Monday, 20 February 2017

Three, count them, three different extruder mounts

[Note: This post is not in chronological order]

After my previous failures, I decided it was time to get this mount right. So let's try no less than three ways to get a good mount!

Way #1: After drilling out the filament hole with a 4mm drill, use a heated metal stick to smoothen the inner surface. Would have been easier if I had had an exact match. This might make my self-printed mount smooth enough that it doesn't drag the filament, but I doubt it - it might simply be too wide by now, so there's room for the filament to bend a bit and cause extra friction. Giving this a try, I noticed that the extruder heat setting was 200C, not 220C as I usually use. Weird. So I delve in a little more, and find that it's been using some default settings rather than the ones I changed - the changed ones were still correctly changed, just ignored. Computers! Slic3r's config system is still the most horribly overcomplicated and confusing one I've seen (outside of the BIOS/Xorg/Window Manager stack which is mysterious beyond belief).

Changing the settings to think my filament is not 1.75mm but 2.85 - which it is - made a print that smacks of over-extrusion:

Cube #57: Look at that oozing!
Cutting the extrusion multiplier to .75 reduced the overextrusion a bit:

Cube #58: Somewhat oozy
And getting down to .5 brought me to a very nice piece, quite acceptable:

Cube #59: Ooze-b-gone! Nice and straight, maybe a tad extra on the corner here.
So it's entirely possible that that was the problem, together with some early clogging and dragging. That's what I get for switching extruder mount and computer and firmware version and Arduino app version and Pronterface version and Slic3r version at the same time. Don't do that.

Way #2: Order a print online. I went to and had a print made using ABS (technically PCABS, which is a Polycarbonate/ABS blend, but I just wanted the heat resistance and solidity of any kind of ABS). This got done by Kinamico's Hub on a Zortrax M200, much more expensive than mine, for €17 + €12 (!) shipping (international shipping, even from not that far away, adds up, alas). The service was really good, calling me back quickly after uploading and asking pertinent questions. Getting it took a while, though, partly because I delayed answering some of the questions, partly probably because the security conference in Munich caused chaos. I got it one working day late (unlike an Amazon package which got delayed by three working days. So much for Prime). But what a beauty! Smooth, solid, and all the holes totally on the spot.

New extruder, with various nuts and bolts fitting perfectly.
Look at that built-in support piece neatly not touching the side of the hole.
If I get any trouble with my now-working extruder mount, I will switch immediately to this one. Until then I'll just cuddle up with this little beauty when going to sleep at night:)

If you want to order from 3DHubs, you can follow this link to give both me and yourself a $10 printing credit:

Way #3: Here's the interesting variation. The black mount I got from +Thomas Riedl  worked quite well as long as the mount was secure, but the mounting mechanism with two cross screws just didn't work well (hinc illae irae). Inspired by +nop head's mount that uses three bolts from the bottom and washers to keep the E3D in place, I decided to see if I could convert this piece. Actually adding three such bolts would have been infeasible, so instead I got a 1mm thick strip of steel and cut it to shape using my trusty Dremel. Getting the outer shape and bolt holts in was easy:

Slightly trickier was figuring out where to drill the filament hole, until I remembered the trick of using a nail to mark through a hole. Bang! A precise marking that's also a good start for the drill.

Dot never, ever marks the spot

So I'll drill a slot from one side wide enough for the neck of the E3D (gotta love having a proper diagram), and the two bolts holding on the motor will also clamp the plate to hold the hotend. The only remaining part is having a 16mm stop-gap piece above the hotend because this mount design is deep:

Check out that deeposity

That I could actually cut out of wood, it shouldn't get very hot up there. Or I can use washers. But given my success with approaches #1 and #2, I'm disinclined to further pursue this venue until conditions require it.

Since I have pinged +nop head anyway, I will direct his attention to my note at the bottom of my previous post about the nut traps for the idler screws not keeping the screw from getting turned further and digging into the extruder motor.

