Wally-0

Re: Wally-0

Postby Cozmicray » Tue Jan 21, 2014 7:15 pm

Lasercut!!!
Isn't that against the ReRap credo?

So all your designs are just one-offs.
Design and Development just for Design and Development fun?

Your not going to get bought out by Epson, HP, or Microsoft that way.

I guess you are in academia - and producing minds???

:D
Cozmicray
 
Posts: 126
Joined: Tue Dec 03, 2013 10:23 pm

Re: Wally-0

Postby Nicholas Seward » Tue Jan 21, 2014 10:04 pm

Cozmicray wrote:Lasercut!!!
Isn't that against the ReRap credo?


I am a designer of RepRaps but that doesn't make me only a designer of RepRaps.

Cozmicray wrote:So all your designs are just one-offs.
Design and Development just for Design and Development fun?


Well yeah. I design for the fun of it. If it wasn't fun then I would stop. I am not sure what you mean by one-offs. I am providing virtual designs that will evolve and that people can and are using across the globe. RepRap printers are oriented more to self-builders. I will design other printers in the future that aren't RepRap printers (still open source) but are easy to mass produce. Wally and GUS would need quite a few parts injection molded or converted to laser cut parts to make a production run viable. It is a little early in the process for either one to get committed to that level. In that way, these "one-offs" are invaluable in informing my future production plans.

Cozmicray wrote:Your not going to get bought out by Epson, HP, or Microsoft that way.


I am philosophically opposed to their standard business models and have no desire to get bought out by them. I have what I consider a dream job and I will have to truly believe in what I am doing to give it up for a foray into the business arena.

Cozmicray wrote:I guess you are in academia - and producing minds???

:D

I am currently in academia.

Let's be clear here. I have little interest in mass producing and a great interest in enabling collaborative open source hardware design. I am working to that end. I won't bore you with my whole business model and some of what I am working on in currently secret (later to be open source) but it is far from your typical 3D printer upstart. I want to help enable divergent design. Much of what is out there encourages convergent design. Here is where you can call me a hippie.

I am sure Epson and HP will come out with some printers and I even expect them to have some decent advancements in the technology. If you want to be able to buy a printer that works out of the box and needs no tweaking then that is what I would wait on. Just expect to pay a small fortune in consumables. The will probably need 2.3mm filament with nano unique RFID chips every 1m or so in the plastic. (Well that may be extreme but I don't expect it to be easy to use 3rd party filament.) There is nothing wrong with someone going after this business model. I just want to stay far far away from it.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Wally-0

Postby Nicholas Seward » Thu Jan 23, 2014 5:16 pm

More good progress!

I finished my first draft on Wally's new segmentize. It seems to address all the problem areas. It makes calibration easier. It scales speed correctly. (Thanks Charles Steinkuehler for the speed fixes!) It does bed leveling. It does bed arc adjustment. The one problem is it slower than molasses. I can fix that. (The problem is that I have the inverse kinematics but not the forward kinematics. This is needed to some calibration. My bandaid solution is to get from the machine coordinates to the reference coordinates using numerical techniques.)

WALLY SQUIRREL.jpg
WALLY SQUIRREL.jpg (175.85 KiB) Viewed 25315 times


There is one band around the middle. I am not sure what the cause is. It could just be a bad value in the calibration table. If other prints have the same band then that will be confirmation for that theory. Other than that, the surface quality is top notch and the retract settings left little to clean up. (At this point it is leaving less to clean up then my Printrbot. I think I have a slight blockage.)

WALLY LISA SQUIRREL.jpg
WALLY LISA SQUIRREL.jpg (193.51 KiB) Viewed 25315 times

Wally squirrel on the left and LISA squirrel on the right. Pretty dang close if I do say so myself.

WALLY CUBE.jpg
WALLY CUBE.jpg (155.88 KiB) Viewed 25315 times

20.6mm x 19.6mm x 19.8mm
The height is within the layer height so I am good there. The test cube printed about 7.5mm from the location I expected so I have a lead on getting this fixed.

