top of page

Case Study: Testing different Support Structures on a 3D model with Thick Walls

Updated: Aug 6, 2020

In this blog, I share the tests I did to determine which support structures would work best when 3D printing a thick flat 3D model with resin on a Gizmo 3D printer.

The model I printed is the Hands-free door handle (Coronavirus prevention) by ippe on Thingiverse. You can download the STL here:

For support generation, I used the following software: 1. MiiUtility version 4. Get it here: 2. Zenith. Get it here:

To check for or solve errors in my 3D model after generating supports, I used the following software: 1. Netfabb. Get it here: 2. Windows 3D Builder. Get it here:

Test 1: In the first test, I created supports in MiiUtility with a top cone of 133% and a radius of 0.25mm, mid cone of 133%, and a bottom size of 1.1mm. I printed the part using the hybrid-continuous printing method and a display time of 3200ms. The following screenshots I took from MiiUtility show the settings I used for this test. Each screenshot is of a different set of settings in the software.

Print result and observations:

As you can see looking at the pictures above (click on the pictures to enlarge them), many of the supports didn't print or broke loose from the print, causing warping in some regions of the model.

At this stage, I didn't know why the supports didn't print or break loose, but I suspected I made the supports too thin for this projection area, or the display time might have been too low.

I also knew that printing a solid piece generates a lot of heat, which is the number 1 cause for warping.

The possible solution for this would be to angle the part differently when printing. The video below is a snippet for the part being printed using the Hybrid-continuous printing method.

Test 2:

Using MiiUtility, I changed the support specifications of the top cone from 133% to 200%. I noticed some of the supports overlapped, which would cause inconsistencies and gaps in the supports of the final 3D print. To prevent this, I uploaded the model to the software Windows 3D Builder to fix it. I always run a model through Windows 3D Builder to check for errors in the model. If a model doesn't show any errors in Windows 3D Builder and errors do show up in Gizmetor (when I slice the model), I run it through a second option called Netfabb.

Print result and observations:

When I compared the previous print with this new print that is pictured above, the surface area appeared to have fewer bumps or roughness. I believe this is thanks to more supports having printed fully - although some were still incomplete.

This can be fixed by increasing the display time, increasing the size of the supports, or making the projection area smaller so that we can print greater detail. I decided to keep the projection area and display time the same for the next test, but to increase the size of the supports.

Test 3: As I wanted to make the supports thicker and stronger for this test, I decided to create it in the software Zenith instead of MiiUtility this time. Supports created in MiiUtility are independent, while Zenith supports are connected and thus supportive of each other, making them stronger. Click on the picture of Zenith supports below to enlarge it - you will see the supports are connected to each other as mentioned.

I created the supports with an end point size of 0.9mm% and a thickness of 1.1mm and a bottom size of 2.0mm. The density of the supports was set to 1.1.​

I usually use Zenith supports for large 3D prints. I find the thicker and stronger supports from Zenith also ensure that large prints don't sway in the resin, which is especially useful if you're printing with flexible resin.

Being thicker and connected, support structures created with Zenith can be a little more expensive, but supports from MiiUtilities will probably end up costing the same when they are made thicker and print fully.

In this case, I fixed the model using Windows 3D builder. Sometimes Windows 3D Builder will not see that there is an error in the model, in which case I would run it through Netfabb too just to be sure.

Print result and observations:

As you can see from the picture above, the print didn't warp as much because the Zenith supports all printed. The display time was still set to 3200ms as it was with Test 1 and 2. When the model was printing, it seemed the layers didn't have time to set completely, causing some the layers to curl a bit. This is also why there is heavy layering in the print. The following photo shows the prints from Test 1, 2 and 3 - it's easy to see by comparing them how much the print has already improved.

Test 4:

This time I increased the display time to 5200ms when printing it again.

Print result and observations:

As pictured above, using a longer display time resulted in less layering, which looked better. Unfortunately the warping in some areas of the print caused some supports to detach from the build plate. The extra display time also made the print much stronger physically vs. the 3200ms print, but it generated a lot more heat, which caused the warping.

At this stage, I wasn't concerned with having visible layers in the final print as my focus was testing what supports would work best for this model as fast as possible. The layering is to be expected when printing thick parts using the hybrid continuous 3D printing method at such a low angle.

It's best to print thick parts like this using the standard speed printing method as it gives the resin more time to flow where it needs to before printing the next layer. I also recommend rotating the model to print it at a higher angle as it will reduce the amount of heat generated and result in less visible layering. Excessive heat causes warping during the printing process.

Using the wiper add-on will reduce the heat even more because it moves some of the hot resin away as it sweeps over the part being printed and coats it with cooler resin. The following pictures show the heat of the resin in the area being printed before (pictured left) and after (pictured right) the wiper swept over the build area.

As a 100-micron layer hardens during the printing process (when its image is displayed onto the resin), the temperature of the resin in the area heats up to over 60 degrees C. That temperature drops quickly to 44 C after the wiper passes over it.

Reducing the layer height from 100 micron to 50 micron will also help to reduce the heat because less resin then needs to be solidified.

Test 5:

The display time was changed from 5200ms down to 4200ms.

Print result and observations:

This print (pictured above) appeared to be just as strong as the one from Test 4 that I printed with a display time of 5200ms. It seems, in this case that the resin reaches a certain level of strength at some point, which doesn't increase much more even when printing it with a longer display time.

The print didn't pull away from the supports, and much less layering was visible. I'm pleased with the result - the print looks great. The print time without the wiper was 53 minutes when I did Test 3. This time, the print took 3 hours 38 minutes to print with the wiper. We can print 4 at the same time in this projection area of 200mm x 113mm. If we increase the print angle, we can probably fit two more prints in, but then we probably also don't need to use the wiper.


My tests showed the best supports for this print were Zenith supports with an end point size of 0.9mm%, a thickness of 1.1mm, and a bottom size of 2.0mm. The density of the supports was set to 1.1. These tests were also useful for showing how good prints can look when printed using the wiper add-on. Every print is different. Let's have one last good look at the print from the last test:

Advanced Notes:

  1. Zenith and MiiUtilities don't use a boolean union when creating the supports. This means Gizmetor will have some errors when slicing.

  2. Gizmo 3D Printers Pty Ltd will release an active resin cooler add-on for our machines in the future that will further help reduce the likeliness of warping in a print.

Most prints with thin walls don't require the use of a wiper or thick supports.

Liked this blog? Save it to Pinterest by clicking on the image below.

227 views0 comments
bottom of page