Monday, November 11, 2013

Having a Vase Phase! (with video)

Earlier this year Slic3r introduced an experimental feature called "Spiral vase". I took some time recently to explore its capability, and was very satisfied with the results.
Above... Dandelion Vase from Thingiverse.

I've often noticed people printing vases and other similar shaped objects and asking questions about how to make them hollow, what wall thickness to choose or how best to avoid seams and and other such issues. This "Spiral vase" check-box in Slic3r simplifies the whole process and produces great results.
Above...Slic3r, Print Settings Tab, Layers and perimeters... Spiral vase check-box.

Once the "Spiral vase" box is checked, Slic3r will set a single perimeter, ignore any Infill settings, automatically making the object hollow, ignore support settings if enabled, and will not print a top solid layer. 
Above, my first vase, standing 200mm tall which is the full printing height of my Mendel90. I set a layer height of .2mm, width of .5mm matching the nozzle width of .5mm. I set 5 bottom layers to give a good base. It printed the base layers quickly (60mm/sec) but once it started into the wall it slowed right down. I'm unsure why it seems to ignore the speed settings once printing the wall commences.
(Above) The first one printed so well, I printed some more, all in PLA, with default fan cooling enabled within Slic3r.
My next object was a Gear Vase by Halalan. It's part of a set of vases which have been cleverly generated by Python script. I simply downloaded the Gear.stl file but changed it's Z scale reducing it to .75, making it shallower to become an LED tealight holder. The single wall thickness allows the light to shine through, with the shape giving it unexpected strength.

What the photos don't show is how the "Spiral vase" feature in Slic3r handled the Z movement of the printer. In typical 3d printing, when a single layer has been completed, the Z motor(s) move by one layer height and printing then resumes. With the Spiral vase box checked in Slic3r the Z movement is actually continuous, so that in the time the printer takes to print a full layer (or vase circumference) the Z height has risen by one layer. The best way to illustrate is with a short video.

This short video (below) illustrates the continuous upward Z movement. For each 360 deg. travel around the object's single wall perimeter, the Z axis rises one layer height. In the case of all the objects in this post it was a .2mm layer height. Also of note at the end of the short clip is the retreat of the print bed so the large cooling fan can do its job. 
With one Gear Vase printed I thought it would be a nice idea to have a second one with the "swirls" going in the opposite direction. The handiest way I could find to do this was to set the X Scale to "-1" in Repetier Host, then slice the .stl file. (Illustrated below.)

The completed pair of tealight holders is shown below... The flickering tealights give a lovely effect through the thin single wall.

Finally, this is the Double Twisted Vase (again from Thingiverse), the blue PLA contrasting nicely with the yellow flowers (the flowers are real!). There's a little glitch in the upper section of the vase. Slic3r did give warnings, but I printed anyway. If you are planning to print this object, it might be best to slice in something else or run the STL through Netfabb Cloud Service to see if that fixes it.

Technical Notes: .2mm layer height, .5mm width, .5mm nozzle, PLA at 185Deg C, heated bed at 60 Deg, with glass surface. I had just fitted new glass so I simply cleaned it with window cleaner and the PLA took to it nicely and clicked right off when the glass cooled. I'm not sure how long the "grippiness" will last.

Conclusions: The "Spiral vase" feature in Slic3r is worth checking out. It simplifies things and produces excellent results, with it's continuous Z movement approach to printing. There are no seams and no stop/start ooze print quality issues as a result. It does print slowly though, and seems to ignore the speed settings chosen in the set-up. This warrants more examination to see where the speed constraint lies.

This Slic3r feature is not restricted to "open top" vase style objects, and so if carefully selected, closed hollow, single wall objects can also be printed. An example I printed (below) is this ornament by Ben Malouf.
The continuous movement and flow of plastic, due to the non-stop Z movement employed, leads to an amazing blip and blemish free, smooth surface finish. This was evident in, and common to, all the objects I printed using this feature.

Hope you enjoyed the post. Thanks for viewing!

Sunday, November 3, 2013

Print quality issue [Resolved]

With a good number of hours under my belt, a print quality issue began to show up on the new printer. You can see it evident in the side walls of the battery holder pictured below.
A closer view (photo below) shows lots of gaps and a general poor finish.

I entertained a number of possible causes: Too much "retract" when executed, resulting in extrusion not resuming flow in time for printing. I even entertained ambient temperature changes being an unlikely but possible cause of the problem, as Winter was setting in, and considered increasing the hot-end temperature as one of the things to try. But before changing anything, I set another print under way, with a slow print speed (20mm/sec), and had a close look at the printing process in action. (I've made a cut-away fan-duct just for easy inspection, which I must post about.)

On observing the "gaping" happening a number of times during the test print, it was obvious that the large gear that drives the hobbed bolt was slowing down and sometimes stopping, when it should have been turning at a steady rate. Closer observation of the smaller gear revealed the issue. The extruder motor shaft was slipping within the small gear.

I knew from past experience that simply tightening the grub screw was most likely not going to be a long term solution, as had the grub screw been resting against the "flat" on the motor shaft then it wouldn't have slipped even when it had worked a little bit loose. The reversing action of the motor would have resulted in more of a clicking noise than full slippage.

On removing the small gear, it was now clear that the "flat" on the motor shaft was not long enough for the position of the grub-screw on the small gear. I should really have spotted that on first assembly. The grub screw needs to tighten down to a flat surface on a shaft to best secure it in place.
The above photo shows how the "flat" on the stepper motor shaft doesn't extend enough to align with the grub screw in the small gear.

The "flat" was easily extended using a flat needle file. The bearing of the stepper was protected from filings with some 'blue tack'. The shaft was gripped in a small vice. (See photo below.)
The new "flat" doesn't have to be perfect, just enough for the grub screw to seat against. (photo above)

The small gear was refitted, and the grub screw tightened down, ensuring it aligned with the newly filed "flat".

I then reprinted the battery holder and the difference in print quality was immediate (photo below). The plastic flow was consistent and the print finish excellent.

What is also evident, with consistent print quality is that the z-movement on the Mendel 90 is so smooth. Each layer is laid down perfectly above the other on a vertical wall, such as that seen in photo below.
Technical Note: .2mm layer height, with a Width over Height Ratio of 1.8, sliced in Skeinforge. Print speed 50mm/Sec. Now to tidy up my rechargeable batteries!
Thanks for viewing.