ABS “Die Swell” from hell!

(Update : See note below ** for update on the confusion over what materials I was trying to print with)

 

The learning curve on this project just keeps on rising!

 

I have noticed during calibration that my large roll of natural ABS** seems to suffer from extreme Die Swell feeding a 3mm filament through my .5 mm parcan hotend results in an almost exactly 1mm extrusion filament!

 

So in essence I am getting extreme die swell, If I setup Pronterface & SFACT as I would have done for PLA I am printing with a 1mm filament which leaves a lot to be desired in terms of print quality (I was hoping for approximately 0.5mm) after investigating (on the web) how I might deal with this annoying trait of what appears to be just this particular batch of ABS I found that  “Die Swell” is much more common on ABS & HDPE that PLA (nonexistent on PLA) it was suggested that you may be able to stretch the filament whilst printing by increasing the speed (although after further reading  this only seemed to work on non 5D based firmware’s as if you increase the speed on a 5D based printer the software/firmware adjusts the flow rate accordingly) so after running several print tests I have come to the conclusion that the only way I can deal with such extreme Die Swell is to setup the Software as I would for PLA which results in a 1mm print extrusion (too large for any sort of decent prints) and then half the flow rate which gives me a nice .5mm print side filament, I will now need to adjust the “Carve” settings to bring the extrusion diameter down until the extruded filaments begin to fuse.

 

As I only managed to get my head around this last night I have yet to test this in anger but it is looking like a promising potential solution to this extreme Die Swell issue.

 

Other people have just dumped the offending ABS and bought some that does not suffer from this, I also read that neutral or white ABS seems to suffer worse than Black ABS so my next reel of ABS might end up being a dark colour to hopefully avoid this issue.

 

As I am not made of money I can’t really afford to dump 200 meters of filament so I am hoping my software “Fix” might allow me to at least use up this old ABS reel before investing in more (also it has been a very good lesson on the trials and tribulations of trying to use ABS instead of PLA).

 

** I have since discovered that this was HDPE not  ABS after all, hence all the problems I was having trying to print with it, I have also discovered that my entire roll of what I thought was PLA has now been chemically proven to be Polycaprolactone  AKA PCL or CAPA (trade name) http://www.biodeg.net/bioplastic.html.

 

I proved the CAPA by chucking an early failed print into a saucepan of water on the kitchen hob and heating as soon as the water started to boil the whole lump just turns to goo which you can reform in your hands and once cooled sets hard, this material is also known as friendly plastic or polymorph (tradenames) and added bonus as I should be able to heat and extrude back into 3mm filament for reprinting, I must get my plastic recycling extruder built.

 

So in essence every PLA/ABS reference previous to this post has incorrectly assumed that I was printing with PLA/ABS when in fact I was printing with PCL,CAPA and HDPE respectively, what a dummy! Oh well you live and learn.   

1st Child Printer comes to life!

(Update : See note in ABS “Die Swell” from hell! for update on the confusion over what materials I was trying to print with)
Printed the Z motor mount bar clamps & end-stop holders and finished the mechanical assembly, wired up all the axis motors, end-stops and ran a quick functional test to see if everything was working as expected, apart from a few firmware tweak’s it all did, once again buoyed by this success I rushed headlong into stripping down the Darwin and fitting the ABS extruder mechanism & Parcan hotend onto the child printer.

In less than a few hours I had hot plastic coming out the hotend and was ready for a first print, went for the widebar just to see if I could get anything resembling a decent print and to my surprise the first layer came out almost perfect, I was still printing with PLA at this point and decided that before I calibrate (yes I rushed into printing without calibrating first, not recommended!) that I would switch to ABS.

ABS requires the hotbed that was still currently fitted to the Darwin, stripped this of the Darwin and removed the Prusa’s build plate (I used a 6mm piece of MDF for this as I had some handy) drilled mounting holes and fitted the hotbed to the Prusa build plate.

Refitted the Prusa’s build plate and ran my first attempt at printing ABS on the newborn child printer, after a few unsuccessful prints of the widebar (unsuccessful due to the ABS not sticking to the hotbed) I had my first failure.

I had melted my beltclamps! Due I guessed to the fact that PLA melts at a very low temperature compared to ABS and the MDF Prusa build plate was getting  too hot.

Stripped off the build plate and stuck a sheet of aluminum foil to the MDF underneath the hotbed, I also doubled up on the PLA belt clamp’s to effect a bodge until I can print some ABS belt clamps, remounted the hotbed (a little higher this time to give more space between the build plate and the hot bed) and tried again.

Several failed attempts later I decided it was time to investigate “ABS not sticking to hot bed” on Google, after reading lots of potential solutions I found that lots of people were recommending that you dissolve some abs in acetone and prime the hotbed with this before printing.

