De meningen ge-uit door medewerkers en studenten van de TU Delft en de commentaren die zijn gegeven reflecteren niet perse de mening(en) van de TU Delft. De TU Delft is dan ook niet verantwoordelijk voor de inhoud van hetgeen op de TU Delft weblogs zichtbaar is. Wel vindt de TU Delft het belangrijk - en ook waarde toevoegend - dat medewerkers en studenten op deze, door de TU Delft gefaciliteerde, omgeving hun mening kunnen geven.

Posted in September 2010

Meshing Madness

After yesterday’s scans the next step was to convert these files to printable STL files. Therefore, we used an applications that could convert these files for us. After a lot of trial and error that we experienced over the weekend, this was done in much less time. Viewable three dimensional files were produced very quickly. Because we used the medical CT-scanner (the one with a lesser resolution) the result was slightly dissapointing. But the recieved files were a lot smaller (say, ten times smaller) and can be processed much easier.  We even managed to smoothen the surface, so it doesn’t not look like the previous test glass. This was done by first exporting the files to a different extension then STL, and then loading it into Rhinoceros 4. The results of the scans and the converting process, which we are very satisfied with, are shown below.



After some tweaking and editing it was time to finally export the file to STL. To our dissapointment the result after exporting was very unsatisfying. The now printable glass looked nothing like the original shape. The surface was distorted and edgy. This was not how we imagined how the printed glass would look. After consulting with the 3D printer opperator, he confirmed the print would definately look like what we saw in the modelviewer. This is how the glass would look if we didn’t do something to improve it.


Because what we are attempting is not done before very often, there are not many people that could assist us in this matter. This resulted in again many hours of trial and error. Ing. A. de smit (our Rhinoceros expert) explained us that what we saw in the rendered image was not the real thing, but an interpretation that Rhinoceros made for a better looking render. After almost haven given up, we contacted Dr. Y. Song. With his broad knowledge about situations like this one, he managed to ‘repair’ the file and produced an amazing looking printable three dimensional copy of our glass. From which the result is shown below.
We are very happy with this result and will 3D-print this glass in plastic for further testing. The result will follow tomorrow.

First real scan with old glass work.

Today we made a test scan of three glasses. These glasses are also from the 17th century Henkes collection so we could make a good test scan. We made this scan with the medical CT scanner (Siemens Somaton series) at the faculty of Civil Engineering. After making the scan we discussed what we could do next. We decided to make two types of molds of these scanned cups; a negative mold made by the 3d plaster-printer, and a positive mold made by a 3d-plastic printer. This way we can see if the print material is resistant to porcelain and how the result will look in porcelain.

On Thursday or Friday we have the results of the test with the plaster 3D print. If this works we can continue by making more scans and creating 3D prints. If the result is negative we have to look for another solution. We already brainstormed about coating the 3D print with different materials to see what works the best. First we have to see what the results of the test are before looking further.

Henkes collection

Plastic-positive molds:
The main plan is to create a 100% perfect plastic mold (positive), around this model a plaster mold can be made to produce a 100+ series of porcelain cups. The negative molds will be made out of 3D-plaster to test is porcelain can be molded directly in this material. This type of mold is more of a test and will only produce one porcelain cup.
This afternoon Sisko worked on the 3D files which rolled out of the CT-Scan. Tomorrow morning we will print one glass out of plastic. The plastic printer which will be used is a Objet Eden 260V. This machine works with liquid and UV light and produces pretty high quality surfaces.
This is a demonstration of the printer.

After the plastic molds are made, Maaike Roozenburg will create plaster molds so the first porcelain cup can be molded. The negative molds will be filled directly with porcelain sludge.

3D-Plaster molds:
The negative plaster mold will be printed before Thursday. After a short conversation with Martijn Stellingwerff, we found out the plaster mold, directly out of the 3D-printer, is way to weak and will dissolve if it gets in touch with porcelain. Therefore the first 3d-plaster test molds will be impregnated with all sorts of materials.

Another plan didn’t make it today: Milling a negative mold out of a block of porcelain-mold-plaster. This could be a way to make a durable mold in which more then 100 cups could be made, but the milling machine can’t handle plaster; the powder damages all the components.

Glass photography:
When Boymans & van Beuningen will bring their precious glasswork next week for scanning, we will also photograph the glass cups. We can use these pictures for presenting the project. Today Michel used a beer-glass with a cracle effect. Afther all kinds of extra lightning, the pictures below shows lighting from below, this way only the glass and it’s cracks will be shown on the picture.

What’s next?

Before the next meeting on Thursday, the first molds will be ready for use! Sisko will place an extra blog later on about the 3D-STL files, in this blog Sisko will show a few very positive results.

We’ve done it.

It worked. After even more hours finding out how to combine the slices of de scanned model, after countless more computer crashes, we finally managed to load the entire model into Rhinoceros 4. We now know that when we scan the real 17th century glass we can definately convert it into a printable three dimensional file. What is left now is the process of making the model of the glass into a mold. This is not done by simply turning the object into it’s negative. Figuring out how this will work, is our next task. Again the result of our work is shown below. It is a render made with Rhinoceros 4 from the scanned test glass.



DICOM to STL Breakthrough!

