Skeinforge Support.. Part 2

Trying out the Support structures in Skeinforge again this afternoon. If you missed part 1, you might want to give that a look first.

Today I’m using an actual, decent model, the Jewelry Hand at Thingiverse made by the the folks at MakerBot Industries. If you look at the pictures of this hand, the thumb is some serious overhang and, hopefully, perfect for support testing. The MBI folks suggest putting a book under one side of the bot (changing the direction of “down”) but that’s too easy.

Cross hatching is off and I’ve got the support gap set to 5.0. Skeinforge says print time will be 51 minutes. (I also skeined it without support enabled: 43 minutes)

It’s printing now. Third try actually. One thing I noticed right out of the gate is that the raft area under the support structure is not only not larger than the area of the support structure, it actually seems a little smaller. And it’s the very first thing to print.

Here’s a picture of try #2, about 30 seconds before it tore the support structure up. I had been crossing my fingers and hoping it would stay down but it just was not meant to be.. 😦

My printing routine is to load some gcode into ReplicatorG (the software that controls the bot), sorta eye the center position and manually move the build platform and Z stage to that point then click the Build button.

Unless you’re printing something with a large footprint, the X & Y centering isn’t really critical but the Z always is – you really want this to be as close to “correct” (which is a whole ‘nother blog post) as you can get it. Normally, this isn’t a problem because the raft is larger than the object and I’ve got maybe 5-10 seconds to manually turn one of the Z pulleys and get it right, before the real printing starts.

With the support raft not being oversized and being the first thing to print, I no longer have my 5-10 seconds of grace time – I’ve got maybe 1-2 seconds, the time it takes for the build platform to move over to that spot.

My first couple tries at this fail. The Z is a bit too high and things don’t stick or there’s a little blob of plastic on the nozzle that drops off and catches on the print head the next time around. This try, the 3rd one, I’m a lot more careful and it’s looking pretty good, 15 minutes into the build.

22 minutes in: As I watch this print, it occurs to me that my prints always start on the left of the build platform and that’s where the troublesome support raft is. I wonder if rotating the object 180 degrees around Z would have made this easier. I guess it depends on if skeinforge just likes to print left-to-right, which seems likely, or if it likes to print the support raft first, which seems less likely. I’ll have to remember to test this.

30 minutes in: I think skeinforge probably wants to (or should want to, if it doesn’t) print the support parts of each layer first, before the object parts of the layer. The idea of the support structure is to have something for the filament to hold onto instead of printing just over open air so doesn’t it have to do the support parts of each layer first? It’s not clear that this matters for the raft parts, though – still have to test that out.  (update: more testing has revealed that yes, skeinforge prints the raft from left to right (regardless of support stuff) and that support structures print first on each layer)

40 minutes in: I really wish I had enabled Comb and had gotten around to getting Oozebane dialed in or, better!, had a stepper extruder. It looks like this print is done with support structures and is down to printing the less agressively overhanging bits. RepG says 24 minutes left..

50 minutes in: I think RepG lied to me about time left – it now says 7 minutes left and is counting down faster than 1 second/second. I think it’ll be about 3-4 minutes. The thumb finished a while ago and the pinkey finished a minute or 2 ago..

54 minutes in: done!

The things I wanted to look at today were disabling cross hatch and bumping up the gap between the support and object. I was also going to drop the flow rate (PWM) ratio bit more but, well, forgot about that.

Disabling cross hatching made a HUGE difference. Last night, with cross hatch enabled, the support structure was pretty tough stuff. Not as tough as the object itself but I had to put a bit of work into it to break it up with my fingers. Today, cross hatch disabled, I can easily crush the support structure between my thumb and forefinger.

Also, as expected, looking through the support structure, shows lines instead of tiny boxes. This is looking through about an inch & 1/2 of it:

The gap ratio got turned up to 5 this time and you can see there’s definitely more space between the two support structures and the object. I think this is probably LayerHeight * WidthOverThickness * Ratio + X, where X is likely one perimeter width.

Overall, I’m close to thinking the things I was thinking last night. Unless you know you need a strong support structure (more than just an anchor for filaments from the object), disable cross hatch – it’s far, far easier to remove that way.

For the support gap, last night I said:

I think the optimal value here really depends on your feed rate – set this value so you get a distance that you can easily cross (easily cross = not have the filament drop much, if at all) at the speed the platform is moving. I think I’m never going to use 1.0 again; 2.0 is probably my lower limit now. I tend to print at fairly fast feeds and predict I’ll keep this around 5.0 or so.

I think I’m sticking to this. The farther you can get the support from the object, the faster it prints (though not a huge difference here) and the easier it will be to clean up later. I think oozebane (or stepper extruder) might also factor into this but I’m not sure how..

Here’s the finished object, as always, not really cleaned up very well..

That’s about it for this round. I’ve got an idea of what to try next but haven’t decided if it’s stupid or not. I suspect it is. I’ll see how I feel about it tomorrow…

Advertisements
This entry was posted in Makerbot, skeinforge and tagged , . Bookmark the permalink.

5 Responses to Skeinforge Support.. Part 2

  1. ppirilla says:

    I unfortunately do not have a makerbot of my own to try it with but I wonder if cross-hatching the supports will allow you to get away with an even larger gap, and thus use less material. It probably won’t save you any time, but could still be an advantage if you’re printing something with a more expensive plastic.

    • Dave Durant says:

      I don’t think hatch/no-hatch helps with the gap it can span. That is, when you’re printing an overhang, you’re more concerned about moving quickly enough (feed rate) to get to the support structure before the filament droops than you are about how the support structure is built.

      Having cross hatch enabled or disabled is more, I think, about the strength of the support structure.

      For part 1 with the cubes, cross hatch was definitely overkill. The support structure was really big and even with the lower density that non-cross hatch produces, it would have been pretty sturdy.

      For the hand on part 2, the support structure had a pretty tiny foot print. Even though I managed to pull that one off without cross hatch, I think I’d at least think about enabling it in the future if an object had support areas that were also that small or smaller. That’d at least drop the odds of a stray blob of plastic knocking it over and ruining the print. With bigger support structures, that’s not as much of a concern because they have a solid base.

      I think that’s right anyway.. I’m hardly an expert here. 🙂

  2. Pingback: Skeinforge Support.. Part 1 | DaveDurant's Blog

  3. NanoZap says:

    Pardon in advance for my crytic mood. I’ve found the Durant Directives useful, so thank you. Part variability, an affordability trade-off along with builder capability, combine to create a unique techno-sherlock puzzle, meaning the nature-nurture phenomenon applies to our bots.
    But the brains, Skeinforge has many facets, each written no doubt by the code junkies with a specific purpose. If there is a purpose summary for each variable we can tweak, written by the gray matter alogorithms of the conjurers, where would such arcane information be hidden? Yes, an official manual.

  4. davide says:

    Hello! I don’t know if you still answer to this page, but I try…
    I have some problems with Skeinforge and bridge speeds. I was trying to find out the best speed to extrude bridges and I’ve found out that speed wasn’t changing. Even if I write “0.1” in bridge speed, the printer still extrudes at the same speed. Is it possible that skeinforge doesn’t recognize that it’s a bridge? How can I solve it? Thanks

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s