is still an integral part of the process - for now.

Not all types of work are equally automated. The level of automation depends on the type of work being carried out and what type of machinery is being used for production. The blades on graders and bulldozers can for example be fully automated requiring little input from the operator, while excavators require more of their operators, who receive cues from a 3D model and machine guidance system to arrived at the correct elevation.

The Robots are coming

Google, Volvo, Tesla and most automotive manufacturers are already well on way to putting self-driving cars on the roads permanently. With time and as the legal environment adapts to unmanned systems, more and more processes will be automated. Construction machinery will join self-driving cars in ridding them selves of operators.

The change is starting to happen within the mining and agriculture fields, as these are ideal to begin the implementation of unmanned systems. Huge areas can be designated to the machines with little risk of inadvertently harming humans. Semi-automated unmanned solutions have already been implemented within the mining and agriculture sector where a single operator can either monitor or remotely control multiple excavators and dumpers.

As automation becomes more widespread, the reliance on 3D models will increase. The unmanned systems will not be as good at filling in the blanks as human operators (at least not initially). It will therefore be necessary to model projects in higher detail in order for a project to be solely produced by unmanned systems.

It is not unreasonable to expect that vehicles such as dumpers will get data from terrain models to optimize the wayfinding and choose a suitable driving style for the terrain. Each unmanned dumper would receive routing orders based on a model of the earthworks, to minimize the kilometer-tonnage, reducing cost and environmental impact.

Like in other fields vulnerable to automation, the move to unmanned systems will require a change in culture. It will require new competences from foremen and project managers on site, not to mention the machine operators. There will be a heavier reliance on models and this will require more of the designers. There will be a need to educate designers and buyers about the increased reliance on 3D models in order to reap the benefit of the unmanned systems. Perhaps this will result in more design-build contracts, by aligning the interests of designers and contractors with the increased reliance on models, which so far, have not been recognized as the contractual basis for projects.

There will be lots of changes, not all of which can be foreseen. We might have 5 years, 10 or 15 years before the robots come, but they will come. It is inevitable. Resistance is futile.