Forum

Welcome Guest 

Show/Hide Header

Welcome Guest, posting in this forum requires registration.





Pages: [1]
Author Topic: My Finished Pol-itzler
Morbo
Newbie
Posts: 10
Permalink
Post My Finished Pol-itzler
on: March 17, 2012, 20:57
Quote

It's cut out of 6mm Acyrlic completely and mounted on a (very cheap) and easy to make easel.

The software is actually Alex's from tinkerlog, as it was the easiest to get going with the hardware I have available (A4988 Pololus, breadboard, arduino UNO) and draw any path that's given to it in a .svg (I'm also useless when it comes to programming ;P)

However, The Gondola, which I'd like to dub "Braces' Gondola" is my design and improves on some issues. Mainly secure pen mounting, weight distribution and gondola pen 'twisting'

It's using a T2,5 6mm belt with pullys. Nema17s and some random skateboard bearings. The entire body can be cut from a single peice of ponokos p2 cut template. And is simply super glued together. (threads need to be cut for the blots, of course)

Version two will be coming soon.

Pixahs' here;

http://flic.kr/s/aHsjyBSheh

Bearings -should- be included but because of a number of shipping issues...they've yet to arrive, and I couldn't wait, so the belts are just zipped on the bolts.

There is a significant issue with the software -somewhere- mainly that immediately after the first row of lines (which is -always- straight (in relation to the easel), it...well squishes the rest of the drawing the further it goes down...I'm not sure what the issue could be. Whenever polargraph gets software running for arduino/pololus I may switch :).

But if anyone might have an idea, I'm around to listen! Incase it becomes a curiosity I can share some of the code that's up on github.

My website is at smoothoctopus.com
Thanks for the view! I'd love to hear your comments, good or bad 🙂

Thanks Sandy for the inspiration!

woodpecker
Beginner
Posts: 33
Permalink
Post Re: My Finished Pol-itzler
on: March 17, 2012, 22:01
Quote

Hi Morbo,

very nice machine, proper pictures (even with the "flow")! I like the way you guide the belt. Let us know where to find your code!

Morbo
Newbie
Posts: 10
Permalink
Post Re: My Finished Pol-itzler
on: March 17, 2012, 22:09
Quote

Hey, Thanks woodpecker! The git link is;
https://github.com/tinkerlog/Kritzler
I have modifed one of the .sh files only to tell it where dependencies are. Otherwise, I'm using all original code from there. I 'organized' everything in eclipse, however, I run everything now in terminal. (which works really well!)

sandy
Administrator
Posts: 1317
Permalink
sandy
Post Re: My Finished Pol-itzler
on: March 18, 2012, 11:27
Quote

Hey Morbo, brilliant build and very useful set of pictures - thanks for sharing them! I'm curious about if there's a noticable delay when the bot receives a new command from the PC - on mine, it's quite a pause while it does a back-and-forth, but I'm not being very clever/efficient about it. I get the impression that der Kritzler software is somewhat more streamlined than mine.

Re: the issues you're having with a gradual bend appearing in the drawing - I don't know what to make of that. I get similar kinds of distortions on mine sometimes, but always in the opposite direction (I think) - so I get a distortion that leads to a convex bottom (ha). This is when the home point has been misplaced. I don't recall how Alex does the calibration/homing on der kritzler, a default belt length on startup I guess, so I would look there first to make sure you have your measurements as accurate as is practical. Certainly, when the gondola gets to the bottom of the page, the belts are not as extended as the machine thinks they are.

Does your easel slope backwards or is it vertical? Looking good.

Oh right, code, I've just made a branch of the polargraph_server code that I believe will suit you. It was actually for the ITead dual step motor shield that has A3967 drivers on it, but these are serial chips, so I'm pretty sure they'll work ok with the pololus. I've had a very quick test with some stepsticks here, and it makes whirring noises and things, but haven't tested it any further than that.