Now, I plan on getting a fresh computer and doing video documentation of the following.

*Loading firmware
*Setting up the host
*Calibration
*STL to print workflow

Exciting times. GUS and Wally and getting pretty close to being able to print their own parts.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Wally-0

Postby cdsteinkuehler » Thu Jan 23, 2014 5:27 pm

Nicholas Seward wrote:More good progress!

I finished my first draft on Wally's new segmentize. It seems to address all the problem areas. It makes calibration easier. It scales speed correctly. (Thanks Charles Steinkuehler for the speed fixes!) It does bed leveling. It does bed arc adjustment. The one problem is it slower than molasses. I can fix that. (The problem is that I have the inverse kinematics but not the forward kinematics. This is needed to some calibration. My bandaid solution is to get from the machine coordinates to the reference coordinates using numerical techniques.)


Great news, and great looking prints!

I was reviewing your changes to the segmentize code this morning, and it does look like there's room for some optimization. ;-)

I need the forward kinematics as well, so post them if you get it worked out. I'll let you know if I make any progress.
cdsteinkuehler
 
Posts: 74
Joined: Tue Nov 26, 2013 1:53 am

Re: Wally-0

Postby Nicholas Seward » Thu Jan 23, 2014 6:16 pm

@cdsteinkuehler: A closed form solution is going to be ugly. Option 2 is to make a 2D lookup table. I will try for the closed form but my math mojo wasn't working last night.

So if you detach the arms in the middle but leave the motors set you get each arm sweeping out a cardioid curve. The problem is that the mechanical advantage isn't a whole number so the curve never ends. Worse case scenario is that every point within 300mm from the shoulder is reachable. We have to find the intersection of these two curves. So a direct attack is out. The next approach would be to find a close approximate curve for the useful range and intersect those.

WALLY ARM SWIRL.png
WALLY ARM SWIRL.png (25.21 KiB) Viewed 25313 times


The task is to take this curve and a mirror of it and center them on the shoulders. Use the stepper rotation to set the how the are rotated. Lot's of intersections. Ideas?
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Wally-0

Postby cdsteinkuehler » Fri Jan 24, 2014 1:38 am

Nicholas Seward wrote:So if you detach the arms in the middle but leave the motors set you get each arm sweeping out a cardioid curve. The problem is that the mechanical advantage isn't a whole number so the curve never ends.


I'm confused. The only way you could get something that looked _remotely_ like that is if you swing the arms around multiple rotations (basically turning the arm instead of turning the motor). And my wally won't do that...there's this back-board thing in the way! :)

The way I figure it, you get two modified arcs and have to figure out where they intersect. Given a motor position, figure the angle (c or d in your kins png) assuming a 90 degree angle between the base (L) and the first arm l. This gets you the distance from the shoulder for your arc (a or b), but that distance has to be modified slightly as the base arm moves around the shoulder pivot. But there's no way to get a full cardioid spiral that I can see! AFAIK, what you'd get from a physical Wally would be at *MOST* the outermost loop in quadrants I and II of the cardioid image (limited by how far the other arm will reach). What am I missing?!?

Also, I reviewed the new Wally code and I can't find where you're rotating the reference 3D space to account for the tilt of the bed. It looks to me like you're just tweaking the Z dimension to account for the bed not being level. It seems if the bed is so far out of level you have to compensate for it, you ought to be pivoting the reference coordinate space so you don't end up with non-square parts. So what am I missing?
cdsteinkuehler
 
Posts: 74
Joined: Tue Nov 26, 2013 1:53 am

Re: Wally-0

Postby Nicholas Seward » Fri Jan 24, 2014 4:29 am

cdsteinkuehler wrote:I'm confused. The only way you could get something that looked _remotely_ like that is if you swing the arms around multiple rotations (basically turning the arm instead of turning the motor). And my wally won't do that...there's this back-board thing in the way! :)

The way I figure it, you get two modified arcs and have to figure out where they intersect. Given a motor position, figure the angle (c or d in your kins png) assuming a 90 degree angle between the base (L) and the first arm l. This gets you the distance from the shoulder for your arc (a or b), but that distance has to be modified slightly as the base arm moves around the shoulder pivot. But there's no way to get a full cardioid spiral that I can see! AFAIK, what you'd get from a physical Wally would be at *MOST* the outermost loop in quadrants I and II of the cardioid image (limited by how far the other arm will reach). What am I missing?!?


