Post has attachment
I would like to encourage you to go over to the Hackaday.io site and like/follow the project to help it get the standing it deserves. It does require you to have an account or link up one of your existing accounts. It looks like I could win real money which would go a long way to my mad scientist fund.

Update: I am done with firmware for a while and I am finalizing HELIOS v0.2 and hope to have full renders and animations out early this week for peer review.

Post has attachment

Post has attachment
For those of you that are disturbed by my dislike or lack of endstops. This is one sensor that will rule them all.

Post has attachment
Everything is pretty much falling into place. I am literally down to my last component. The piece of plastic between all the steppers is giving me fits. The biggest problem is how should I do the belt tensioning. That back stepper especially need some redirects at the leading edge of this piece. None of that is really hard but none of the solutions are really speaking to me. The real trick is to keep the redirects/tensioners from limiting the arm movement any more than the steppers already have. I may have to bite the bullet and further limit the arm range. It won't really change the printable area although it will move the arc where you have to arm switch.

BTW, if I can pull this off then the whole printer can fit in a 200mm/8" cube.
Photo

Anyone have any Make contacts? I want to get into the Bay Maker Faire but I obviously didn't know this would be so awesome until it was too late.

Post has attachment
Prototype take 2 is coming along. I was able to shrink all the pulleys down but keep my mechanical ratios. (Magic...don't ask. Actually, I used the compound approach. There are a crap ton [crap ton=4] of 50x100 compound pulleys.) That gives me a super small stowed envelope and an even bigger print area. (20% bigger and it fits boring squares and rectangles better too.)

I essentially have the arms fully redesigned. I have to spend a lot of time on the electronics box and top support. I think I can shave over 50mm off the z with a aha moment I just had.

Edit: On the build platform I was able to move the arcs where arm swaps have to be forced way back. That is what the two semicircles are on the drawing. I thought it was import to explain why there were circles there. :-)
Photo
Photo
4/18/17
2 Photos - View album

Post has attachment
It pains me a bit but I have decided to make OpenJSCAD the parametric modeling language for the project. I was super stoked about ImplicitCAD and I ultimately prefer its modeling approach...however, the one online tool I found for it had severe resolution issues. At the end of the day I don't want to make people install Haskell just to modify a electronics board mount hole pattern. Additionally, all the parts are quite simplistic and don't need the infinite resolution of ImplicitCAD. Thank you +Julia Longtin for telling me about ImplicitCAD. Thank you +Mark Moissette for offering OpenJSCAD services.

I hope to start the OpenJSCADing in about 2 weeks. If anyone else wants to jump in leave me a comment below. I will be producing beautiful pdf's of all the parts prelabeled with parameter names and default values. It would be very easy for work to be done in parallel or piecemeal at that point.

1: will be just to have all the parts modeled.
2: will be to build a static assembly for visual fit verification.
3: will be to build a dynamic assembly because because.
4: will be to build in check and verification to tell the user if they are picking a combination of parameters that is not advisable or doesn't work.
5: add sliders to the interface for people to quickly examine different machine configurations.
6: print out a list of stats to help people pick a design. (print volume, print area, biggest square, worst resolution, rough estimate of arm deflection, etc)
7: print out a BOM for the vitamins which are a function of other parameters. This is a big one. Some design might require a screw that just doesn't exist. Sometimes the design has to be tweak one way or the other to make sure screws can get to nuts or they don't stick out too far past the nut.

I think 1,2, and 5 are reasonable to get done before launch in a month. I will probably take a step back after launch and suck in all the info from peoples build problems to make one more iteration before I go super crazy on all the build tools.

Let me know what you think about what else you might like in the automated design tools or what you think would be a waste of our effort.

Good and bad news. Printing across arm modes is a none starter. If the effective arm lengths differ by delta then the registration error is 2*delta*sqrt(2) at the worst place. (If I did the math right. My tests seem to line up with this.) You may say "Hey Nick just make the arms the same length." And I would say the arms are the same length but with droop and rail misalignment the effective lengths may vary slightly. As you can tell from the expression it actually get magnified. You may also suggest I just calibrate the machine. I could but it would only be valid for the first layer.

All that said, there is good news. I will re-institute lazy arm mode switching. In this case when you go to do a skit on a part that lays across a singularity it will do one mode switch and then stay in the mode until it goes to do a part all the way on the other side of the build platform. If you break your huge prints into at least 3 zone, you will be golden. The other trick is bed leveling is specific to arm mode. To handle this, I will do probing across the platform in both modes and maintain two leveling solutions depending on arm mode.

Post has attachment
Have HELIOS will travel.
Photo

Post has attachment
I went looking for some artifacts. To find them, I printed a square spiral vase at 90mm/s so there was a chance it would hit just the right motions. I definitely found some. The ones pictures are just noticeable to the touch. When I get some time I will print at 30mm/s and see how the artifacts compare. (To be clear, I personally don't care one way or the other but I know some people really like a perfect surface so I want to make sure I do my due diligence.)
Photo
Wait while more posts are being loaded