Evaluation of the project

From the beginning of the project, we had a very clear goal compared to other projects. Creating eight crowns for a monument sounded straighforward. Due to this clarity, we could start right away and this helped us a lot in the time planning.

After coming up with a plan, were able to start scanning in the first week and throughout the rest of the project, we spent most of the days really working rather than thinking or waiting. Even though the crowns are not done yet, we have not been wasting time and worked on the project as well as we could.

The three of us worked quite well together as a team. While Rohan is clearly the person with the most knowledge of moulds and 3d models, Vivian and Tom were able to learn a lot and contribute with new ideas or perspectives. Because we did almost everything with the three of us, from scanning to pouring the wax, we all thoroughly understand the process and agree on what is the best method to recreate the crowns.

There was a big difference in experience, Rohan and Vivian having studied design while Tom studied Electrical Engineering. On top of that, Rohan has a lot of hands-on experience due to his work. Even though the differences, we understood each other very quickly and were able to efficiently work together without arguments or a lot of misinterpretations.

During the project we had at least weekly contact with our coach Maaike. This varied from meetings in which we could show our progress to mailing her about what we were planning to do. Maaike really let us decide on what would be the best for the project and with good feedback every time, we could progress really quickly. She was very involved so it felt like we did it together.

Unfortunately, we were not able to work out all ideas equally well. The digital model of the crown, created from scratch was finished too late to be sent to Groningen. Even though we agreed on that it would look too smooth and that the newly cast crowns would lose the historical touch, it would make a good comparison. On top of that, maybe it would regain the historical touch because of the loss of detail due to casting the wax. Looking back, we could have planned it better so we would have it in time.

Personally, I think that the project went really well. The atmosphere in the group has always been really nice and being able to make something quite important with two just as enthusiastic teammates is really satisfying. Where Rohan always was full of ideas and Vivian never afraid to try out new solutions. Even though we still have a way to go to finish the crowns, I am really glad with the results so far.

We are looking forward to finishing the crowns and seeing them on the royal monument.

-The End-

Retouching the print and receiving the metal samples

Currently we are in the final few days before the Science Fair which will be held this Tuesday. On this day all the groups from the Advanced Prototyping minor will present their work. Even though we got a fair amount of prints and models to show at the fair, we are not quite finished yet.

This morning (Monday) we took out the cast-able resin print we started Friday. We were very happy with the quality of the print, but there is still some room for improvement. That is why we will edit it manually.

A full-size cast-able resin print

After cleaning and taking off the support, we used a dremel tool to start cleaning up the edges. Even though we still have a long way to go, the difference is clearly visible and the lines are getting sharper and more detailed.

Using a Dremel tool on the cast-able resin crown

During the cleaning, we noticed an email from Maaike that she received the casts of the samples we sent to the casters in Groningen. We decided to get them as soon as possible, so we took the first train to The Hague.

The metal samples from Groningen

After receiving the metal pieces we noticed straight away that the marks on the back were not visible anymore! Luckily we cut the pieces a little bit different so with some puzzling we could to figure out which one was which. We also saw that all the pieces are very similar to each other. The digital and manual restoring is visible, but the usage of different materials (cast-able resin or wax from the silicon mould) has little to no influence on the result.

This lack of difference tells us that we can use both methods to restore the crowns, and since both methods differ a lot in time and money, we can choose the most suitable one. This means that we will chose wax instead of cast-able resin, because this is much cheaper.

One other observation we can make is that we still have to work on the details. Due to the casting, the crown will lose a little detail. We therefore have to overcompensate for it, which can be done by continue using the Dremel tool on the cast-able resin print.

For now we will continue to prepare for the Science Fair by making posters and other materials to explain our process as clear as possible to the visitors.

Rebuilt from the ground up

One of the methods for restoration that we wanted to explore was completely rebuilding the crown digitally. We did this in Solidworks by reverse engineering what we could from the existing crown. With this method the details will be much clearer than with digital retouching. However, the historical charm could be lost with a digital made model. That is why we will make a model to see how accurate we can get.

We knew the overall dimensions and could work out many of the angles and curves by taking multiple measurements from a datum point (ie the height and distance to centre).

