Wednesday, March 18, 2009


ABS Prints

I finally managed to get hold of a roll of ABS from Ian, he has been printing ABS for some time but I have only done a few test prints.

To run the higher temperatures I have made a new hot end, the old one had PVC wire insulation in the hot zone. It also uses PTFE insulator with backup ring to stop it pushing out of shape with the high temp.

I have also experimented with the finish and material of the print table to see if I could get away with using MDF. I found there was good adhesion for the first one or two prints then you have no chance. The ABS seems to stick to a fresh clean surface but when the first print gets peeled off, so does the surface finish. Options then were to turn the board to a fresh spot for the next print, this works but is tedious. I also tried taking an “old” board and resealed the surface. I tried PVA adhesive, a very thin layer (undiluted) all over the surface with the excess wiped off. This worked very well, probably better adhesion than untreated MDF. But, the first print peels off the surface again.

I have now switched the board to Acrylic, this seems to be working, the first few prints were very difficult to remove, but now it seems to be settling down, I will let you know if this is a long term solution. I should also say that I am very late getting into ABS, have any others tried Acrylic for more than a couple of prints? The same spot on the Acrylic has been used about 8 times.

Note, on the new board I ran the first prints at 200C and these were very hard to get off, now the surface is matt finish I can run the first layer at 225 to 230C.

ABS boards are always an option but price is an issue plus we are already cutting acrylic so to run another board is very easy.

Print Temp is 235DegC measured at the tip, print speed/extrusion speed – 16mm/sec
No raft, continuous extruder running.
(V3 BfB machine with PIC32/SD control)

Its still early days with ABS, these first items show promising results, I need to get some more material through the machine to fine tune the settings and explore higher speeds.
The new V3 machine is capable of running 64mm/sec, Ian has printed at this speed, quality definitely suffers but the machine and software keep up. I don't intend to go anywhere near to 64mm/sec it is extremely hard on the machine. 24 to 32mm/sec would seem a more realistic maximum, I will post results as time permits.


Whew! VERY nice print quality!
where do i go to get ahold of that wineglass model?

also, i'd love to look over the firmware. is it in subversion?

one more question: what are you using to generate the gcode with no extruder on/off toggles? are you using skeinforge?
Yes, that looks great!

As for the acrylic bed durability, I printed about 100 parts on a 1/4 inch sheet of acrylic. I found that it cracked pretty much everywhere from the thermal stress, but the final limit was the point at which the acrylic sheet was too warped to get a reliable raft laid down.

Warping became an issue on the larger, flat parts, but I haven't yet got around to building a heated build chamber.

Thanks again for the comments.

Wade, good feedback, I think I will stick with the Acrylic board and see how it fairs. Did you have issues with the print being difficult to remove from the new board?
Do you think if the surface is lightly sanded to remove the polished finish it would help reduce the adhesion?
And what temperature are you putting down the first layer?

Sorry for all the questions!

Zach, the wine glass is attached, see how you get on with it. Maybe you could add it to the parts repository for me.
I use skienforge for all the G-Code files but when I interpret the code I do not switch the extruder off. I find that once it has been turned off there is a lag in getting reliable flow going again. I know this can be compensated for in skienforge but when I tried it it did not seem to work as well as running it constantly. If you look back through my blogs I have never turned the extruder off during a print and made a point of mentioning this in the notes. For me its less of an issue to deal with a few hairs than a void in the print. Give it a try and see if you get the same result. I have a feeling that Chris runs the extruder all the time, maybe he can confirm this.

Regarding the firmware, see the previous blog, nothing has changed since then.

Hi Tony,
If you mean me, then no I don't leave the extruder on.

I turn it off at the end of the thread and note the time. At the start of next thread I start the extruder and pause for an amount of time which is an exponential function of how long it was off for.

