Sailfish + Print-O-Matic = Wow

dutchmogul's Keep

dutchmogul’s Keep

Last night and tonight I have been tinkering with Flexo, my Thing-O-Matic, to install Jetty’s Sailfish firmware update.  The setup text is very helpful, but not particularly optimized for a Thing-O-Matic with a MK6 extruder.  I also found that I had to heavily modify the “start.gcode” and “end.gcode” to keep my bot from shaking itself to pieces. 1

However, the results of combining Sailfish with ReplicatorG’s Print-O-Matic are impressive.  Dutchmogul’s Keep, pictured above, was printed with a layer height of 0.15mm, 10% infill, all with a stock 0.5mm nozzle.  That’s the very best resolution I’ve managed with my Thing-O-Matic, ever.  At 0.15mm thick layers, I have a difficult time seeing the layers if the model is more than 6 inches from my face and the ridges on the vertical parts is difficult to discern with a fingertip.

Now that I’ve finished that print I think I can push my ‘bot’s print resolution farther.  I’m willing to bet I can print down to 0.10mm layer height if I re-applied the Kapton tape and did a better job of leveling the print bed.  Also, one of the acknowledged challenges with 3mm filament based extruders is the pressure buildup that can lead to imprecise plastic deposits when dealing with non-contiguous parts. 2  Dialing in the Skeinforge “Retraction” or the Sailfish “Deprime” settings dialed in would really help with complex prints.

  1. The default “start.gcode” tried to home to the XY maximums and Z minimum – where there are no endstops.  The default “end.gcode” did the same thing for some reason.  I also had to modify the “start.gcode” so that the print head was better positioned for starting a print. []
  2. By this, I mean that when you’re printing two or more features that are not connected, say for example a table that is legs-side-up, an extruder will naturally ooze a little bit of plastic as it travels between legs, leaving a thin spider web like strand.  The “retraction” setting in Skeinforge is used to combat this, and does so fairly effectively in the case of a stepper based extruder.  That setting reverses the extruder motor quickly just after the print head leaves a leg and then quickly moves extra fast forward as it gets to the next leg, which prevents the spider web effect. []

ProfileMaker update

I’ve made a few slight modifications to ProfileMaker.

First, after testing it seems that the apparent optimal W/T is approximately 1.67 or so.  Using this value for target W/T setting has lead to my cleanest and most sturdy prints.  Admittedly, I haven’t done proper science-testing to determine the optimal value – but I do not believe it could be far off.  Thus, I’ve removed this setting option and ProfileMaker will calculate the optimal thread width based upon a W/T value of 1.667 and whatever layer thickness you choose.

Secondly, I’ve removed the automated profile creation system.  Not many people seemed to be using this and this feature gets quickly outdated a new versions of ReplicatorG are released.  While the five critical settings provided by ProfileMaker have been unaffected by recent changes to ReplicatorG’s Skeinforge engine, lots of the other settings have been juggled around.  Skeinforge is too much of a moving target for me to try to keep up.  Ideally, you’ll find a good set of profile settings you like and will still be interested in using ProfileMaker to adjust the five critical settings.

It is my hope that by making these minor changes everyone will have an easier time using ProfileMaker and get better results.

Please let me know what you think!

Alas, poor ProfileMaker, I knew you well…

I launched ProfileMaker on March 20, 2011.  After serving up more than 3500 awesome Skeinforge profiles, my website hosting company has actually started to complain that I’m using too much of the shared hosting environment’s resources.  As a result, I’m going to have to disable temporarily and potentially discontinue ProfileMaker.

I suspect it is the Skeinforge profile creation feature that has caused the most trouble.  Every time this feature it used, it creates a temporary file and incorporates it into a full Skeinforge profile, which is then ZIP’ed for e-mail delivery.  It’s possible that disabling this one feature may dramatically reduce my site’s server load.

ProfileMaker: More than 209 profiles and 58 people served!

More than 58 people have generated profiles with ProfileMaker1 !

Seriously, you need to stop printing calibration cubes and fiddling with Skeinforge.  Find out more than 58 people have discovered and start using ProfileMaker to create your Skeinforge profiles.  You can either manually enter the values provided by ProfileMaker into Skeinforge or have it e-mail you a ready-to-slice Skeinforge profile.

Always wanted to try printing at 0.25mm per layer or with 1.75mm filament or in PLA?  Your ‘bot can do it.

  1. Powered by SCIENCE! []

I would like to thank my agent…

…the Hollywood foreign press…

Actually, huge thanks to Dave Durant for the math and answering questions, Renosis for exhaustive testing and feedback.  Thanks also to all of the other beta testers of whom there are too many to name.  Um, they’re playing the music,…  uh, uh, Honey – we did it!  Um…  Free Tibet!  and… um…  I’m King of the World1 !!!