After spending many hours on the internet, we finally managed to convert some of the data of the scanned test glass. The scan of the glass (that was done by the smaller Nanoton) has produced several slices. Because of the high resolution and high accuracy of the scanner it overwhelmed us with well over 1000 slices, resulting in a 3GB DICOM file. Converting this data to a 3D-printer extension (like STL) is not as simple as it sounds. The slices are horizontal scans, pictures you might say, made every fraction of a millimeter. These pictures do not contain any data that represent a coordinate system. Therefore, boundary edges for every slice have to be made and a computer program that can combine these edges has to be used. When this has been done, this program will combine every boundary edge and shape them into the final model. This model (now consisting out of points rather than image data) can be exported in STL format. This can be loaded into a modeling application, like Rhinoceros 4. Again, this is more difficult as it sounds. The 3GB of data, or 1152 slices, is too much for a home pc to handle. Therefore, sections of maybe a 100 slices at a time have to be made. Yesterday, as a test, we managed to successfully convert one single slice into a STL file and loaded this Rhinoceros 4. The result is shown below. Our next step is to try and combine a set number of segments into a complete model.


Raw project planning

In order to make this project succesful, it is important to test this technique with a glass which is more or less identical to the Boymans en van Beuningen glaswork. Therefore we will scan a glass from the Delft archeological archive next monday. After this, the DICOM scan model will be converted into an STL file which can be used to 3d-print the negative mold including runners. When this mold is made, Maaike Roozenburg will make a porcelain model. After this proces it will be clear if 3d printing plaster is good enough for porcelain fabrication. The molds for the inside of the cups (inner mold) will be made out of plastic, so also for this part we need to do some work to get everything clear. If the 3d-print plaster is nog good enough for a negative shape mold, plan B is needed. In plan B, the glasswork will be directly printed in 3d as a replica. After this, a mold of real plaster will be made around this replica.

7 oktober will be the day for scanning the real glaswork. After this, time will be a determing factor because on the end of october, the porcelain cup has to be ready for the Delft science fair.

First introduction with CT-scanners

Yesterday we visited Ing. W. Verwaal at the faculty of Civil Engineering. Ing. W. Verwaal is a specialist in scanning all types of samples like soil, concrete, sandstone and asphalt. However, we will use the CT-scanners this time to scan the 17th century glasswork. Before we can do this we need to understand how the scanning process goes and what file/software we need to use. Ing. Verwaal showed us the scanners and showed us a model of an old glass which was scanned with the small scanner. This small scanner has a higher resolution than the big medical scanner but can’t be used for the real objects later on. That’s because of the fact that the objects needs to be glued on a rotating part for perfect scan results, ofcourse we can’t use glue on the museum pieces!
  We now have a DICOM file of the scanned example. Sisko will try to convert this file into a STL file, which we need. STL files are perfect for 3D-printing.
  Also, the other group members now will try to print a few examples on the 3D-printer. With these examples the behaviour of the plaster/glue can be tested. A possible big problem will be the use of porcelain sludge into a plaster mold made with a 3d printer. The sludge could dissolve the mold. The goal is to 3d print a negative of the glass object. This negative can be used directly as a mold for porcelain.
  It’s important in this project to do all the tests in advance, because the glass from Boymans & van Beuningen will be here at 6/7 oktober and the final product needs to be ready on October 26th (baking Porcelain will take 2 weeks).


Introduction post

On this blog, a project named ‘ Vindplaats Delft’ will be described from start to finish. The final goal will be to show ceramic drinking cups to the public at the final UNIC meeting, which is planned on May 2011 in Delft. These cups will be made with techniques used in rapid prototyping and pottery. The examples which will be used to develop the ceramic version will be glass cups which are unique and were made in the early 17th century. These archeological pieces are now part of the Boymans & van Beuningen collection.


Delft is more than 750 years old. The city owes its name to the world ‘delving’, digging the oldest canal, the Oude Delft (digged in 1100). As early as 1355 the city reached the size it would remain until the 19th century. The glass objects were found on locations like the Oude Delft and the Pieterstraat. Designer Maaike Roozenburg wants to use these pre-industrial cups with a rich history as a starting point for a new design, this way the old heritage can be used for daily use.


Maaike Roozenburg will cooperate with TU Delft students to create a low-tech product, with high-tech prototyping processes like CT-Scanning/3D scanning and 3D-printing. Eventually Maaike Roozenburg will use her porcelain expertise to create the final product in porcelain named ‘Vindplaats Delft’. This product will be supplied in a casket with information about the origin, create history and founding location.
This project will combine old crafts and high-tech technology together. It combines all the qualities which resemble Delft; rich history, traditional pottery and TU Delft knowledge and experience.



The group of TU Delft students are following the minor Advanced Prototyping, Vindplaats Delft is a project for ‘Augmenting prototypes’ under supervision of Ir. Jouke Verlinden. The group of five students consists of four Industrial Design students and one Architecture student.



Architecture bachelor

Sisko Roosenboom


Industrial Design bachelor
Elsa Noske
Astrid van Til
Rob Scharff
Michel Haak

This project is a cooperation between: Studio Maaike Roozenburg, Gemeente Delft, Working Group Delft ceramic City, TU Delft faculty of Industrial Design and Galerie Terra Delft.


This project is coupled to the UNIC congress(Urban Network for Innovation in Ceramics), a project of the EU.

© 2011 TU Delft