Skip to content

Archive for May, 2016


Just spoolin’ with ya

After having constructed more Kossels than the average person, one thing that annoys me about them is the spool handling setup, or lack thereof.

Normally what you get is something like this:

Kossel Peg

Which in all fairness does work, but it sticks out some place, usually in the wrong place. Filament spools come in every size width and you need a long peg to handle those Kg spools.

Thus, I’ve never been a fan of them.

Next up we have the spool rollers like this one:

PLA spool holder

Now one would think these would be the perfect holder. And they are. Right up to the point that your Kossel does a massive retract and extract. And then this spool has a nasty tendency to unspool itself from the momentum. I built one of these but used skate bearings on the ends. I was checking the printer and found it had unwound several feet all on its own. With bearings it was frictionless. Not good.

The variation of this is the horizontal spool roller that sits at the top of the Kossel.

Kossel Horizontal

Again it works but I have the power supply mounted on the top of my delta so I don’t want the power supply fan restricted by a big reel of filament.

So this brings me what I worked with, when I actually worked for a living, a straddle winder. Course when I worked with it, the rolls were a lot larger than a 1Kg roll of filament. But the principal remains the same.

Stradle winder

Scanning through Thingiverse will give you any number of these types and one by AirTripper has been one of the best.

Airtripper s Pocket Filament Reel Rollers by Airtripper  Thingiverse 2016 05 19 15 59 11

I built one of these early on and have been using daily ever since. When I built my second delta printer I decided to see what else I could come up with. Something that used less bearings for example.

This lead me to a design called “Spool Foot” on Thingiverse.

Spool foot

It uses four 608ZZ bearings, some 8mm nuts and bolts. So I printed one. And immediately found that with the foot print of the cradles it would not go close enough together to fit my narrow spools from eSun. Opps.

Over to Tinkercad and made some modifications.

IMG 1454

With the offset foot design, I can move them very close together. I also decided to use 5/16″ x 1 1/4″ bolts. The final design looks like this:

IMG 1455

The only thing that moves is the bearing, the filament spacers either side of that clamp it in and keep the filament spool in alignment. It works super.

If you want to build your own, here’s a link to my design


Kossel Fan Update

The dual hot end cooling fans have been installed on Rocky & Bullwinkle for quite a while now and they are working to cool the prints better than without fans.

Having said that, remember the old adage that you don’t get something for nothing? Well, turns out there is a side effect to using fans, or rather adding anything to increase the mass of the hot end. A perfect example is in the photo below. Specifically the vertical lines you see on the right side of the print.

IMG 1386

I normally have the perimeters set to 3 and what you see is a result of the rapid movement of the head for the perimeter and then the acceleration as it moves away from the “jerk” point. Call it a ripple or ringing effect. What happens is in a very fast move, the whole mass of the effector jerks, and then the vibration tends to die down and you end up with that little wavy effect in the photo.

Even without the fans, when I went back through some of the prints I’d made with a bare minimum hot end (just a fan shroud and fan), there were still some vibration artifacts.

Truthfully, I hadn’t really even noticed this until I was printing out a CODEX that consisted of a numbered wheel. I have rotated the picture 90 degrees clockwise so keep that in mind when you’re looking at it. As it comes off the printer it looks like this:

IMG 1426

The “4” on the left is with some Marlin firmware adjustments, the one on the right with the ripples is where I was normally printing.

IMG 1427

Another example is the “7” on the same wheel.

IMG 1424

The only way to fix the ripples is not really to slow down the printer per se, but to tinker with the “jerk” and “acceleration”. In the Marlin firmware the section you want to tinker with is (configuration.h):

// default settings
// delta speeds must be the same on xyz
#define DEFAULT_MAX_ACCELERATION      {3000,3000,3000,3000}    // X, Y, Z, E maximum start speed for accelerated moves. E default values are good for Skeinforge 40+, for older versions raise them a lot.

#define DEFAULT_ACCELERATION          3000    // X, Y, Z and E acceleration in mm/s^2 for printing moves
#define DEFAULT_RETRACT_ACCELERATION  3000    // E acceleration in mm/s^2 for retracts
#define DEFAULT_TRAVEL_ACCELERATION   3000    // X, Y, Z acceleration in mm/s^2 for travel (non printing) moves

// The speed change that does not require acceleration (i.e. the software might assume it can be done instantaneously)
#define DEFAULT_XYJERK                20.0    // (mm/sec)
#define DEFAULT_ZJERK                 20.0    // (mm/sec) Must be same as XY for delta
#define DEFAULT_EJERK                 5.0     // (mm/sec)

The values for default acceleration and XY and Z jerk on the “wavy” examples are shown above.

When I started testing, I changed first the Jerk, and it helped a little, then changed it back and change the acceleration default and it helped a little. What I found at the end is that it’s a combination of both that worked best. And of course it depends on how badly you have those vibration lines to start with…

For the numbered rings with the least amount of ringing, I used a default XY and Z jerk of 10, and a default acceleration of 1000.

#define DEFAULT_ACCELERATION          1000    // X, Y, Z and E acceleration in mm/s^2 for printing moves
#define DEFAULT_XYJERK                10.0    // (mm/sec)
#define DEFAULT_ZJERK                 10.0    // (mm/sec) Must be same as XY for delta

The result:

IMG 1423

Yes, there is still a seam, but the ringing is a LOT less than the original.

The downside to having cooling fans on the hot end is that when you change the jerk and acceleration, you slow down the print job. The first print was 1:31, the second was 1:42. But if it’s quality that you’re after, 10 minutes seems very little to give up.

While I was testing, I also decided to change the perimeter settings in Slicer from my usual 3, down to 2. The idea being that it might help the hot end jumps. Safe to say after testing a number of prints, the number of perimeters doesn’t matter squat as far as I can see. It does help slightly with heat on the overhangs but that’s about it.

Regardless of the printer type, the head is moving in one direction and changes to another, there’s always some backlash. Vibration and ringing on the print are the result. If you want quality, especially with an object with writing on it, speed is the enemy of quality.

Maybe someone will come up with a vibration attachment so one can really fine tune for maximum speed while maintaining minimum ringing. Until then, like most others, I’ll just bumble along the learning road.


3D Printer Nozzles Part 2

A few days ago I was testing out some printer nozzles and at the time I had some perfect results with the largest clunkiest one in the lot. I.e. the left most one.

IMG 1347

All of the nozzles are .4, 6mm thread, typical MK8 style. The difference in all is the inner chamber and the outside nut size.

  • Smallest ones, the two on the right, inner diameter of 1.85mm, nut size of 6mm.
  • Middle two, inner diameter 1.91mm, nut size of 7mm.
  • Left “monster”, inner diameter 2.25mm, nut size 7mm.

I did a lot of test prints and some various “worst case” prints that I use for testing everything from coverage to bridging on the printers.

And the term, “one size fits all” started to quiver on shaky ground. While the monster nozzle did some designs well, as in large designs, with smaller designs that had small circles and such (think standoffs) that blunt nozzle tended to heat the surrounding area too much and even with fans the result was poor. At best.

I swapped out the monster nozzle with one of the middle ones, 1.91mm inner diameter, 7mm nut size, 12mm length. And run many of the same prints the first nozzle had problems with. While the same issues did appear, they were far superior. So much so that I left the nozzle in and am still using it.

So there ya go. Spend about 16 hours testing nozzles and what I can say is that the best all round nozzle I have found is the middle ones in this photo:

IMG 1349

To recap the specs on it, inlet diameter 1.91mm:

MK8 Nozzle

Length is 12mm. Seems that on eBay now most of the ones I found are 13mm in length.