Clear Filament Comparison

Jamie from sent me some samples of his Crysta-Line Clear and Ruby Red translucent ABS filament to try out.  So far translucent ABS has been rare, although translucent PLA is common.  Clear has popped up in a couple of places, but 3DPrinterGear is the only supplier who has colors, currently on pre-order.  I was interested in comparing it to the clear from Repraper that I reviewed earlier.

Crysta-Line is on the left, and RepRaper is on the right.  The Crysta-Line filament ranged from around 1.75mm to 1.73mm, so I was pretty comfortable using 1.75 as the filament diameter in my slicer settings.  The RepRaper filament ranges from 1.63 to 1.73, but mostly varies around .05mm.  You can feel some waviness running a couple of fingers along it.  The RepRaper filament has bubbles visible in it, which are completely contained. The Crystal-Line Clear didn’t have any bubbles but some were visible in the Red, though they were much smaller than the bubbles in the RepRaper clear.

I found that the Crysta-Line Clear is a bit brittle.  While untangling the coil and trying to determine how much I had left, it kept breaking into shorter lengths.  This wouldn’t be a problem with a spool, but a coil would need to be handled very carefully.  The Ruby Red didn’t seem to have this problem.

UPDATE: Jamie said that whenever he tried to break the filament, it took a lot of bends to fatigue it enough to do so.  I went back and tried to break a piece and found the same thing.  Then I tried what I had done before, which was measure it by stretching it out and counting arm spans.  After I stretched it, it would snap cleanly on a moderate bend.  I would then try to snap it at another nearby point, and it wouldn’t snap until I stretched it again.  It seems that moderately stretching the filament will weaken it for a few seconds, so if you need to measure out a portion, go slowly.

I printed some single walled cubes with a thread width that came out to about .38mm.

Here is Crysta-Line at .3mm layers-

The RepRaper Clear at .3mm layers-

I’m not seeing a clear difference between them, and the bubbles I saw in the first RepRaper print I did aren’t evident here.  These tests were extruded at 210 rather than the 195 I used for the first review, and that might have made a difference.  I went with a high temperature because I wanted to see if it affected the clarity (it didn’t appear to).

Here are both clear filaments at .1mm layers, Crysta-Line on the left-

Crysta-Line .1mm-

Crysta-Line Ruby Red in .1mm

Red in a .3mm cup.  If you look closely, there are bubbles visible in the print, though they did not appear in the .1mm print.  Maybe if the bubbles are larger than the layer height, they won’t show in the print.  The walls of the cup are thicker due to the filament being bents into a curve.  They are around .6mm thick.

All three filaments as cups, Crysta-Line left, RepRaper Center, Crysta-Line Red Right-

All three cups at .1mm layers.  Again, no bubbles visible at this resolution.
Crysta-Line Clear at .1mm

Crysta-Line Clear at .3mm

My daughter had asked me for a dining room table to go with her miniature LaLaLoopsy dolls, so I thought the red would look really nice for that, especially since it leans a bit toward pink-

The main reason I am so interested in translucent color in ABS, when they are so available in PLA, is that ABS can be solvent smoothed.  Here is the table after a couple of dips in the vapor bath-

Here are some before and after pairs of some of the test pieces after a couple of dips in the vapor bath
It would be interesting to see what the results would be like with a little pre-sanding, ideally from both sides.  If you needed some clear windows, I think you might get good results from printing 2-3 layers flat on the build platform at .3mm thick and wide threads, then sanding and vapor smoothing them.

Switching translucent color filaments between layers of a print can create some interesting effects too.  Check out the experiments one Thingiverse user did with PLA, to get an idea of the possibilities-

I’m working on a clear acrylic enclosure for my Solidoodle, and I think re-printing accessories like the fan duct and thumbscrews in clear or a translucent accent color would look pretty cool.


Gcode Controlled Extruder Fan

Printing small features can be challenging .  Sometimes a layer is so small and takes so little time to print that it is still melted and soft when the next layer goes on.  The plastic gets pushed around, and quality suffers. Another problem is overhanging perimeters.  Sometimes loops that are only partially supported curl up as they cool.  Later after they have hardened, the nozzle can bump into them with enough force to break fragile pieces, or knock them from the bed.

One solution is slow-down cooling.  You set a minimum time per layer, and the extruder will slow down enough to take that much time to print, giving the plastic more time to cool.  This can dramatically increase print times depending on the model.

The other way to cool the plastic is with a fan blowing into a duct that directs the air out through a ring around the nozzle.  This cools the plastic as it is extruded.  Slow down cooling may still be needed as well for very small featuers, but you can turn the minimum time down to something like 5 seconds when a fan is there to help out.

You can find a fan duct for the Solidoodle at Thingiverse – .

With ABS, keeping a fan on all of the time might not be a good idea.  The layers need to be hot to bond together, and if they are cooled too quickly the print may turn out very weak.  It’s best to save the fan for the very small layers, and also unsupported bridges.  If a thread needs to be pulled over open space, cooling with the fan can prevent it from sagging.  Slic3r and Cura can both turn on fans for layers under a minimum time, and Slic3r can also run a fan during bridges.  To make this happen however, you need to be able to connect the fan in a way that allows it to be controlled by Gcode.  This capability is built into RAMPS and Azteeg controllers, but not the Sanguinololu.  It can be added, however.