Saturday, 18 February 2017

Printing woes with my self-printed extruder mount

Of course Slic3r had put config files elsewhere. Merging them together made everything make sense, and I was able to print again!

...with severe under-extrusion (and poor adhesion due to no hair spray). Changing the speed settings back to what they used to be didn't change anything (so once this is sorted out, I should change again to the more modern settings).

Symptoms are those of total stripping. Probably there is enough error in the width of the filament hole to push back and cause stripping. Alternatively there is more resistance in the gears., but that should cause gear skipping if it was too much for the motor.

G0 Z80 is appropriate for taking off the gears.

After opening: No sign of stripping, but the extruder idler was really hard to get off. Switched it out for the one I'd printed.

If there's no skipping and no stripping, we should be getting an even amount of filament out, but the first print was good in the beginning and then pretty crappy afterwards.

Let me calibrate extrusion - yes, way too little filament gets taken in. Something is wrong with either my E_STEPS_PER_MM calculation or the way it's used. It extruded 1.5 mm where it should have extruded 5. Wut??? I just copied over my old calculation. Just to see if it reacts properly to changes, I changed my correction factor by a factor 3. That got it pretty close. Weird.

Printing a test cube with this caused actual, confirmed stripping. Sigh.

Nice print suddenly stopped due to stripping.

I have to admit, the Pronterface UI works a lot better under Linux, the Mac version has weird UI issues.

I don't know what possessed me to think that, with all the screw holes in the extruder mount being too small, the filament hole would be the right size. I drilled it out with a 4mm bit, and now it seems to have the same amount of space as the old one. Just need to get the drilled-out bits out of there. First a couple of cold pulls, then hopefully I won't actually have to remove the whole thing again - though there could be a rim at the bottom that will need drilling from the other side.

After three cold pulls and checking that the hobbed bolt wasn't too full of red filament, I made this (double-height) test cube:

The clogger/stripping is not in evidence, but the amount extruded is not perfect - it is severely and systematically underextruded. Now that could either be because the extrusion multiplier is wrong, or because there's enough pushback to prevent full extrusion. Giving the consistency of the piece, I'm thinking the former. I thought I used to have an extrusion multiplier of 1.1, but it was showing up as 0.9. Increasing to 1.1 to see what happens. And what happens is ... stripping. Ok, time to take this bugger apart again and make it as smooth as possible on the inside.

As I feared, the nut traps holding the other end of the idler screws are not holding the screws fixed, when I mount the idler the screws turn back into the motor, which is probably not good. A nut trap with room for a nylock nut would have worked better. There is almost but not quite room for a nut on the motor side.

The mount is a lot easier to take off and on than the previous one, and the hotend mount part is pretty damn solid and easy to get aligned, if a little problematic if the nuts can't go far enough into the nut traps.

I'm concerned at how easily it strips. I was trying to calibrate the extrusion rate, and it pretty much stripped immediately, possibly due to too little idler pressure. If I can't get that to stop, I'll probably order a professionally printed mount off Thingiverse, to be sure my print quality isn't getting in the way of being able to print.

Saturday, 11 February 2017

The $200 usable ARM-based printer

We interrupt our (ir-)regular post of trials and tribulations with my Mendel90 to bring a short commercial:

There is a usable, open $200 3D printer on the market. The Monoprice MP Select Mini sounds like quite the coup if you're looking at getting started. This Hackaday review finds few issues with it, mainly poor temperature control, and also points out that it is the first cheap open 3D printer with a 32-bit processor. The time of Arduino-variations might be over. That's probably what's most exciting about if, apart from being a good cheap printer. There are new developments to be made.

The fact that unlike numerous other commercially-made printers this is compatible with all the standard 3D tools makes it extra interesting for hackers on a budget. The 120 x 120 x 120mm build volume is not huge, but frankly almost all my prints so far have fit within that.

Ok, so this is news about a year out of date, but only now saw it, and I frequently get asked about what a decent printer costs. Now I have a good answer: $199. And watch out for those ARM-based accelerations.

Note: I didn't get anything for this commercial. Monoprice doesn't even know I exist.