Spare parts

The last week has been a fruitful one for me, at least as far as printing things on my MakerBot is concerned.  I’ve designed, created derivatives, printed upgrades, updated the firmware, installed the latest (super sweet!) RepG, and even performed some maintenance.

One amusing and useful experiment was the serial production of the Z Axis Wobble Reducers by MakiYoshida.  Using my current settings I printed one.  Then I installed it and printed a second.  The second was a visibly better product.  I installed the second and printed a third.  The third was no better than the second.  I installed the third and printed a fourth.  I then installed the fourth… and didn’t bother to print with it.

Right now I have two of these wobble reducers installed on the front right and back left Z axes.  Installing one on the most warped rod (front right, for those of you playing at home) helped significantly.  Installing the second on the back left helped a little more.  Installing the third on the back right didn’t seem to make much of a difference, and may have been worse than the second print.  I didn’t realize why until I tried to install the fourth Z axis wobble reducer.

Once all of the wobble reducers were in place I found that the entire Z stage was too mobile – and would easily shift from left to right and back and forth.  When you have no wobble reducers one or more warped Z axis rods will force the entire stage out of whack.  When you have four wobble reducers, there is nothing keeping the Z stage steady.  Thus, the optimum number is between one and three.  However, I noticed that having two diagonally across from one another gave me enough wobble reduction to improve the print, but not so much flexibility that it could get pushed around in the middle of a print.

Part of the problem is that this particular wobble reducer will allow nearly unimpeded side-to-side motion.  However, these wobble reducers are also fantastic at compensating for a badly warped rod.  I suspect that mixing different kinds of wobble reducers might be my best bet.  I’m not really that interested in installing one of versions that requires precision rods, since it’s just one more thing I have to source for my MakerBot.  Ideally I can mix and match these two types of wobble reducers to improve the overall print.  The Z-axis floating arrester appears to keep the captive nut more … captive than the Z axis wobble reducers I have installed.  This may provide the exact combination of rigidity and flexibility I need to optimize my print quality and minimize layer shift as I print.

The upshot of all of this, and the reason for the post title, is that I now have two Z Axis Wobble Reducers by MakiYoshida that I’m not using.  Do you need one or two of these and feel like saving yourself the print time?  If so, drop me a line.

PLA versus ABS, initial thoughts

I’ve yet to print something big as I am still working out some of the black ABS specks from the extruder.  Of the things I have printed so far, it looks like the PLA feels more brittle than the ABS.  ABS will flex slightly when pressure is applied, whereas PLA seems to resist until it snaps.  I like that I’m not wasting PLA by building rafts any more, now that the raft option is turned off in Skeinforge.

PLA takes a little longer to cool which has a few effects on a build.  If the layer underneath is still molten and saggy, the layer on top will sag too.  When an ABS part is done printing you’re basically ready to remove it.  I find that a PLA object can still be a little soft for as much as 15-30 seconds after a build.

I’m loosing finer details on prints.  I just reprinted the toy fire truck swivel.  It turned out really nice except for the actual swivel part.  Unlike the ABS model where the swivel snapped off immediately, the PLA model’s swivel was pretty solid (once it cooled and hardened).  However, it came out too thin!  As I watched it print, the PLA was so warm that it just stretched back during parts of the print – meaning the extruder nozzle made the full circuit, but the plastic stretched a little and pulled back closer to it’s last position.  Uh, that may not sound like the clearest example.

Suffice it to say that small round details such as the swivel didn’t turn out very well.  Which is pretty odd since similar small round detailed parts on the teapot did turn out well.  I suppose part of the reason is the swivel was one small round part where the next hot layer was set down on the prior still hot layer, unlike on the teapot where the thin round spout would be drawn, then the body, then the thin round part for the handle and back over the body – so that by the time the next thin layer was set down the prior layer had cooled.

How to caliberate and tune Skeinforge for your MakerBot

Update 4/21/2011: This post is more than a year old.  Since then I’ve created ProfileMaker which should eliminate the need for almost all of these steps.  Share and enjoy!


A MakerBot Operators group member, Cliff Biffle, has posted a very comprehensive list his Skeinforge tuning and calibration procedure.  Cliff owns MakerBot #69 from Batch#4, named Kallisti.  Although Cliff has a MakerBot Skeinforge tuning page, it looks like he hasn’t had a chance to transcribe his lab notebook yet.  However, Cliff has been patient enough to type up his Skeinforge calibration information twice for the benefit of the MakerBot Operators group.  Here are links to his original postings on 1/6/2010 and another on 1/15/2010.

The other day I wanted to go back and find his calibration procedure, but had forgotten who had posted it, when it was posted, and which thread it was posted.  Having now found it, I don’t want to lose it again.  It turns out that Cliff had posted it once in response to my own question about dialing in Skeinforge settings in order to print up lego pieces!  Thanks Cliff!

I’ve mashed his two posts together to put together this checklist:

  1. Adjust Temperature:  Skeinforge -> Raft
    1. Find a temperature that allows you to extrude quickly and easily without too much popping or smoking
    2. MakerBlock: I have one of the 1mm thermistors, use the Zach 1 temperature table, and print reliably at 220
  2. Adjust Layer Height: Skeinforge -> Carve -> Layer Thickness (mm)
    1. Cliff uses the default of 0.36
  3. Adjust Extrusion Speed: Skeinforge -> Speed -> Flowrate PWM Setting (if PWM Setting is Chosen)
    1. Make the extruder go slower and lower the PWM value
    2. Cliff uses 230
  4. Adjust Feedrate: Skeinforge -> Speed -> Feedrate (mm/s)
    1. Too fast and you lose the fine detail and introduce errors.  Too slow and it will be too… slow.
  5. Adjust Skeinforge to understand how much plastic is being extruded
    1. Raise the Z axis and perform a test extrusion into the air
    2. Do not pull or stretch the extrusion
    3. Using calipers, measure its diameter, D
    4. Go into Skeinforge and look up the configured layer thickness, T, from Skeinforge -> Carve -> Layer Thickness (mm)
    5. Divide the diameter by the thickness, D/T
    6. Set Skeinforge -> Speed -> Extrusion Diameter Over Thickness (ratio) to D/T
    7. Print a test piece with thin walls
    8. Measure the width, W, of a single extrusion somewhere above the first few layers
    9. Divide the width by the thickness, W/T
    10. Set Skeinforge -> Carve -> Extrusion Width Over Thickness (ratio) to W/T
      • Cliff said this setting may be in Fill, but I think he may have meant Carve

    11. Nick Ames has suggested setting Skeinforge -> Carve –> Extrusion Width over Thickness (ratio), Skeinforge -> Inset -> Extrusion Perimeter Width over Thickness (ratio), and Skeinforge -> Speed -> Extrusion Diameter over Thickness (ratio) all to the same value.  His post also has a lot of information and links about other potential Skeinforge tuning resources.
  6. Adjust other Skeinforge settings
    1. Overlap of grid corners
    2. Infill density
    3. Unpause
    4. Stretch