Unfortunately due to the degradation of the existing crown there were a number of elements that we had no reference point for, such as the flower details that are repeated above the collar of the crown. For this we tried to find examples of similar details in other royal crowns but ultimately had to create our own interpretation of the details and so are not sure if it is accurate compared to the original.

This seemed to be the most challenging part of rebuilding the crown digitally so we started with it and once we had the flower details figured out we moved on to bulking out the rest of the crown. Screenshots of the process can be seen below followed by a digital render of the finished crown.

Although the model is very accurate it still looks very clean and modern. That is why we decided not to use this digital model, but continue with the digitally retouched model from before.

Printing full-size models

At the end of last week, we decided to print a full-size version of the crown. This was a challenge on its own, because the crown barely fits in the print area. After rendering the support structures, some of the material appeared outside the boundaries. Because this material would only have an effect on the bottom of the support, we tried to send it to the printer anyway. Luckily, the software was able to send it to the printer even with some parts sticking out and the printer had no problems starting the print.

After 28 hours of printing, the result turned out perfectly, a beautiful see-through model of the crown was hanging in the printer.

A full-size model of the crown, printed in clear resin

We were very happy with the successful print, since this print was not only meant to show at the science fair, but also to test if the full-size crown could be printed with the more expensive cast-able resin.

Removing the clear print from the support

Due to the transparency of the print, small details are hard to see and because some digital and manual modeling still has to be done, we painted the crown in the red and yellow of the original cast-iron crown.

Painting the clear full-size model

After finishing the digital modeling, we wanted to start a cast-able resin print on Thursday. Unfortunately, the program we normally used to make the models hollow, Netfabb, was not up to the task. After waiting for a long time and trying different methods, we decided to use Windows 3D Builder. This worked a lot quicker, but left some unwanted artifacts in the middle

Luckily, this did not matter for us, since the inside is not used in the casting, so we went on with the process. When we finally imported the model in Preform, the program used for preparing Form2 prints, and made it generate the print file, it got stuck at some point. Maybe due to the size or due to the artifacts, the program just could not prepare the print.

Removing detail from the inside of the model. Most detailed is on the left side and the least detail on the right side

On friday we tried to reduce the file size by removing detail on the inside of the model and removed the artifact by hand. After another long loading time, we tried to generate the print file again, and after a little less than an hour, we were able to upload it to the printer!

Now the printer is still busy printing and it will take 1 day and 7 hours to complete. Hopefully we got a cast-able resin print afterwards that we can use to make a silicon mould from and cast the wax models from.

Making the wax models

Today we collected our silicon mould and started to remove the walls. The silicon was firm and smooth so that was good.

Removing the walls from the silicon mould

After removing the bottom of the box we saw that some of the silicon slipped beneath the edges and filled up parts of the mould. Luckily it was very easy to remove those edges with a pair of scissors. The quality of the mould is good enough to make wax moulds with it.

Silicon ould without the wooden box. Some silicon got into the moulds

We bought a cheap pan with a teflon layer to melt the wax in. First we wanted to melt the wax with a heat gun, but we found out that the wax needs to melt slowly and evenly. Then we decided to just use Tom’s furnace to melt the wax with. There was almost no smell while heating the wax.

Heating up the wax

Then we poured the wax in and let it cool down for an hour.

Silicon mould with the wax poured into it

When the wax moulds were remved they looked like this:

Finished wax moulds

We see a little bit of detail loss on the surface, but not too much. The wax moulds are somewhat smoother than the original printed test pieces. We will send the four wax moulds together with the digitally edited cat-able resin piece to the caster this afternoon and hopefully we will recieve them somewhere next week.

Printing the cast-able resin test pieces

On Wednesday, we arranged to change the current grey resin for the newly arrived cast-able resin. This resin is used a lot for investment based jewelry casting. Because we want to use a similar method, the resin should fit our purpose perfectly. When heating cast-able resin, it will melt out of a mould, leaving space to pour liquid metal. We discussed this method in this post.

