ShadowRam Wrote: ------------------------------------------------------- > Pretty awesome > > Firmware could also be a challenge since I have > never modified firmware before. > > Wait what?!? > > What are your programming skills like? > > You will need to implement all the complex math > Annirak posted, but just not get it working, > but write the code iby nicholas.seward - Delta Machines
Here are details on everything but the filament drive which will be easy enough to finish up this weekend. The 12 pieces that make up the arm will get laser cut on monday and I should have enough vitamins by Monday or Tuesday to do a simple motion test. Very exciting times. Here I placed some rough boxes to make sure I can fit the power supply and the controller. I will have to print some moby nicholas.seward - Delta Machines
CAD is done!!! Most all the orange plastic parts are already printed. The white arms will be laser cut on Monday. All the vitamins have been ordered in duplicate (gotta replicate) and are racing towards me as I type. I am so excited. This thing might breath as early as next week. It was 20 days ago that I seriously started thinking about making a 3D printer that has no linear rails and lowby nicholas.seward - Delta Machines
First, I have ditched the quadlink idea and have gone back to the simple two link arms. I just couldn't get my head around a simple design that didn't send my vitamin count through the roof. Not a rendering! (I probably shouldn't have removed the background in GIMP so it would seem real.) I couldn't just let myself wait to get my Polypropylene laser cut. I had to CNC out some 1/4" wood to seby nicholas.seward - Delta Machines
I just realized that if I go with the quad-link arm above that Simpson will look like Bart Simpson. I am ordering yellow filament. Update: I have ordered all the non-printed parts. The printing is going great. I just have to finalize the quad-link design and get some oversized parts laser cut.by nicholas.seward - Delta Machines
I came up with an improvement that can get rid of the the big spring. I was thinking about the arrangement above with strings in tension along a and b. Ideally a+b would be a constant but that only works if c/d is infinity. However, I did happen to notice that as c/d increased the amount that a+b would vary over the range of motion would decrease. This got me thinking that maybe I would setby nicholas.seward - Delta Machines
Annirak Wrote: ------------------------------------------------------- > I think I may have come up with a way to get rid > of the spring. It means adding another piece to > the arm. I'm not convinced this is a good idea. > My test showed it performing well on the lift, but > poorly on the drop. This would probably be helped > by gravity on the toolhead. > > I also caby nicholas.seward - Delta Machines
ShadowRam Wrote: ------------------------------------------------------- > Couldn't all the cons of a bowden be solved by a > small light incremental encoder at the hotend? > > So many joints, > > I think the number of tolerances will add up, and > I'm worried you will have a tough time getting 1mm > accuracy at the tool end. > > Interesting project though, keepby nicholas.seward - Delta Machines
I couldn't contain myself. Here is a snapshot of the 3rd redesign of Simpson. He is looking pretty good. I need to add in some string, springs, magnets, hall effect sensors, spacers, bolts, and nuts. The spring will attach to the short side of the upper arm and extend down to somewhere close to the stepper. The two short ends of the arms will be pulled together with a block and tackle setup.by nicholas.seward - Delta Machines
I have been agonizing over the extruder setup as of late. I have been waffling between 3 options: 1) Bowden, 2) Lightweight all-in-one unit, and 3) Flexible shaft driving a hobbed bolt on the effector. Option 3 is not proven. I would like to get this flexible shaft. The one that is .187x20". It costs $130 and still the shaft will have an angular deflection while loaded of 5 degrees over theby nicholas.seward - Delta Machines
Annirak Wrote: ------------------------------------------------------- > The problem with the spring is just the range of > force required to move it. When the arm is at > either extreme, the line is short and the spring > is compressed. When the arm is in the middle, the > line is long and the spring is extended. > > Because F=kx, if you want a reasonable force at > thby nicholas.seward - Delta Machines
Annirak Wrote: ------------------------------------------------------- > Well, I rigged up a test with meccano. It looks > like the spring is going to be a problem. A > constant-force spring would solve the problem > nicely, though I thought about constant-force. I initially ruled it out due to price but it isn't that expensive. What kind of problems did you have with the spring?by nicholas.seward - Delta Machines
I have been working hard on the 2nd major redesign of Simpson. The design will utilize Annirak's genius. If you look at his first diagram, I will have a block and tackle setup along B and an extension spring roughly along C. I will have the string make 2 round trips along B to give me a mechanical advantage of 4. I am going to make E=8" and F=2". These numbers I hit upon happened to produceby nicholas.seward - Delta Machines
Annirak Wrote: ------------------------------------------------------- > Okay, this is going to be a long post. > > I should make clear before I go any further that > this didn't work out quite how I wanted. I was > looking for a simple geometry that would work > along the same lines, but it didn't come out how I > wanted. It will still work, but the math becomes > moby nicholas.seward - Delta Machines
You could theoretically stretch it during deposition if you have an elastic layer on the top of the build plate. PVDF would need to permanently bond to it (or at least for the build phase). I can see that this isn't going to be easy by any stretch (pun intended). I think the moral of the story will be that a lot of work would be required for a response that is much less than PZT. Side note: uby nicholas.seward - Developers