I discovered this from Nophead’s blog post –

You will need a Logic Level Mosfet such as this one –

You will need to add some headers to the Sanguinololu.  This isn’t too difficult, but if you haven’t soldered anything to a PCB before I recommend these tutorials- Part 1 Part 2

The host is a bit eccentric, but the info is good.  I had a go at it myself with a $6 solder iron and whatever solder and had a hard time of it.  Then I watched the videos and followed his recommendation of a solder station, a chisel tip and the proper solder, and found that it was quite easy.  If you are inclined to modify your printer to do great things, then some equipment and basic soldering skills are a good investment, and nothing to be afraid of.

One more thing you will need is a solder sucker, which is just a bulb with a tip.  The holes for the headers are filled with solder which will need to be removed.  Heat the solder from one side of the board and hold the sucker against the other side.  When the solder melts, suck it up with the bulb.  This would be easiest with a pointed tip that will fit through the hole.  This way you can push the melted solder through and suck it out at the same time.  Do this a couple of times from both directions to really get it cleared out.

You will need some straight headers, like .  Any self respecting electronics store should have them (which rules out Radio Shack).

At minimum you will need to solder headers into the four pins at the left end of the expansion area.  I recommend you solder all of the pins while you are at it.  At least add headers to the pair at the far right.  This will allow you to plug in an SD card reader, should you choose to buy one.

Note the shaky work here due to diving in without the proper equipment or education.  Still functional, if not a little ugly.  You will need a couple of connectors like these,%20Interconnects&WT.medium=cpc&WT.campaign=Connectors,%20Interconnects&WT.content=text&WT.srch=1&type=Phrase&WT.source=google&cshift_ck=452CDBE6-2610-4AC2-BEEE-57A4932EE148csDGEoepNp

This photo from shows where the MOSFET will connect to the board-

Solder some wires onto the legs of the MOSFET, with some shrinkwrap tubing at the ready-

The right leg is Source and it will go to GND on the board.  The middle leg is Drain and will go to the black wire on the fan.  The left leg is the Gate and will go to PWM B12 on the board.  The red wire from the fan goes directly to 12v on the board.

You can switch the fan on and off by manually entering a M106 code in Pronterface, such as M106 S255 where S is the speed, expressed as a number from 0 to 255.  Repetier host has a button and a slider to control the fan.

The fan may not already be activated in firmware, so follow the directions here for how to download and update the firmware.  When you have the firmware open in Arduino, change to the Pins.H tab.  This will probably be too far down to appear at the top of the screen, but on the right side there will be an arrow that pulls down the full list.  Scroll about halfway down, looking for the section for Sangiunololu.  In that section look for

#define FAN_PIN            -1

and change the -1 to 4.

DIY Smoothing Station


The most effective way to smooth the surface of an ABS print is with acetone.  Acetone will melt the plastic and let the surface flow and even out, eliminating layer lines.  It can be a little challenging however.  If you dip the model, you can get bumps where the acetone drips off.  If you submerge it too long, the acetone can get inside and quickly melt all the fill,  causing the model to collapse or remain soft for a couple of days. You can brush it on and easily create a glossy finish (with glossy layer lines).  Melting it enough to remove the layer lines can be difficult without leaving brush marks.

The most effective method I have found is with vapor.  Stratasys has a smoothing station for finishing ABS parts from their pro level printers that uses this method.  The acetone (or other solvent) is heated until it boils creating a pool of heavier-than-air vapor.  The part then gets lowered into the vapor for 30-40 seconds.  The warm solvent vapor condenses on the  room temperature part, coating it more evenly than a brush can, and thinly enough to be controlled.

If the idea of boiling acetone seems scary, you can try this at room temperature with about 1/2 inch of acetone at the bottom of an empty paint can.  Simply place the model on some kind of metal stand in the can and leave it for a few hours, hoping you don’t leave it too long.  This is the first approach I tried.

This was the result of an hour or so in the can.  The plastic has become glossy, but there isn’t any noticeable melting.

This was about 3 hours later.  You can see that it is almost completely smooth at the bottom, but the lines are clearly visible at the top.  It is sitting on an overturned 2oz cup, which places it about 1″ above the acetone.  There is clearly a significant difference in the density of the vapor between the bottom and top of the model.

After another four hours, the model is completely smooth.  This was allowed to dry out between steps.  The first test had been a Yoda left in overnight, and it melted so much the head fell off.  Smoothing can be done this way, but it takes a very long time, and there will be a much greater loss of detail at the bottom.

I decided the best results will come from boiling the acetone and creating a pool of vapor dense enough and deep enough to allow exposures short enough to judge and control the amount of smoothing.  Stratasys’s smoothing station is based on vapor degreasers, which use solvent vapor to clean electronics and other parts.  They include a pool of heated solvent, a chamber to hold the vapors, and a condenser coil along the top.  The condenser creates a layer of cool air which will condense and contain any vapor that rises into it.

