Archive

Monthly Archives: May 2013

Now that I’m free of final exams – and not a small amount of unpacking – I can write a bit again!  To start, about three weeks ago this got published.  It’s the first of two pieces I was asked to do for Make’s 3D Thursday series, and I was excited to share what we’ve learned from running that many machines with such a high throughput at the Invention Studio.  The second piece should be up in a few more hours, and covers a bit of a different subject.  Additionally, I hope to have another big post on what I’ve been doing with the printer-formerly-known-as-HPRINT online soon.

In the first piece I wrote, I got some comments regarding what I believed to be oversights and deficiencies in the design of MendelMax-style printers.  Without further ado, here’s a list of the aforementioned issues:

  • Insufficient Constraints: In some design variants mechanisms are left underconstrained and are able to move excessively, causing artifacts to occur in prints.  Some machines have under constrained Z-Axis screws, while others may have under constrained X-Axis carriages and beds.  Taking care to ensure sufficient constraints exist on a machine is critical to keeping undesired movement from occurring, impacting print quality.  In MendelMax designs, it’s usually the Z-Axis screws that have this issue.
  • Excessive Constraints: In other cases, MendelMax variants were made with care to avoid issues that arise from insufficient constraints and instead went too far.  More than three constraining features is usually excessive for a plane, particularly if two are hard constraints with no allowance for movement and the third is able to make up the difference with a small degree of movement.  This can get tricky because you can have more than three constraints, and they can all be capable of small amounts of movement or deformation individually;  care must be taken to only allow this movement in certain directions which are irrelevant to the positional accuracy of an assembly.  Most often, this pops up in having too many bushings, bearings, shafts, and very occasionally nuts or bolts.
    In MendelMax designs, this can occur in all of the above areas depending on the machine; in AO-101s, I believe some of the issue crops up most frequently in the nuts for the z-axis screws and the leveling screws for the bed.
  • Ill-Planned Drive Systems: Many printers use very small neoprene belts for their drive systems strung over long distances with tensioners included as an afterthought, if they are included at all.  This leads to a great deal more backlash than would otherwise be present in machines which could be avoided by using wider belts across shorter distances with spring tensioners, or at least printed clip-on tensioners (and if you wanted to splurge, fiber reinforced belts that don’t tend to stretch as much).
  • Poor Material Choices: Some hot ends are made entirely of aluminum; while this is great for thermal conductivity, this is bad for both assembly and disassembly.  Aluminum parts have a tendency to bind as they rub against other parts made of aluminum, sometimes badly enough that they’re impossible to separate without destroying both.  Switching to materials that either bind less or more easily slip against each other can reduce this annoying occurrence.  In a similar vein unrelated to MendelMax machines, anything using wood as a main structural element will change shape ever so slightly under varying humidity; this means calibration may need to occur much more frequently than machines that use metals and plastics for their structure.
  • The Combined Effects of the Above: Let’s say you have an under constrained Z-Axis screw and shaft setup and over constrained nuts for the Z-Axis screws.  This means the screws both wobble slightly as they rotate, and as a result the nuts drag the XZ-Assembly along on the XY-Plane as a result.  This is going to result in worse prints than either problem individually.  Add to that some stretchy belts, and the problem is magnified even more.

That isn’t to say these can’t be mitigated in machines that have these problems.  With a little bit of adjustment beyond the setup work already done by LulzBot on the AO-101s before shipment, we were able to almost entirely eliminate these problems with no additional parts except for a few printed tensioners.

-Jeff Out