Not a perfect solution by any means and my current extruder oozes quite badly. I plan to do more work on this when I have an extruder that can be reversed.
Chris, sorry I should have used your tag'. Yes, I thought I had read that you tried running the extruder continuously, but thanks for the comments its interesting to know how you got round the problem.

I had two issues when I stopped the extruder, first off was getting a reliable start, you are using a more sophisticated delay algorithm than I tried.
The second issue was the mess that oozes out is dragged across the object sometimes causing damage, at other times messy stringing. I don't pretend running continuously is a perfect solution either, it just worked a little better than my on/off delay.

The new extruder design we are running uses a stepper motor so reversal is something Ian has had a look at. His first bench test seems positive, reversal for a few steps halts the flow immediately. We have yet to shoehorn this into the firmware to see if it returns a better result, I will feedback our findings.

Your print results are extraordinary, especially the prints using the 0.3 nozzle. Does your on/off method banish all the hairs on the print? Also the material that does ooze out, how do you handle it, a nose wipe or does it get knocked off as the head re-enters the print?

I am confused, how does not turning it off cause less ooze?

I turn it off during head moves. It still leaves lots of string, but it is much finer than leaving it on, where it would be a full thickness filament rather than a fine hair.

When I turn it on at the start of the next track the delay is long enough to get the filament flowing again. I tend to have it a bit on the long side as it you can cut off a blob, but you cant replace missing material.

The ooze either gets pulled into a hair or smeared very thin on the top of the object (I don't lift the head), so I don't need to wipe the nozzle after a move.

I do wipe the nozzle when I start a layer after a warm up or cool down, because in that case there will be some cold oozed plastic that will not stick.
Actually leaving it on would not lead to a full width filament because moves are faster than extrusions of course.

On HydraRaptor they are only twice as fast, but on Darwin I expect to be able to get ~ 10 times as fast, so maybe not turning off would be better in that case.
I didnt mean to give the impression that leaving the extruder on causes less ooze. Leaving it on maintains full flow ready for a reliable touchdown.

The excess material is as you surmise pulled very thin with a rapid head move.

I have used values between 5:1 up to just over 10:1 for rapid head moves, the higher end of the scale gets rid of virtually all stringing but is harder on the machine. The lower end of the scale reduces stringing to an acceptable level without the machine jumping about!

I am interested in your speed estimations for the basic Darwin design. We have printed on the Version 3 machine at 64mm/sec, this high speed is very hard on the machine, moves are very aggressive. A one off move is not an issue, rapid reversals on infill are.

I have slugged my machine back to 16mm/sec for normal duty, it keeps it quiet and it does not shake itself off the bench! Ian tends to run his machine a bit faster.

The version 3 design has been strengthened, particularly around the x-axis where the original Darwin is very weak in torsion, also the new frame is fully cross braced and we run linear bearings to get friction losses down.

What modifications are you doing to the standard Darwin design in order to run continuously at high speed?
I haven't made any modifications or run it in ernest yet.

I simply observed that with a signal generator the pull in rates were about 120mm/s for Y and 150mm/s for X. That was with no accelaration, but also with no extruder fitted.

For extruding I will add minimal acceleration to keep the stress down. For moves I will acclerate all the way to the mid point, or till I hit the pull out rate. That accelaration can be relatively gentle, so less stress than the extruder on moves.

The 10:1 was based on 32mm/s 320 mm/s, just a wild guess till I cook up some electronics.

Could you post your skeinforge preferences somewhere so I can try your settings on my machine

Is there an easy way to do this or do you have to go through all the plugins and list settings?
Enrique - if your looking in - do you have or can you add an export settings option to make sharing settings easy.
Ie; copy all settings into a text file I could e-mail with an stl file

The set up is basically standard
Using 0.5 nozzle
235 DegC print temp

Export - delete comments
Fill - Hexagonal
Inset - as default
Material ABS
Polyfile - Execute file
Speed - 16mm/sec

Unchecked and not used

Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to
Posts [Atom]