qharley Wrote: ------------------------------------------------------- > I am not giving up on Bowden just yet. The latest > version of Slic3r has quite a couple of quality > enhancing features built in, and the prints I am > doing today are not displaying the endless oozing > and stringing and blobbing of my first attempts in > February. > > Remember that Kossel is a seby nicholas.seward - Delta Machines
Annirak Wrote: ------------------------------------------------------- > I think I may have an optimization for your > design, which also gets rid of the arccos. I was > thinking about the wheels impinging on the work > area and I have an idea which will replace them > with a pair of pulleys and a cantilevered beam. > I'll try to get it drawn up and posted here later. The wheby nicholas.seward - Delta Machines
I would definitely rather not have a bowden tube. A local company, QU-BD, is making a printer with a 170g complete unit. I am a high school teacher and our school was in the market for a printer. I found theirs and they are only 1 hour down the road. However, the printer sounded too good to be true. I was dubious at first so I scheduled a field trip with my kids and we went and checked outby nicholas.seward - Delta Machines
New video The 100% safe work envelope has been downgraded to 6x6x4.5". You can go much higher and much wider but you have to be careful that the wheels don't hit the part. I might need to write a gcode analyzer that will throw a warning if it would produce a collision. You could easily print an 11" Eiffel Tower or make a 8" diameter by 4" tall bowl.by nicholas.seward - Delta Machines
@Annirak: I agree whole-heatedly with the idea that math should be second to rigidity. However, I like how much easier the math is if all the arms connect on one plane. (I also think people will be more accepting of this but I don't really care about that.) I think that the rigidity differences between a 1 plane connection or a 3 plane connection would be academic. Anyway you cut it, those areby nicholas.seward - Delta Machines
johnrpm Wrote: ------------------------------------------------------- > I proposed this ages ago, but I do not think I got > one response, (I can't be bothered to find post) > my understanding is that the polymer is held at > about 100v whilst it cools to polerise it, my > thoughts were to have electrodes on the nozzle and > build sensors/ actuators directly. I did a google sby nicholas.seward - Developers
Annirak Wrote: ------------------------------------------------------- > Hah. Couldn't stop thinking about it so I did the > math. You may have some trouble with the speed of > the controller, because the coordinate > transformation takes some heavy lifting as far as > the math goes. > > ... I haven't gone through to check your math but the geometry that I picked has theby nicholas.seward - Delta Machines
Gerard Choinka Wrote: ------------------------------------------------------- > have you looked at kossels fishing lines pullys > and the ideas behind? > > the lines are fixed with screws on the wheel and > doesn't depend on friction for driving the wheels Thanks so much! I had a similar idea but they have really refined it. Here is what my basic setup up will be based on thaby nicholas.seward - Delta Machines
> That makes sense. Failure-resistant sensing > methods are definitely better and a magnet + hall > sensor is a good step in that direction. I would > ask, though, what kind of precision you can expect > with that setup. What about a magnetic angular > encoder? > ? > lang=en&mpart=AS5163-HTSP-500&cur=USD I imagine the repeatability will be fantastic. qharley'sby nicholas.seward - Delta Machines
Above you guys can see that I will use cheapo hall effect homing. There will be no need for a full feedback loop. I like the idea of using the encoded desk or potentiometer but I am opting for the simpler design.by nicholas.seward - Delta Machines
Here it is complete with hall effect sensors, extruder, adjustable heated bed, string, bolts, and nuts. All I have left to do is pick a filament drive and make a nice package that includes the power supply, electronics, filament drive, and spool. Here is the home position. The arms are in the 180 degree position. During normal operations there is no need to go past this position so it is aby nicholas.seward - Delta Machines
I just came across this polymer. It exhibits a larger piezoelectric response than quartz does. It melts at 170C and it is cheap. Do a "Google Shopping" search for "Polyvinylidene Fluoride" and you will see many things that are made from it. (There is one downside. You have to "pole" it, apply a very large magnetic field, to get it to exhibit piezoelectric properties.) Why should we care? Wby nicholas.seward - Developers
I forgot to add that to my todo list. I have a few ideas. I think I will use your idea. Instead of a hole, I will make a slot. During homing each arm with travel to the light/dark interface. You can have adjustment built in that will allow you to rotate the slot to different places on the wheel. I am going to use the nozzle touching the center of the bed as the homing position.by nicholas.seward - Delta Machines
Well, I have been getting lots of good feed back and am almost done with my second iteration. I have remodeled all the parts. All bearings are now captive. (They don't rely on being pressfit.) The motors have been upsized to Nema17 and moved up onto the arms. I kept them as low on the arms as possible. This will let me have a longer smaller diameter driving pulley. I also increased the sizby nicholas.seward - Delta Machines
@qharley: I wish I was going to be in Gauteng. I am in the middle of nowhere USA. Great work on your project. Without people pushing the envelope, life would be pretty boring. Check out http://forums.reprap.org/read.php?178,206458 if you get a chance. It has a long way to go before it can be printed but it is Morgan-inspired even if it doesn't look like it. I have been meaning to ask you hby nicholas.seward - Developers