Bought some acetone (nail varnish remover) tried to dissolve some of my ABS in it to no avail the acetone did not seem to touch the abs after several days I was starting to think that I was not using abs at all and that maybe this was some other sort of plastic filament (don't forget I had been gifted the 2 rolls of feed plastic with the Darwin and so I was basically guessing at the actual materials composition)

After several more failed prints and me getting more frustrated at the refusal of my ABS (if it is ABS?) to stick to the hot bed, I was rapidly trying print after print tweaking the 1^st layer height mechanically by adjusting the Z end stops position and noticed that I could peel the failed partially stuck print off the hot bed whilst it was still hot and during one failed attempt I also noticed that if I rubbed the hot failed rolled up ball of ABS across the hotbed it seemed to leave a trace of ABS behind.

In a flash of inspiration I decided to cover the whole hot bed with a second layer of Kapton tape and use a piece of failed ABS print like an eraser across the whole build surface, to my surprise (and relief) this seemed to do the trick.

I discovered that if you use a blob of ABS like an eraser, rubbing it rapidly and fairly forcefully (not to forcefully as you don't want to break the hot bed) it leaves a trace of ABS behind essentially misting the whole surface with a fine layer of ABS.

This seems to aid the first layer in sticking to the hot bed greatly and allowed me to move onto calibrating.

During this second round of ABS tests I also managed to partially melt my X carriage, powered off everything and rapped the entire x carriage in aluminum foil and kapton tape in an effort to avoid the need to print another x carriage.

This appears to have worked and so far I have had no more PLA dripping onto the hotbed!

20/07/2012

Calibrated the E-Steps per mm today (in my lunch break) and discovered that it was pushing out 12mm more feedstock than it should have been, seeing as I am using 3mm feedstock and have a 0.5 mm hot end this means that over 100mm of input feedstock I was extruding approximately 68mm worth more output plastic out of the hot end than I should have been, demonstrating why calibration  of the e-steps per mm is so important.

Can’t wait to test this tonight , hopefully I will be able to achieve my first successful ABS prints, and just in the nick of time before I melt my PLA X carriage.

First prints will definitely be an ABS x carriage and some ABS belt clamps, shortly followed by ABS X idler & motor mount ends (I have noticed that my current PLA ones flex some considerable amount) in fact I may just print a whole new set of parts for the Prusa just in case of failures.

Child printer woes!

14/06/2012
After assembling  the basic frame of the prusa and realizing that all of the parts I had previously printed before upgrading Pronterface were not going to be good enough I decided to attempt reprints of the Z motor mounts, these worked remarkably well and were immediately useable with no manual remodeling required at all (Cool) once again proving that software is king when it comes to quality prints.

As these 2 Z motor mounts came out so clean and accurate compared to previous attempts I thought it best to bite the bullet and go for new X carriage end as well, rather than just reprinting the standard ones I had a good look at the alternatives on Thingiverse and found some great looking advanced versions (ref) that appeared to use less plastic, have bolt in LM8UU bearing holders for the Z axis smooth rods and a much improved way of clamping the X carriage rods in place (rather than the captive nut method which is ok but quite fiddly to assemble.

I went for the motor end first as this was the one I was having most of the problems printing with the old software version/configuration, I was stunned by the result it printed perfectly and the quality puts my other parts to shame, buoyed by this success I followed up with the idler end same design, at this point my Darwin decided to remind me that it was not a happy bunny and began to rear its ugly mechanical failure head again!

First off the X axis end stop decided to fail (whilst I was demoing to an interested friend (embarrassing) this was a simple re-soldering job.

Secondly (and more critically) over half way through the X idler end (4.5 hour print on the slow old Darwin)  the X axis began to slip mechanically, affected a repair and tried again, same problem half way through this large print and again the X axis slipped ruining the print, (at this point I am beginning to worry again about how much plastic I am wasting and wondering if I have enough left to finish the 1^st child.

Attempted another repair and broke the captive nut completely out of the resin cast X carriage drive gear (whoops) luckily there as already a second grub screw hole drilled so I used this instead and tried again ~(in fact this was about my 4^th attempt at fixing this X axis slippage issue each time wasting a lot of plastic and time)  but to my relief the 4^th attempt worked like a treat and my perfect looking X idler is now sitting proudly on the build bed waiting for me to crack it off and carry on with the prusa build.

The more I use the Darwin the more it reminds me of an old car, it requires constant maintenance and really needs replacing or upgrading  but at the same time it has the feel of a solid piece of engineering hardware and is remarkably forgiving of the number of bodge’s I have affected to make it print.

Although I can’t quite describe it as “Old Faithfull” due to the constant maintenance, when treated well it responds with great prints.

I have spotted a number of parts remaining to print for my 1^st child, namely  the Z motor mount bar clamps x 2, end-stop holders (I haven even decided what sort of end-stops to use yet!)  also I have only just remembered that in order to get the Darwin printing I used one of the steppers from the Prusa for the geared extruder on the Darwin, this leaves me short 1 motor, I think that I will just remove the entire geared extruder from the Darwin when the Prusa is ready for it and once the Prusa is printing I will probably buy 1 or more new Nema 23’s for the Darwin and other printers/projects.

Although at first this seems a fairly expensive hobby, my total build cost for both printers bearing in mind that I was gifted the old nonfunctional but essentially mechanically complete Darwin is running at about £300-£350 not including purchased tools.