The actual pins you use are different though, so if you want to try it

change line 16 from

int const PEN_HEIGHT_SERVO_PIN = 14;

to

int const PEN_HEIGHT_SERVO_PIN = 8;

and lines 62-70 from

const byte motoraStepPin = 2;
const byte motoraDirPin = 3;
const byte motoraMicrostep1 = 4;
const byte motoraMicrostep2 = 5;

const byte motorbStepPin = 6;
const byte motorbDirPin = 7;
const byte motorbMicrostep1 = 8;
const byte motorbMicrostep2 = 9;

to

const byte motoraStepPin = 9;
const byte motoraDirPin = 10;
#const byte motoraMicrostep1 = 4;
#const byte motoraMicrostep2 = 5;

const byte motorbStepPin = 11;
const byte motorbDirPin = 12;
#const byte motorbMicrostep1 = 8;
#const byte motorbMicrostep2 = 9;

The motor*Microstep variables can actually be taken out entirely, because they don't get used in your circuit. (They don't in mine either, but they _can_, which is why the variables were still there - just to remind me not to use those pins.)

This uses the AccelStepper library, so you will need that, but it's easy enough to run through the main arduino environment rather than eclipse.

You might have to reverse your motor coil wiring to get the wheels turning in the right direction, but I don't _think_ there's anything else. (Other than defining your machine size etc through the controller.)

Oh, one other thing though, the microstepping nature of the pololus means that it takes thousands of steps to move from one place to the next rather than just hundreds, and this makes it pretty slow with accelstepper - lanthan certainly noted this as a problem in his build.

Great stuff, thanks
Sandy Noble

Morbo
Newbie
Posts: 10
Permalink
Post Re: My Finished Pol-itzler
on: March 18, 2012, 13:27
Quote

Hey, thanks Sandy. I will definitely try the polar software out, I really appreciate the help.
Hmm, I wouldn't say there is much of a delay at all. The software calculates everything prior to. From what I recognize, it draws generally all the shortest lines first, then goes to longer ones. and literally sends one command at a time to the arduino, and waits until the next command is finished before it sends the next one. So in that aspect, it's relatively quick. But whole dependent on your computer instead of the arduino. hope that answers your question...? 😛

As far as the easel goes, it does tilt back, somewhere between 5-10 degrees. I'd guess.

I realize, as far as the drawing goes, because the easel itself isn't perfectly flat, it'll cause some distortion..but not the kind that it is making. Alex's software depends on a default length. 1060mm from the pully to the pen when the machine is turned on...although I'm pretty sure it's where it should be, I'm likely off a bit, I'll see what I can do to try and get it more accurate. Thanks for the tip.

I wonder then...if it's a possibility that the machine is loosing steps and that is why the belts arn't where the machine thinks they are?

Alexs software seems to make the motors...'jitter' which produces some of his drawing effects? But I'm not sure. I've never had the chance to see anyone elses version in action. often the belts are doing ALOT of shaking as the machine draws...maybe thats another part of the problem? I'm not that sure...

We'll see what happens if I get the polar software going to see if it still does the shaking, as if it's a software thing, controller thing, or just my motors.

Thanks for the help and the code! 🙂

sandy
Administrator
Posts: 1317
Permalink
sandy
Post Re: My Finished Pol-itzler
on: March 18, 2012, 14:38
Quote

Mine is ok for the bitmap stuff, but it's slow where vector graphics have lots of little lines - the actual command transfer is a second or two, so ends up taking longer than drawing the actual lines. If I can figure out how to do CRC checks on the arduino, I'll be able to make this a little more efficient.

Your distortion - I don't think it'll be a problem of dropping steps - otherwise I guess it'd happen everywhere, but it looks like it's definitely related to the distance down the page. Could it be the width of the machine is off? That would that effect too - if the distance between then motors is physically slightly bigger than the software thinks it is.

Part of the reason I was so enamoured of AccelStepper library is that I hoped the slow/steady acceleration and deceleration would take out a lot of the shaking and resonance that will always be a problem with these kinds of machines. It probably has a bit, but I don't really know - it's more a matter of faith 🙂 The problem is that the accelstepper doesn't do well with high speed moves, so microstepping goes sooo sloooww.

Cheers!
sn

Pages: [1]
Mingle Forum by cartpauj
Version: 1.0.34 ; Page loaded in: 0.025 seconds.