Here is a homemade version of one-

There is a steel pot placed in a larger tub.  Acetone goes into the steel pot, where it is heated by boiling (or very hot) water poured into the outer tub.  Acetone boils at 134F, so not much heat is needed.  In this photo there is a small pot placed into the acetone which provides a place for the model that is lower than the surface of the solvent.  Using something like a steamer rack above the acetone would work well too.  Next to the pot is a tub of ice water with a fountain pump running water through a copper coil.  It would have been neater with a smaller diameter pipe, but this provided an easy connection to the pump.

Much is made about the flammability of acetone vapor.  With this method, the only nearby ignition source would be the pump, which is submerged.  While it is setup outside to provide maximum ventilation, I still recommend using a respirator while working with it.

This method works best on models printed with .2mm or .1mm layers.  Some sanding in advance is a good idea as well.  The less the surface has to be melted to remove the layer lines, the more detail and edges will be preserved.  Set or dangle the model in the vapor for about 30 seconds at a time.  When you pull it out, there will be some melting already visible.  It will continue to smooth as it dries, so it might take 20 minutes before the effect can really be judged.  Let it dry for 30 minutes or so to avoid over melting.  Two 30 second exposures will often be enough to completely smooth a model.

Before smoothing, sand down any blobs, bumps or strings because those will still remain as bumps on the smoothed surface.


Tumbler Finishing

I’ve been wanting to try out finishing printed parts in a tumbler. This is a polishing process often used for things like rocks and metal parts, where the part is placed in a barrel with some media and rotated for a few hours. The media consists of chunks of some kind of material, like ceramic, plastic, or walnut shells. During tumbling the media abrades the part, polishing it.

I’ve heard mixed reports of the effectiveness of this for ABS and PLA, and have been reluctant to spend $150-$200 for a small tumbler than might not even work that well. Then, thanks to some YouTube links, I realized I already had a large tumbler next to the washing machine.

I went to the local Harbor Freight and bought a 5lb jar of ceramic media. Ceramic is recommended for aggressive deburring of metal. Plastic is much softer than metal, but removing the layer lines requires removing more material than you would usually want if you were just polishing. I also bought a 2 gallon barrel and a cheap doormat. I cut the mat and used it to line the barrel so the ceramic wouldn’t make so much noise rattling against the plastic barrel.

To test it, I used a ring that I had printed at .1mm layers. It is about 2mm thick, 40mm high and 70mm across. I wrapped the barrel in pillows and quilts, and placed it in the middle of the dryer. You can position it so that it rolls, or turns end over end. End over end will provide more aggressive action if you feel like you would need it.

I set the dryer to Air Fluff (no heat) and set the timer as high as it would go, which is only 1:20:00.  I ran it enough times to total about 8 hours of tumbling.

It came out smoother than I expected.  It has a matte finish, and a feel that isn’t exactly soft, but very friendly to the touch.  The layer lines are only visible if you look hard for them.    It would take only a quick exposure to acetone to smooth it the rest of the way, and if you still wanted a matte finish rather than glossy, an hour or so in the tumbler would be enough to take the shine off.

The downside to this method is that it will round off sharp edges.  You might want this effect for instance if your part is something that will be held and handled.  Also it works better with high resolution prints.  The one below is .3mm, and was only in for about 2:40.  The edges are already rounded, and the dust is getting caught in the layer lines.  It might be able to reach the same finish as the ring, but take much longer.


A vibratory tumbler should be more gentle, though maybe less effective because of it.  However it would be easier to leave on for long stretches like overnight.  These results give me a little more confidence in the thought of ordering one such as this one from Harbor Freight –  I think it would definitely be good for preparing prints for acetone vapor finishing (next post).

Alternative Print Surface

I’ve always had trouble with ripping my Kapton tape. Prints are can be removed easily, without damaging the tape if the print bed cools down, but it takes so long to heat up, I hate to let it cool too much. Also, if I want to cancel a print and restart it right away with tweaked settings, I’m too impatient to cool it down first.

The solution is to forget the Kapton altogether. I went to a frame store and had a two pieces of glass cut to 6″ (a hardware store might be able to to this for you). I used small binder clips to hold it to the aluminum bed and heated it up. Originally I put Kapton on the glass, which was a lot easier to do outside the printer. However someone posted a tip on Thingiverse to use hairspray, and I found using that on the glass without Kapton worked great for ABS.

The great thing about using glass, is that when the print is done I can take the glass off and let it cool on the table without needing to turn off the heat. When I put a piece of glass back on to the bed, it only takes about a minute to warm up again. One thing to be aware of is the Z offset screw. My glass was about 3mm thick, so I needed to turn the screw down that far to keep the glass from hitting the nozzle when homing Z. If your screw doesn’t have that much space available at the top, you might need to get a longer one.

The problem with covering the aluminum with glass is that you can’t get to the leveling screws.  This is easily solved by printing the thumbscrews here –  Just drop a little superglue inside and twist them on, letting the bolts cut the threads.  The glue will hold them well enough to turn the bolts, but if you need to take them off, you can still break the bond.