A comparison of filament drive bolts: The good and the better [Update]

I recently had the opportunity to compare three different filament drive bolts for Wade-style extruders with each other because one of our extruders began to slip occasionally at 60°C air-temperature in the heated chamber while another worked absolutely reliably under the same conditions.
To verify that my oldest Hobbed Bolt (bottom) made by Wolfgang (reprap-fab.org) isn’t just worn out after over 2500 hours of operation I ordered a replacement (top) from Lazlo at arcol.hu:

The Hyena v1.0 is manufactured very precise and has quite sharp edges – although no knife-like sharp teeth. The comparison in operation, however, unexpectedly showed no better grip than before.

Thereupon I took a closer look at the bolt of the functioning extruder (also from reprap-fab.org – somewhat more recent) and saw a significant difference from the others: the (tap) teeth are cut less deeply so that there is virtually no circumferential groove in which the filament gets pressed.

This marginal difference seems to be important – here are two pieces of ABS from the extruder in comparison (top: Hyena Hobbed Bar; bottom: working Hobbed Bolt):

The outer teeth circumference of the working bolt is about 17% larger than the others (25.56mm vs. 20.10 mm) and therefore bites with more distance and particularly deeper into the filament by an (almost) straight tooth flank.
In the long term we will need a bigger drive bolt radius to distribute the extruder’s force onto a longer section of the filament in order to deal with the conditions in a heated chamber (softer filament, more resistance by a longer melting zone in the Hotend). This will probably be the application of two “Mini Hyenas” that we ordered via Lazlo’s IndieGoGo campaign. Possibly along with another “big thing” on which Jonas and I are currently working…


I didn’t emphasize enough above that these were actually three different drive bolts that i compared (#1 Old hobbed bolt by Wolfgang, #2 Hyena v1, #3 more recent hobbed bolt by Wolfgang). The older #1 hobbed bolt performed equally well (or: not so well) as the #2 Hyena v1 – as epected due to their similar geometry. Both are grooved so that the inner section of each tooth flank is positioned closer to the center of rotation than the outer sections. That’s why they move at different angular velocities – which is a bad idea, it burns a lot of work just damaging the filament instead of pushing it downwards.

The second hobbed bolt (#3) has teeth flanks that are almost parallel, not grooved – all points on the tooth flank are moving at the same velocity resulting in cleanly cut filament, a LOT more torque and thus less slipping. Also the spacing of the teeth is wider and the teeth themselfes are longer (cutting deeper into the filament). As Wolfgang told me he is aware of his improvements and intentionally changed the bolt design due to the the named reasons, that he also already observed.

I had the chance to also test the pair of Lazlo’s Mini Hyenas (v1) lately that unfortunately but understandabley showed an equally bad performance. Don’t get me wrong – these are fantastically machined, beautiful pieces of hardware. I appreciate that a lot. Their geometry is just counterproductive for their intended use as a filament drive gear.

If the Easter Bunny had a RepRap…

…i’m sure he would print an eggbot (or two).

Just in time for this year’s Easter i finished building a copy of Glasswalkers Revised Fully Printable Eggbot from surplus parts i had lying around – an Arduino, a pair of discarded Steppers from my RepStrap, “EasyDriver” Stepper Drivers and a small servo.

The Arduino is running zaggo’s SphereBot firmware that is hosted on https://github.com/zaggo/SphereBot. GCODE is generated in Inkscape with Marty McGuire’s awsome unicorn export plugin (originally created for makerbot’s unicorn pen plotter attachment, but the generated code works great with the SphereBot firmware) and send to the bot via zaggo’s Processing app from https://github.com/zaggo/SphereBotSenderProcessing.

Unicolored eggs are so 2011 :-)

Gyroid Vase – beautiful Math Art

Thingiverse user Dizingof recently postet a host of beautiful Math Art that is very well suited for 3D printing. I must say the model below is generating exceptionally pleasant sounds when printing (mostly just perimeters and organic shapes).


Gyroid Vase 2012 – By Dizingof (50% smaller, with base) by garyhodgson

RepStrap on a diet: losing weight!

Over the last two months our RepStrap has undergone some serious upgrades to improve both print quality and speed. Meanwhile, one can hardly call this a RepStrap anymore considering its self-transformation into an even more Prusa Mendel look-alike.

The first part of the upgrade also caused the most noticeable effect on printing performance: Losing weight. I managed to reduce the moving mass on x- and y-axis by all in all over 1kg! My best guess before was about half of that… Our original x-carriage weighted 499g with all its 608ZZ-bearings, M8 screws etc. The printed carriage is only 70g including all four LM8UU bearings. On the y-axis the cutback was even more – it is now 610g lighter than before (using our slim LM8UU bearing holders).

The new x-axis is now Prusa compatible using a redesigned x-carriage with LM8UU linear ball bearings that Jonas and I build. We implemented several improvements over the default Prusa design like screw-clamped bearing holders, four bearings instead of only three and a more rigid base to allow for a higher belt tension.

The printed x-ends are an early version of another improved Prusa part we designed – take a look at the final version on thingiverse.

