Saturday, 11 August 2012

Printing PLA on Paper

A few days ago I started doing tests of 3D-printing PLA using paper as a build surface, since I previously saw it briefly trialled with some success.

My first test mimiced Doxin's test by simply using cheap white A4 printer paper. The first things that I tested printing on the plate were: a new and more robust axle support for my spool stand, and a couple of the ball-joint pole-end-caps for EFFALO's 2V geodesic dome connector set. The paper wasn't clamped down very well since I could only find 4 bulldog clips at the time, and had wrinkled up either end of the paper slightly when attempting to clip it with its longest side in the X-axis before settling on having it along the Y-axis.
Paper set up for first test.
I pre-heated the printbed to something around 90C at full power, and then set it to half-power before starting the print. (I should probably note here my peculiar heated bed setup for anyone who hasn't already seen it; after a kit from mendel-parts that came with two 70W variable-voltage laptop power supplies didn't prove sufficient to print with ABS on its own, I got a simple standalone 12V-fixed supply to dedicate to the gen6 control electronics, while connecting the laptop supplies to the heated bed in series. With them set to put out a total of about 32-26V, and a total of roughly 8Ohms on the printbed heating resistors, I am usually just over their maximum working amperage for 12V, which gives me a temp around 100-110C needed to keep ABS stuck down to the printbed. However, I have no automated feedback control over the temperature during printing since that gen6 board doesn't support it.)

Initially, the adhesion between PLA and the paper surface seemed to be just great, much better than to kapton (polyimide) tape anyway, which has been a good surface to use for ABS.
Laying the first layer down.
However, several layers in, one of those small ball-joint parts started to give way from the paper surface.
Unlucky that adhesion was lost, but lucky to catch this moment on camera as the nozzle kicked one part up on its way past.
I left it alone to see how it turned out, partly to want to do a full test and partly because the part that had broken loose was tiny while my frame part that was most of the print had already committed a fair amount of PLA. You might also see in the picture above that the paper had started to curl up at the edges due to some shrinkage of the PLA as it cooled.
Finished test-print.
I was somewhat lucky that by the end of the print, the other two parts hadn't been destroyed, as strings of PLA being extruded into air next to the displaced part ended up anchoring it in place, hence the neat dome on top of a mess above.
When I allowed the parts to cool, I found them far easier to remove than was suggested by Doxin
First paper test part removed, without any cleaning.
The paper barely seemed to stick to the PLA part at all, and although it had a faint tinge of white, I didn't have to wipe any paper off of this part or the ball-cap that survived. The part was slightly warped once it cooled due to the amount of slack given by the paper, but it is still completely usable for its intended purpose.

For my second test I clipped down a sheet of printing paper again, making sure to keep it as flat and taut as possible with the bulldog clips. Once again I pre-heated the printbed at full power, this time making a note of its temperature as 91C across the paper, while 95C on the kapton surface at the edge, then switched it to half-power when starting a print of two top vertices and two footed vertices from Prusa's 2nd iteration reprap parts.
2nd test 1st layer, notice that for some reason these newer parts have arrows that should point away from gravity when assembled, or in a bunch of directions people have called 'up' ever since uninformed goat-herders declared the earth to be flat.
I took a bit of footage of this test, showing in rubbish phone-quality video how well it got on at the second layer:
By about the 5th/6th layer I started to notice that the paper was increasingly wrinkling up and that at least one of the vertices was starting to peel away from the paper at one end. I quickly measured the bed temperature at about 55C through kapton tape at the edge. My best guess at the time was that having the printbed on at 18V was not keeping it hot enough to keep the PLA sufficiently adhered on the paper surface, due to differential expansion and contraction between the two materials as they cooled, since I had seen the paper wrinkle up a lot when I switched the heated bed off at the end of the last print, before the previous parts came off easily. So, I set the heating back onto full power til the end of the print, went out to do some gardening and hoped that the parts would remain stuck down.
They did.
The part quality was pretty good, but after I switched the bed-heating off and allowed them to cool, I could see the paper wrinkling up a lot as they warped from cooling contraction.
Reverse side of the paper, with parts still attached, gives a nice visualisation of the stress/strain being put on the paper by contraction of the cooling PLA.

The amount of warp wasn't huge, but it was enough that I probably won't sell these parts to anyone, and just use them if I'm building some frame of my own, such as for a base for my rotary-hydroponics unit, since it could be annoying when trying to get a printer frame nice and square. The amount of paper that remained stuck to the parts instead of peeling off cleanly probably covered about 10% of their area, but it was pretty easy to scrub/scrape that off afterwards, although they would be completely usable with a smudge of paper stuck to one side.

Next I wanted to see whether newspaper could be used as a printing surface, since it can be easily procured for free and would be better than wasting higher-quality printing paper. I was wondering though whether the dye on the newspaper might stain the parts, and whether that would make them look ugly.
First newspaper test sheet clipped down. +10 points for recycling and +9000 points for something other than the 2012 olympics on the front page.
Not wanting to risk having trouble with adhesion due to heat changes during printing again, I swapped out the voltage-setting block on one of the power supplies so that I now had 18.5V+12V. After letting this warm up for a while I took quick temp readings of 80C through tape at the plate edge, 75C at the paper edge, and 81C in the middle of the paper, then set a print running to built two Prusa-i2 footed vertices.
The first layer just finished, adhesion looks fine.
By the end of this print those temperatures had just crept up to 93, 90 and 97C respectively, and the newspaper barely curled up at all during the print.

