Very Poor Layers

[inventabuild]

Well, I've been running a lot of tests and made significant improvements to the print quality of the vase (http://www.thingiverse.com/thing:25278) I've been trying to print with KISSlicer by doing the following:

1. Printing from the SD card.
2. Setting Prime = 10.6, Suck = 10, Wipe = 5, Zlift = 0.5
3. Configuration.h Motor Current Settings (195, 195, 195, 195, 0) which are the stock settings in github.

Throughout all my testing I already had a fan mounted to the effector to blow on the Peek and I tried blowing a fan over the build plate, but this did not make any noticeable difference. By the way I'm only printing at 30 mm/sec.

Here is a pic of the vase using the above mentioned settings. Although there are still some issues at the seams near the edges of the vase and a couple of small blobs near the top, the rest of the vase looks beautiful:

Here is a side view of the vase w/ no seams...looks beautiful:

When I lower the prime so it's closer to the suck, for example if I try anywhere from 10.0 to 10.5 I get worse results. Does this mean I'm getting too much bowden backlash or I have too weak a motor for the task at hand? What other extruder motor settings can I try to adjust besides the current? I'm using the Nema 1:1 and the EZStruder both of which come stock with the current Rostock Max kit.

[lordbinky]

The motor isn't a problem, and that difference can easily be the flex in the bowden tube. There isn't much to do about that without shortening the bowden length. Check out the flying extruder mod if you want to go that route though

It's possible to reduce that flex by keeping a constant force on the tube (effectively making it more rigid), or by adding a more rigid support to the tube such that retraction does not relieve any stresses in the bowden tube that has to be restored before the filament moves the intended amount. That carries the possibility of affecting your hotend's positioning if it is too rigid, so you'd be hunting down the sweet spot for optimal bowden tube rigidity next.

The entrance to the rabbit hole was WAY back there. So don't worry there's no point in stopping now.

[626Pilot]

Are you using the old side-mounted extruder or the new top-mounted EZStruder? The new one is worth getting if you don't have it. The length of the Bowden tube is shorter, and you can put more stuff next to your printer!

Take a look at the PTC fitting coming out of the extruder drive while it's retracting. Does the little plastic piece that keeps the tube in place seem to move up and down when it retracts and extends? If so, you can try supergluing the tube to the connector. I did that when a fitting died and I didn't have any spares. It doesn't bob up and down anymore. It took three coatings to get a sort of gumdrop shape. If it doesn't bob up and down, don't bother.

[Flateric]

You guys do know you can minimize tube flex by putting the 1.75mm tube inside the tube made for 3mm.

It's a small tweak, but they all add up right?

[626Pilot]

It's possible to reduce that flex by keeping a constant force on the tube (effectively making it more rigid), or by adding a more rigid support to the tube such that retraction does not relieve any stresses in the bowden tube that has to be restored before the filament moves the intended amount. That carries the possibility of affecting your hotend's positioning if it is too rigid, so you'd be hunting down the sweet spot for optimal bowden tube rigidity next.

Maybe if you lash speedometer wire to it? It's flexible, but only sort of. It resists torsion.

[kbob]

You guys do know you can minimize tube flex by putting the 1.75mm tube inside the tube made for 3mm.

It's a small tweak, but they all add up right?

What? Are you putting PTFE sleeve inside PTFE sleeve? The sleeve I got from SeeMeCNC is about 4.2 mm OD. I can't see that fitting into a 3 mm ID sleeve...

My sleeve's ID measures about 2.1 mm, and the filament is quite loose in there. I can pull or push it 3-4 mm with light finger pressure. I should try to get a tighter sleeve, I think.

[626Pilot]

Maybe the next thing to try is mounting the extruder directly to the hot end. No Bowden tube then. Someone mounted a QUBD extruder under the platform of their Rostock and had very good luck with it. Maybe the next thing to try is sticking the EZStruder on top of a groovemount print head, since it has that capability.

[Flateric]

This is a bad idea in so very many ways. You want to keep that effector as light as possible. Any added weight forces slower less accurate movements.

[lordbinky]

If you had a tiny lightweight extruder motor with tiny lightweight gearing it might work. Although you'd be stuck with slow speeds because your tiny extruder would be limiting how fast you can extrude filament...

[626Pilot]

This is a bad idea in so very many ways. You want to keep that effector as light as possible. Any added weight forces slower less accurate movements.

I think it's worth investigating. It's been done and the print quality is great. The GT2 belts add a lot of parasitic inertia to the system so I'd do a fishing line conversion at the same time. With carbon fiber arms I don't anticipate any issues with flexion, so at that point the limiting factor is probably stepper performance and how much inertia the frame can absorb before it wants to start tipping.

In sports cars it's common to install tower braces that connect the shock absorber towers on the right and left sides so that roll force is more evenly distributed while going around corners. The chassis stays relatively flat. Tower braces also translate the upwards force experienced by the wheel on the inside of the turn into DOWNWARDS force on the outside wheel, which is really good because it wants to lift up as the chassis experiences roll. I designed some tower braces that go with threaded rod and wingnuts to pull all the towers as close to one another as possible last month and that made it SUPER easy to align the printer. Something like that, but designed to keep the threaded rod out of the way (it's normally just for alignment and requires dismounting things) would probably work. One set of braces up top, one on the bottom, and some vertical rods connecting the top and bottom, would make the Rostock significantly stiffer than it is now. To make it complete there would also need to be some "flying buttresses" extending from the tower bracing down to table level and outward, to distribute side loading to the surface the printer sits on. If you want to get all serious business, it could even be bolted down.

Coupled with the EZStruder mounted directly to an E3D hot end, it should be able to run good and fast and at really high quality. The weak spots in the equation are how much torque we can get out of these steppers, whether beefier ones will fit if needed, and how well the extruded plastic can cope with high speed deposition.

[lordbinky]

You can see the results of direct extrusion on a rostock max at http://www.kronosrobotics.com/3d/rostock/. The slapped a QUBD extruder on it and they did have to slow their print speed for it. I suppose if you wanted to find the sweet spot between gearing and stepper motor torque/weight you might have something to aim for.