In seriousness, since the launch of the first ProfileMaker v1.0 less than a week ago there have been 152 profile settings generated and the beta testers generated 270 profile settings through ProfileMaker v2.0.  Version 2.0 incorporates many of the things mentioned in the recent poll. 2  Here are some of those improvements:

  • ABS as well as PLA
  • Works with 1.75 and 3mm or any filament diameter you choose
  • Ability to change the feedrate, the mysterious gear swell, and gear diameter

I’ve already begun work on ProfileMaker 3.0.  If you want to help as a beta tester, or get the math involved, or want to help kick the tires of the user interface please drop me a line or leave a comment.

  1. of web based 3D printing calculators that solve for flowrate for stepper extruders []
  2. Still active as of right now – but get your votes in if you want to let your voice be heard []

Creating an Online Profileinator

I’m always juggling multiple projects – that seems to be the mark of a Maker and/or a Maker with ADD. 1  I’m looking to tinker with a clockwork spider, an open source disc shooter, possibly an open source shooter compatible with Nerf darts, and now…  I’d like to create a web app that aids in the calculation of good profile settings.

I’ve downloaded a bunch of different Thing-O-Matic, Cupcake, PWM and RMP calculators off of Thingiverse and am tinkering with them to get a better understanding of how they run these calculations. 2  Once I understand these, I’d like to make a very simple front end for these calculations.

If you have a good idea of how these things work, please leave a comment or shoot me an e-mail through the contact form.  Thanks!

  1. Which, in my experience, might just be a redundant statement []
  2. I’ve put out the call to Dave Durant to help me get a better grasp on these calculations… []

MakerBot woes

Building my MakerBot and getting it printing reliably was challenging, but totally doable by a technical novice such as myself.  I have lots of people on the MakerBot Operators group to thank for their patience and help in getting my MakerBot online. 1  Looking back, I spent about a month building and then calibrating my ‘bot. 2

It’s easy for me to forget that first month of occasional frustrations and triumphant victories, now that I’ve been printing successfully for more than eight months.  It actually makes me a little sad when I read people writing about their own frustrations and how they’re ready to throw in the towel.  The most recent example was noobcake getting frustrated with her ‘bot and getting ready to sell it off in parts.  Thankfully, Spacexula swooped in to help her out.

This brings me to We Alone On Earth‘s recent post.  WAOE are a group of introspective, philosophically and technologically minded twenty-somethings.  To give you you an idea of their frustration with their ‘bot, the post was entitled, “MakerBot: not very much fun at the moment (caveat emptor)”  Yikes. 3  I realize that WAOE has revised their original post several times since the original publishing date – but they have a lot of legitimate concerns.

WAOE list off seven problems with the MakerBot.  I’m not going to refute these points – but rather offer another perspective on them.  After several updates, WAOE offer additional comments, I’ll include them here in “[]”.

  1. The PTFE is prone to melting.  [WAOE expects the new MK5 Plastruder will resolve this issue]. I have never heard of a PTFE barrier melting.  I’ve heard of them deforming from a blockage and had one develop a clog which I had to remove.  Several people have purchased MakerGear PEEK replacements – but these are far from necessary.  I clogged my first barrier once, cleared it, clogged it again, and am now using a slice of it as an insulating washer.
  2. Inexplicable printing behavior due to noise.  [WAOE fixed this issue by twisting wires and installing a resistor]. I’ve never had this problem, but I know others have.  Like WAOE, I’ve heard of people fixing these issues by twisting wires, using resistors, or ferrite beads.  Perhaps my workstation has less electronic noise, but I haven’t had to do any of these things.  Then again, perhaps my prints suffer from a certain degree of noise?
  3. Printing large objects is hard without a heated build plate.  [WAOE notes this isn’t an issue if you’re good at soldering]. Totally true for ABS, but not PLA.4  However, this is really a problem with the print media – not with the printer, right?  ABS will warp as it cools, unfortunate but true.  I’ve had less warping problems in warm weather or during with a second print – basically when the build platform is already warmed up.  Zaggo’s printruder is one of the largest things I’ve printed.  Interestingly, his design takes into account that certain parts are expected to warp. 5  Or check out Clothbot’s train track – it was designed with a lattice/correlated bottom to prevent warp problems.  Plastic warps – but with careful and thoughtful designing, this shouldn’t be a limitation.
  4. The heated build platform is difficult to build and requires a relay kit. [WAOE notes this isn’t an issue if you’re good at soldering]. I can’t dispute either point.  I just got both and haven’t had a chance to assemble them yet.  I’m assuming the heated build platform, which requires SMT soldering, will be challenging.  Frankly, fear of SMT soldering was the big reason why I didn’t jump into buying a MakerBot sooner. 6  The MakerBot HBP is just one option for a heated platform – there’s several others out there.  Don’t like SMT soldering?  Try out Rick’s platform over at MakerGear.  More into DIY?  Well, use the plans posted for any of several other variations.  As for the relay kit – it’s not a requirement – but it will prevent MOFSETs from burning out on your extruder motherboard7
  5. Calibrating Skeinforge is hard.  [WAOE notes this is still an issue]. I like to use the word, “challenging.”  A better way to look at MakerBot calibration is that you get out of it what you put into it.  I have my MakerBot tuned to the point that I get reasonably good looking durable parts.  Sure, I could spend more time and get even better looking parts.  However, once I got it printing reliably I was much more interested in printing new things than refining the printing process.  I’ll get around to improving the print quality even more – but I’m having too much fun right now.
  6. The Plastruder MK4 feed system is unreliable. [WAOE expects the new MK5 Plastruder will resolve this issue]. Getting the tension on the MK4 idler wheel is just one of those aspects of my MakerBot I had to experiment with and get just right.  I’ve been printing reliably for eight months using the same idler wheel and gear.  With proper maintenance, flossing the extruder, and clearing chips out of the extruder the current setup is serving me well.
  7. The threaded rods are of poor quality. [WAOE are getting new threaded rods, which should fix their problem.]. Of my four threaded rods, one is definitely warped and two have very minor warps.  By experimenting, rotating them just so, and printing a few wobble arrestors I’ve eliminated most of these issues.  You can definitely get more expensive and straighter threaded rods and improve your build quality.