Of course in reality you can not spin the arm forever but math doesn't care about silly things like the backboard. We will need most of the curve in the graph. The curve actually goes on to infinity and fills up the whole circle but I stop the curve generation when the arm completely folded in on itself. At any given time only about 180 degrees of this is usable but it could be any portion of it. I am going to do some circle fitting of this curve over the work envelope and see how close we get.

cdsteinkuehler wrote:Also, I reviewed the new Wally code and I can't find where you're rotating the reference 3D space to account for the tilt of the bed. It looks to me like you're just tweaking the Z dimension to account for the bed not being level. It seems if the bed is so far out of level you have to compensate for it, you ought to be pivoting the reference coordinate space so you don't end up with non-square parts. So what am I missing?


Good catch on my cheat. I need to put something in for that. I found that the arms tilted down much more than the bed. My bed just happens to be pretty dang level. So you can assume the arms are level and do the adjustment like I did before or you can assume the bed is level and possible scale up the y dimension by dividing by cos(tilt angle). However, even if the arm is tilted 5 degrees we are only off by .4%.

Again, this is a cheat. Solutions:
*shim the bed to make it level
*add springs so you can adjust the level
*measure height of each corner of the bed and get the software to adjust for it
(Starting at the upper left of my board at the square_z position and going around clockwise I got these values for the height of the board... 207.2 206.9 203.3 204.3 That comes up to be a 1 degree tilt. That is nothing to ignore but it is far from my biggest problem right now.)

My personal plan is to add the adjustment back. Again, good spot. Thanks for keeping me honest.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Wally-0

Postby cdsteinkuehler » Fri Jan 24, 2014 4:43 am

Nicholas Seward wrote:Of course in reality you can not spin the arm forever but math doesn't care about silly things like the backboard. We will need most of the curve in the graph. The curve actually goes on to infinity and fills up the whole circle but I stop the curve generation when the arm completely folded in on itself. At any given time only about 180 degrees of this is usable but it could be any portion of it. I am going to do some circle fitting of this curve over the work envelope and see how close we get.


I'll grant you that any position in the circle is reachable, but for forward kinematics, you're given the current motor position. Combine that with the fact that the shoulder arm can only move (at most) +/- 90 degrees from perpendicular to the backboard (making a small corresponding change to the effective motor position in the process), and I don't see how you can generate the full spiral. <confused>
cdsteinkuehler
 
Posts: 74
Joined: Tue Nov 26, 2013 1:53 am

Re: Wally-0

Postby Nicholas Seward » Fri Jan 24, 2014 4:52 am

@cdsteinkuehler: You are right. You will only use sections of the spiral at a time. There will only be 1 intersection between the 180 degree sections. The problem is that without going into numeric techniques (This actually might be the best plan. I can probably tailor a root finder that can converge in 10 iterations or so.) it is hard to get a closed form solution for the intersection of these two section without the rest of the spiral mucking it up. So one plan is to get a circle that fits the 180 section as a function of the motor rotation and then do the intersection. There will now be two intersections but we just want the one on our side of the backboard.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Wally-0

Postby Nicholas Seward » Sat Jan 25, 2014 5:10 am

WALLY FIRST LAYER.jpg
WALLY FIRST LAYER.jpg (182.42 KiB) Viewed 25288 times


A first layer to be proud of. Just fixed a sign in segmentize that was causing me leveling grief.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

PreviousNext

Return to Wally

Who is online

Users browsing this forum: No registered users and 2 guests

cron