Compensating for ABS Shrinkage

Summary: Quantify your ABS shrinkage rates and scale your model to compensate...maybe.

ABS is known to contract or shrink as it cools. The amount of shrinkage depends on the quality of your filament as well as your printer settings and capabilities. Heated build platforms and enclosed print chambers can minimize these shrinkage problems that can lead to warping in ABS parts. I never noticed much of a problem with the final size of my ABS parts until I started printing small holes. I was designing a bunch of 3 mm holes into my parts and each needed to be drilled out after printing to get the hardware to fit.

So I decided to print a bunch of ABS pieces and compare the CAD dimensions to the final printed dimensions. I used MatterHackers Red Pro Series ABS filament for these prints. All prints were with 0.2 mm layers at 230°C with 90°C bed temperature, 50% fan, and 107% flow. I also used the brim option for adhesion. I started with cubes 5, 10, 15, and 20 mm square.

test prints

When the prints were finished, I carefully measured the length, width and height of each cube with a set of precision calipers. Ironically, all dimensions actually increased except for two measurements. There doesn't appear to be any correlation between the size of the print and the amount of error. Note: I tried carefully to measure at points on the cube that were clean, avoiding any print fragments or over-extrusions.

shrinkage of cubes

Next I wanted to test hole sizes. I designed a 3 mm thick card with 5, 10, 15, and 20 mm diameter holes in it. It was printed using the same settings as above and measured with the calipers. In all cases, the diameter of the holes were smaller than designed. And this time, there was a clear decrease in print error as the size of the hole increased.

shrinkage of holes

As it turns out, it looks like shrinkage isn't as much of a problem as I thought it was. In fact the cubes were slightly oversized. Perhaps it is the quality of the filament. Or maybe the files I printed are accidentally designed in a way that minimizes shrinkage. Or does the extra flow at the nozzle make up for any shrinkage? I don't know, but so far the dimensions of ABS parts from the UM2 are almost perfect. But, if you do find you are having problems with final dimensions, you can always repeat this test and scale your final prints. As for the holes, you may want to oversize your small holes accordingly or simply drill them out to their final size.

for perspective, this is the average error in the cubes

Ratchet and Pawl Mechanism

servo-driven ratchet and pawl

I've been working on a project that requires some precision motion control, rotating an object 360 degrees in 12 steps. At first, I tried using a stepper motor but it drew too much power, ran hot, and was too heavy.  Worst of all, it wasn't strong enough to turn my load. A servo would be a better solution but most servos are only capable of 180 degrees of motion. How do you get 360 degrees of rotation from the limited throw of a servo? A ratchet and pawl looked like it might work.

drafting the ratchet

I looked online for an existing design but I couldn't find anything that would work for me. So I drew one up in 123D Design. It was pretty easy; simply draw two circles to define the top and bottom of your teeth. Sketch one line from the center point to the outer edge of the circle. Then use the Circular Pattern tool to replicate that line around the circle as many times as you need (once for each tooth). Then use the 3 Point Arc tool to draw the back edge of the tooth and replicate that arc just as you did the lines. Lastly, just extrude and remove any unwanted parts. The arm and catch aren't nearly as critical, just sketch them to fit the profile of the teeth.

the finished design

With the design finished, I printed the parts in ABS at 230°C with a brim. Since the appearance isn't critical, I printed with a 0.2 mm layer height at 80 mm/s. I mounted the pieces to a scrap piece of lexan and used some small springs to tension the pawl and arm against the ratchet. Using an online calculator I determined that for a 30° arc, the servo arm needed to be about 19 mm long to give me the necessary linear motion from the ratchet arm.

servo connected to arduino

I used an arduino to test the mechanism. I simply modified the Sweep code from the Servo library and shortened the sweep from 180° to 30°. To my surprise, it worked on the first try. I modified the code a few degrees to dial the servo motion in perfectly. This is a nifty little design that should come in handy in the future. Next I'll print a hub to mount this to the part I need to rotate.

Files are available on Thingiverse and YouMagine.

Printing Multiple ABS Parts

multiple ABS parts printed "all at once"

Summary: Print ABS parts one at a time, not "all at once".

Lately I have been printing a lot in ABS to make some custom parts for my multirotor helicopter. When I need to print several copies of the same part, I have been using the "print all at once" option just as I did with PLA. With PLA, this setting allows time for each layer to cool and generally gives higher quality prints.

layer separation

However with ABS, I noticed some layer separation in my prints with printing with the "print all at once" setting. There were obvious gaps in some of the layers. With just a little pressure, the layers pulled apart very easily.

layers pulled apart

I think this is due to the fact that ABS shrinks some as it cools. Print continuously and the previous layer is still hot while the next layer is printed. But given extra time to cool, the previous layer shrinks before the next layer is printed and you end up with the layers not touching.

Luckily, print quality has not been an issue; it seems easier to get good prints with ABS. So the print all at once option does not seem to be necessary. I have been printing my ABS parts one at a time and the quality and strength are perfect.

First Prints in ABS

ABS print

Summary: Printing in ABS is easy. Lower default nozzle temperature for best print quality.

I have a number of parts to make for my custom hexacopter. I needed to print them in something tougher than PLA that can hold up to some hard landings and withstand high temperatures like in a car. So I ordered some Pro Series ABS filament from MatterHackers. If you read up on printing with ABS you can easily psyche yourself out. So much advice on getting successful prints with ABS. I chose to ignore all of it and just print the Ultimaker Robot with the default ABS settings at normal quality.

normal quality print at 260 C

The default ABS settings are 260°C nozzle, 90°C bed, 50% fan, 2.85 mm filament diameter and 107% flow. I used glue on the bed to help the part stick. The part printed without any problem and stuck to the platform firmly. Print quality was pretty good but it looked like it printed a little hot. The color also looked to go a little light and matte compared to the bright gloss red of the filament. The first few layers printed a little messy with nubs of material, same for some of the infill. The small details like the antennae and ears looked melted and the bridges looked bad. The Pro Series ABS is rated for 230°C - 240°C so I decided to print again at 240°C. This time I adjusted the filament diameter to 2.90 mm.

260C print (L) vs 240C print (R)

This 240°C print turned out a little better. The initial layers printed clean as did the infill. The color of the part also got brighter. The fine details printed much better and the bridges improved slightly. On another note, many people complain about the smell of ABS during printing. While it does have a faint smell of melting plastic, I found it underwhelming. It was only really noticeable when close to the printer; not a problem. Let's try one more print at 230°C.

240C print (L) vs 230C print (R)

At 230°C the small details and bridges improved a little more. Most noticeable, the sides of the print were smoother. At this point, the ABS prints nearly as well as PLA. Larger prints may not be so easy, bed adhesion and warping may become an issue, we'll see. But for now I'm ready to make some parts in ABS.

ABS (L) vs PLA (R) prints