First test parts removed from newspaper print surface.

This print went fine, albeit with as much warp as the previous test, and my suspicions about staining the PLA were confirmed. Scrubbing the excess paper off didn't noticably take any dye with it, so I think it has soaked into the first layer of plastic. I don't intend to sell these either, unless someone out there sees some artistic merit in a pair of footed mendel vertices with part of "inverness highland games" across one side of them.
This gives me ideas about intentional applications of printing on dyed paper though; perhaps some aesthetic or functional designs could be displayed on the outside of parts by printing them to scale with your intended parts-plate and then zeroing the extruder on a set point on the paper. More appealing to any parents out there wanting to print something fun for their kids, it could even be possible to print a PLA jigsaw puzzle this way, with the PLA soaking up dye from a digital image mirrored before printing on a regular inkjet printer.

On to the second test, I set out a plate of 8 bar clamps from the Prusa i2 set, and 4 of the ball-joint caps that I had a problem with before. If anything was going to let go from the paper here, I expected it to be them with their tiny footprint in the first layer relative to their height. Temp readings here went from around 80C at the start up to 85C when I stopped it, due to this:
Several layers in, things got a bit messy.
I was very surprised when most of the bar clamps started coming loose from the paper, and there didn't look to be much hope of this print producing anything other than a mess, so I aborted it to investigate.
Marks left on newspaper by small parts after removal.
All the parts were easy to remove from the paper, taking no paper with them at all, and it seems to me that what happened here was too much ink in the heavily black-dyed sections interfered with the parts adhering to the paper, acting like a thin layer of dust that quickly gave way, which makes sense if the newspaper was produced by laser printing, in which particles of dye are attracted to the paper by a charged roller and sealed on with heat.
I guess this means I won't be getting any bar clamps with a snapshot of someone's kilt on one side, unless I can find someone to supply me with tartan filament (it'll probably be widely available in web-stores by April now that I mention it).
Speaking of which, I recently received a delivery of more PLA filament from a new-ish UK supplier called FilamentPrint, who provide PLA not only at a competitive price in general, but importantly to me with a cheap delivery service who are actually cheaper than Royal Mail when delivering to the Scottish Highlands, which I haven't seen any other couriers do (most of them seem to loathe all the driving up single-track roads to get here and demand a hefty surcharge). One of their new staff apparently managed to misplace one of a few 100m PLA reels that I ordered when packaging it, but when I asked about this by email I got a very apologetic reply 1 hour later from someone else at the firm saying that they would send out the missing spool of filament the same day, when it was already late in the afternoon, and it turned up yesterday. I doubt they'll be making another mistake like that quickly, so I would recommend them to anyone on this island starting out with 3D printing, on the basis of both price and customer service.

Back on topic, I was almost ready to give up on newspaper due to that mess the other day, but I did one last test shown here to see whether distance from the centre of the paper was affecting adhesion strength at all:
A hub part to make one of my modular filament spools.

While surface adhesion for this part was quite solid, it warped quite badly in spite of my finding a few more bulldog clips to place around edges of the paper, and I ended up trying to flatten it a bit by leaving a book on top of it with a couple of full 822g jars on top while the plate was hot. It seems to be a usable part now, but this warp problem is putting me off using paper as a print surface.

For my initial conclusions, so long as there isn't a huge amount of ink present, it seems that adhesion is more reliable than on kapton tape alone, which I can only get PLA to print on after wiping it clean with a dash of dark rum (yes, no joke, that actually works; my theory is that after the ethanol cleans fingerprints away, a thin brown-sugar-derived residue is probably left behind that aids the sugary PLA in remaining stuck down), while using acetone to clean the printbed has only worked before using ABS for me.
There is a bit of a paradox in that larger parts appear to remain stuck down to the paper better, while those same parts are more subject to warp. Paper might prove effective for small parts that are unlikely to warp, wherever the first layer has enough surface area to adhere well. I've also just been thinking that perhaps warping problems might be reduced to some extent by using stiff card instead of paper, so that it can't wrinkle up - using the inside of dry-food packaging boxes as a print surface, before recycling the cardboard, might be a worthwhile experiment.
Of course, a thin plate of glass cut to the size of a reprap's print-bed is still the best surface that I've heard of for PLA, but when the nearest glazier's shop to me is 100 miles away, I can only make do with what I've got for now, and I think that will be trying to keep a kapton-tape surface scrupulously clean from finger-print-grease.


  1. did my thought of using a fishing reel mechanism to maintain tension on the spool have no merit Andy?

    1. I'm pretty sure it does have merit, I just haven't got round to trying it yet, on account of not finding a spare fishing spool to try it with.

    2. I should probably mention I'm not using my half-arsed bungee cord idea anymore, since that itself got tangled up a couple of times, but now I have fitted nuts and washers either side of the spool on that piece of scrap studding that I've been using as an axle (having now cut a paint-clogged section off one end), with a disc of cardboard between the spool and nearest washer to provide a bit of extra friction, but it's still not as good as a ratchet.