I think that this is damn good value for a system that can essentially print anything in plastic in a matter of hours.

ABS tests

I have been gifted a large reel of natural ABS and recently tried an experiment with the Darwin in attempting to print using it.
Although I got close to it working it was not close enough to warrant any more time spent until I have the new child Prusa up and running.

Here is a brief description of what I managed to establish.

1)     ABS needs a hotbed (which I have already purchased) I connected it directly to an ATX PSU as suggested on the eBay page I bought it from, and after some experimentation with different ATX PSU’s with different AMP ratings I managed to achieve the recommended 115 degrees C

2)     ABS needs a hotend temp of around 230 degrees C and after much fiddling around I came to the conclusion that the Parcan hotend with single heating resistor was just not going to cut it (not sure why as I am fairly sure that I have seen others using ABS successfully with this hotend) it may be more to do with my firmware control of the hotend as however hard I try I have yet to get the hotend to calibrate properly, i.e. my IR temperature sensor always says that the hotend is some considerably lower temperature that the thermistor reading in the software says the hotend is at.

        *Update*  - I can now confirm that there is no issue with the parcan hotend for printing ABS as it works fine on my new 1st Child Prusa derivative, the issue must have been with the depreciated Gen3 electronics set and or mid 2010 5D firmware.

But considering these issues I did manage to get a single print to stick (partially) to the hotbed and was impressed to see my first layer of ABS taking up the form of a printed “widebar” unfortunately due to this being a partial success I was forced to abort the print before the second layer peeled the first and began creating ABS freeform art!

I look forward to trying this again on the new printer as I am hoping that the temperature control issues may vanish with a new electronics set tied to the Parcan hotend.

I will also be removing the Hotbed from the Darwin, which was only added to test ABS and attempt to avoid the warping issues I was having with PLA, and fitting it to the new child printer.

The PLA warping I was getting seems to have been caused by the low ambient temperature in the room that I run the printer in as I have seen a marked improvement now summer is here and the room is warmer.

I will need to look at building a heated chamber in the future as I intend to move all my printing capability into the workshop I have which is not ever likely to be heated.

1st Child printer begins to come together

I have spent the last couple of weekends worth of spare time putting my 1^st child printer together (well the mechanical part of the build at least)
So far things are looking fairly promising, I have purchased most of the vitamins required.

Electronics – I settled on the Sanguinololu set mainly because it was fairly cheap, available on eBay and supports the latest sprinter & marlin firmware derivatives.

Mechanicals – sourced from a number of suppliers, bearings from Hong Kong (eBay) motors & induction hardened smooth rods from Zapp Automation, fixings & threaded rod from RS for simplicity of ordering. In my haste to get things bought I inadvertently ordered 10mm rod rather than 8mm (whoops) and even then after I had ordered the 8mm rod as a replacement (kept the 10mm as I am sure I will use it eventually) I forgot to include the z axis smooth rods so I am still short of 2 x 350mm smooth rods to complete my mechanical build.

Hotend – I am hopefully going to be able to reinstate the original Darwin Hotend back onto the Darwin freeing up the fairly expensive (£45) Parcan Hotend for use on my first child printer. (yet to be attempted) this will involve me tapping out a small length of Peek rod (previously purchased for other tests) and coming up with a different mounting solution for the Greg’s/ Wades Geared extruder. This is not as easy as it may sound as I don't (yet) have any form of lathe (maybe I should try and print the one on Thingiverse) I will be investing in a lathe soon I hope.

AFAIK that should be all the bits I need and it is just now a case of finding the time to put it all together, I am expecting great improvements to the print quality and speed of printing from this first child, for the reasons below.

1)     Massively better overall mechanical design than the Darwin “Prototype” (thanks to the Prusa’s design)

2)     Far better electronics set (able to run latest Sprinter/Marlin Firmware derivatives)

3)     LM8UU bearings on all smooth rods

4)     Induction hardened smooth rods ( I am hoping this will mean I can push the speed boundaries to the max without the risk of mechanical failure)

Another advantage is the size of the printer is considerably smaller than the Darwin meaning that it is potentially portable.

As far as the printed parts are concerned I have had (and continue to have) mixed success with the Darwin in producing good quality functional parts for the Prusa.

I have managed to print an excellent quality Greg’s Accessible Wades Geared Extruder (ref) (4 part’s).

an impressively accurate X carriage with LM8UU snap in bearing’s by (ref)(1 part).

Reasonable Vertice ends by (ref)(1part x 6) although I neglected to notice the need for 4 x “footed” vertice ends so all 6 are the same.

A reasonable Y carriage with snap-in/bolt-in LMU88 bearing housings (this is an optional extra, to ease the need of accurate measurement/placement of the Y carriage smooth rods).

Poor quality X motor end & Idler end’s (had real problems with these due  to the mixed cooling requirements for these parts, i.e. some layers are large and therefore want no cooling to avoid warping, and the last few layers are very small and therefore seem to require some cooling to achieve the required quality, (this will hopefully be addressed with the new firmware on the Sanguinololu which will enable interactive cooling of individual layers).