A MakerBot Cupcake CNC kit is not for everyone – but the kit can be build and operated by anyone who is willing to invest the time to do so.  It is a cheap, hackable machine that is literally going to be just as useful as you make it.  Want less warpage, higher resolution, more reliable extrusion?  You can buy an upgrade, build one from their plans, or design your own solution. 8  Want a CNC mill, CNC pencil, or CNC music box?  Design the very first one!  Then again, you don’t have to do any of these things.

A MakerBot kit is just a platform for your creativity.  It’s just that big. 9

As for you, WAOE, if you want some help – drop me a line!

  1. If I had an acceptance speech, I’d go on and on while the music played. []
  2. My first successful print was on 12/31/2009. []
  3. Don’t get me wrong – I love my MakerBot, but I readily acknowledge its limitations. []
  4. From what I heard.  :)  []
  5. It was designed before availability or widespread use of heated build platforms. []
  6. Well, that and a little thing called “money.” []
  7. Did I get that part right? []
  8. I haven’t installed a single non-printed upgrade. []
  9. Or that small.  :)  []

MakerBot Calibration

After my design-print failure I thought it was about time I recalibrated my ‘bot:

  1. Starting from scratch, I’m using the 0.5mm test pieces from Spacexula’s calibration set.  Before starting I set Skeinforge->Carve->Layer Thickness (mm) to 0.4.  Thus, I came to test piece 815.1.  The piece came out well, but I noticed that where the Z axis raises there is some slight blobbing and just before the blobbing, some sparse areas I can see through.  Otherwise, great interlayer adhesion.  The height of the piece is 10.25, 10.31, 10.30, 10.55.  Throwing out the high and low, there’s an average of 10.305mm.  It is 22.29mm x 22.29mm.
  2. Skeinforge->Carve->Layer Thickness (mm) to 0.38.  Test piece 815.2.  Again, slight blobbing, very small sparse areas and great interlayer adhesion.  Piece height is 10.39, 10.40, 10.18, 10.19, we’ll call this 10.29mm.  It is 22.41mm x 22.13mm.
  3. Skeinforge->Carve->Layer Thickness (mm) to 0.36.  Test piece 815.3.  Slight blobbing, very small sparse areas and great interlayer adhesion.  Piece height is 10.25, 10.31, 10.14, 10.30, we’ll call this 10.275mm.  This one was 22.33mm x 22.29mm.

Spacexula to the rescue!

This poor guy gal known as “noobcake” had gotten so frustrated with his her MakerBot CupCake CNC that he she was getting ready to either strip it for parts to start work on a RepRap or just sell the entire thing on eBay at a loss.  It read like most of his her problems were coming from improper Skeinforge calibration.  Anyhow, that made me very sad.

And then I scrolled down.  Spacexula helped noobcake get his her Skeinforge calibrated and MakerBot printing reliably.  That’s really really fantastic, man.  Kudos.

Edit:  Apparently I have as much difficulty telling girls from guys online as in real life.  Sorry noobcake.  My bad.