We achieve an excellent print quality on both of our RepRaps now. But we are curious how they perform regarding reliability and durability and started to crank out Prusa Mendel RepRap parts to sell on eBay. Our printers run day and night producing a compilation of Prusa parts that incorporates all improvements and revisions we made to our own machines.

This is the RepStrap in its current configuration printing in a heated chamber at ~50°C:

And that’s one of the produced parts:

Household repairs: fix a broken aquarium filter

Last night I moved my goldfish aquarium to my new home. During reassembly of the pump a small clip that secures the sealing bracket on my “EHEIM eco pro” aquarium filter broke. I had to quickly fix this since the fishes were waiting in their transport container and urgently needed some fresh and circulated water.

A great opportunity to test my RepRap workflow!

I removed the broken part, took some measurements, drafted a replacement in OpenSCAD and printed it on my RepStrap in ABS. From idea to completion in 45 minutes. I suspect that it would have taken longer to even find a spare part supplier. Order and delivery not included.

I published all files on thingiverse: http://www.thingiverse.com/thing:16119


RepStrap’s first child – another RepRap family tree

We have reached another milestone in our RepRap Projekt. Our RepStrap finished printing its first child. The Prusa Mendel v2 is assembled and working.

Last two challenges had to be overcome: Because the RepStrap still has no heated printbed the Prusa x-carriage had to be made from PLA. Which is not the best choice when faced by the heat radiation from a heatbed at up to 110°C and extruder temperatures beyond 200°C. Fortunately the carriage survived just long enough to get its ABS-replacement printed on the Prusa itself.

The same applies for the Extruder body: The PLA version did not withstand the heat long enough so we temporarily replaced it with the RepStrap’s extruder. This way we could print an ABS-extruder (seen in the picture below; there are still some warping-issues since the power supply was not capable of heating the printbed up to our desired temperature)

During the assembly of the Prusa, we decided to substitute some of the parts with better replacements than provided by the original Prusa repository. These are:

Replication almost complete

Our RepStrap nearly finished printing all neccessary parts for its first child, a Prusa Mendel v2. Only the Extruder is missing.

Yesterday Jonas made a short video of the RepStrap printing.

The Printer should primarily serve the purpose to make a ‘real’ RepRap-Prusa from printed parts. Since I managed to get reasonably useful (and mechanically robust) print results, step by step set of parts for a RepRap-Prusa was completed. Further calibration and optimization of print quality took place throughout the process.

Initially we started printing with a spool of PLA-90 from Orbi-Tech – a german supplier for 3D-Printer Filament. They sell this special type of PLA as beeing more heat resistant than standard PLA (getting weak at 90°C). Unfortunately this PLA blend is much harder to print, especially without a heated printbed. Although Orbi-Tech claims this “premium PLA” warps less than standard-PLA, we could not confirm that. So after a lot of fine Tuning and experimenting we switched to white Standard-PLA (from Orbi-Tech as well) for the remaining prints.

In the picture above all parts that are a little darker (light grey) were made of PLA-90. The bright white plastic is the standard-PLA.

Btw – Jonas started the RepStrap Build Log with part 1, describing our frame design.

Already printing!

Well, that’s awesome. I have never seen a 3D printer live before – and now this bootstrapped contraption on my desk is fabricating three-dimensional objects from plastic.

Absolutely fascinating.

Time passed quickly last week, so i’ll try to give a brief summary of the final steps to a working RepStrap 3D printer.

When our RAMPS-board and microcontroller arrived last week, the electronics were completed. The PCBs came pre-assembled, so i just had to stick the Stepper drivers in place, add a little cooler on top of each and solder some female connectors onto the stepper cables and opto-endstops. I crimped the hotend heat resistor and thermistor onto their cable instead of soldering due to high temperatures near the heater block.


As a power source i pulled an old ATX computer power supply from the junk box. There is a great page in the RepRapWiki on how to wire those up.

That’s for the hardware. On my Controller i installed Sprinter, a powerful and extensivley tested RepRap firmware. As a computer-toolchain i am testing SFACT (Slicer) and Pronterface (RepRap host software). For now i am quite happy with those.

In further posts i’ll go into more detail on the first extrusion, my first spool of filament and tuning all the parameters for good print quality.

My way into 3D printing

It must have been about a year ago that I (again) stumbled across the RepRap project. If you ask Jonas, this was probably the key moment, when it grabbed me to build my own 3D printer. Since I can remember we have dreamed up technical projects together that failed to realize due to an absence of a full featured mechanical workshop.

But being able to make free-form objects from plastic right on the desk, that changes everything.

A few months later I had read deeply enough - and Jonas infected. We started developing our own RepStrap - strongly inspired by the RepRap Mendel design. And this is the result so far:



The frame is entirely made of materials from a local hardware store. The linear bearings on all axes are based on standard 608-ballbearings, aluminum profile and threaded rods. Stepper motors, timing belts were purchased as well as the extruder and hot end for lack of time.

Only thing missing is the RAMPS-1.4 electronics that should be delivered from reprapworld.com these days.

Jonas and I will continue to document the ongoing build process and portray our work done to date.