Another option is to create silicon moulds from these pieces. Hot, liquid wax can then be poured in the silicon moulds and the resulting wax models will be used for the investment or sand-casting. To make the silicon moulds, both the cast-able and non cast-able resin prints can be used, so in order to compare the two types of prints, we decided to print the same pieces as we did before and make a silicon mould of all four pieces.

First, we had to print the cast-able models, so we had to change the resin in the machine. Because Tessa never changed the resin in the Form2 yet, it was a new process to all of us. Being very cautious to not spil a single drop of resin on the clean glass, we managed to take out the printing platform and replacing the resin tank. Once everything was cleaned and changed correctly, we started our print and the tank started filling with the cast-able resin. Due to the orange cover of the machine, it looked like a black material.

The next day, Thursday, we returned to take out the print. When opening the cover, we were very suprised with the beautiful, bright, blue colour.

The two pieces printed in cast-able resin

After taking out the prints and taking off the support material, we noticed a clear difference with the non cast-able resin. The cast-able material is a lot softer, stickier and can even bend a little. Even though we were very careful taking off the support structure, we could not avoid scratching the surface a little.

The cleaned cast-able resin prints

Once the pieces were released from the support, we went to the city centre to meet Maaike and discuss the progress.

Sculpting the 3d models

After scanning the crown on Friday, we took a good look at the result.

Unfortunately, the original crown is quite old and some of the details have been worn off. The following image shows how the loss of details on the ornaments of the crown is clearly visible. Luckily, some sides have more detail than other sides and therefore we know what shape the details have and what we could do to improve them.

The visible loss of detail in the ring of the scanned crown

In addition to the loss of details, the existing paint layer is quite thick and due to the accuracy of the scanner, the lines where the paint was worn off are clearly visible. If the model is printed and cast without smoothening this, the lines will be visible on the cast iron, giving the final product an unwanted finish.

The cracks and paint lines are clearly visible on the model

On top of this, the original structure to mount the crown on the fence is still partially present. Three ‘feet’ were used to stick the crown to the concrete poles. After scanning, the three feet and a bit of concrete is still present in the model. These will have to be removed, since the new way to mount the crown is going to involve a different method.

In order to mount the crown, three feet were added and are still in the scan.

In order to get rid of the paint lines, mounting structure and bring back some of the details, we have a few options during the project. One of the options is to edit the 3D model. Another option would be to retouch the print of the model and depending on the method of casting, the wax can be retouched as well.

We decided to test how hard it would be to digitally retouch the model as this was the most accessible option to begin with. The only problem with this is that most 3d programs do not like the complexity of the model or cannot edit the ‘mesh’ this model consists of. Therefore, we had to search for a program that is able to sculpt 3d models to digitally edit the crown. The following is a list of the programs we used and why they did or did not work for us:

Slicer (Open source: https://www.slicer.org/)

After investigating in a project a few years ago, we read that this project used a CT scan to scan their objects. Using the images, they were able to create a model using Slicer. We hoped that this program also included some modelling options for pre-scanned items, but we were unable to find them.

Meshlab (Open source: http://www.meshlab.net/)

Meshlab is an interesting program for 3d models, since it includes a lot of different filters and options to edit, smoothen and finalise the model. Unfortunately, due to the unfriendly user-interface and the inability to undo actions, simple tasks, such as removing a bump were really difficult.

Geomagic (Wrap) (Trial version: http://www.geomagic.com/en/products/wrap/overview)

Geomagic is a professional program to create scans, edit them and improve models in general. Because this is a professional program, we could only use the trial version. After installing using the “Mesh Doctor”, the limitations appeared. All options were greyed out and you could not save the result.

SculptGL (Open Source: https://stephaneginier.com/sculptgl/)

This web based program was perfect for us. It allowed us to load the object and sculpt it like it is clay by adding, removing and smoothing material. Using this program, we were able to remove the mounts and make the paint layer smoother.

The image below compares the raw scanned crown with the retouch version from SculptGL. Left is before editing and right is after editing. The result is not perfect yet, but the more we can change in the model, the less we have to change once we have a printed version. Bringing back small details is really difficult in the model, so most of the refining has to be done at a later stage.

The comparison between the unaltered model (left) and the edited model (right)