Ran the Circle Diamond Square test

Talk about all things CNC
twforeman
Posts: 1347
Joined: Tue Jan 29, 2013 4:51 pm
Location: Minneapolis, MN
Contact:

Ran the Circle Diamond Square test

Post by twforeman » Mon Mar 11, 2013 12:45 am

I made up a Circle Diamond Square test and ran it on my ShapeOko.
c-d-s-2.jpg
c-d-s-2.jpg (123.35 KiB) Viewed 12499 times
It looks pretty nice, but measuring it shows that I appear to have a backlash issue.

These are the dimensions that I used to generate the G code:
sq-dia-cir-2a_640.png
sq-dia-cir-2a_640.png (25.96 KiB) Viewed 12499 times
These are the measured dimensions on the finished part:
sq-dia-cir-2b_640.png
sq-dia-cir-2b_640.png (33.62 KiB) Viewed 12499 times
You can see the full details over on my blog.
Ender 3 3D Printer
ShapeOko v3 serial #0004 - upgrade thread
All of my ShapeOko related blog posts

cvoinescu
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK
Contact:

Re: Ran the Circle Diamond Square test

Post by cvoinescu » Mon Mar 11, 2013 1:14 am

The error in the circle points more to a gantry (X rail) that is not square to the Y rails. How are the angles of the square? If my theory is right, the top-left angle is acute and the top-right angle is obtuse. The diamond should have close to right angles.
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

twforeman
Posts: 1347
Joined: Tue Jan 29, 2013 4:51 pm
Location: Minneapolis, MN
Contact:

Re: Ran the Circle Diamond Square test

Post by twforeman » Mon Mar 11, 2013 1:39 am

I'll have to put a square on the part, I didn't really think of that.

Hmm. Squaring up the X axis should be entertaining... Any suggestions on how to start?
Ender 3 3D Printer
ShapeOko v3 serial #0004 - upgrade thread
All of my ShapeOko related blog posts

twforeman
Posts: 1347
Joined: Tue Jan 29, 2013 4:51 pm
Location: Minneapolis, MN
Contact:

Re: Ran the Circle Diamond Square test

Post by twforeman » Mon Mar 11, 2013 1:41 am

Also, as another question, how does one ensure that the two motors on the Y axis are on the same place on the belts, cog wise. Is it important? I have the feeling it may be...
Ender 3 3D Printer
ShapeOko v3 serial #0004 - upgrade thread
All of my ShapeOko related blog posts

WillAdams
Posts: 8489
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Ran the Circle Diamond Square test

Post by WillAdams » Mon Mar 11, 2013 1:59 am

Interesting. I was working on the same thing, but w/ less success --- Y-axis slipping again.

Let it run for ~2.5 hours since some aspects were pretty good and I wanted to test all the elements in my (hand-coded) file.

Didn't realize the elements were inset w/ nice measurements like that --- will have to re-work my file to take that into account.
Shapeoko 3XL #0006 w/Makita RT0701 Router w/0.125″ and ¼″ Elaire precision collets
Nomad 883 Pro #596 (bamboo)

twforeman
Posts: 1347
Joined: Tue Jan 29, 2013 4:51 pm
Location: Minneapolis, MN
Contact:

Re: Ran the Circle Diamond Square test

Post by twforeman » Mon Mar 11, 2013 2:33 am

2.5 hours? Wow. What are you cutting?

I used MDF and the program ran in about 45 minutes.

I just made up the shapes in LibreCad and used CamBam for the programming. I figured if I made the shapes "real" dimensions it would be easier to measure.

The more I use CamBam, the more I like it. I think I'll end up buying a copy when the trials run out.

If you like I could upload the program and/or the dxf file.
Ender 3 3D Printer
ShapeOko v3 serial #0004 - upgrade thread
All of my ShapeOko related blog posts

WillAdams
Posts: 8489
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Ran the Circle Diamond Square test

Post by WillAdams » Mon Mar 11, 2013 10:16 am

HDPE, but I'm still figuring out:

- feed rate --- does G21 F355 == 355 mm / minute?
- step-over --- currently advancing by 1 mm, so much less than half the cutter width --- will probably re-code to 2mm (should be 1.5875mm, but doing it by hand the even number is easier --- I guess 1.5mm would be easy enough, should likely do that)
- depth of cut --- 1mm, w/ 0.5mm for a finishing cut
- path --- an ever-decreasing spiral seems optimal --- clockwise or counter-clockwise?
- is it best to do an entire level, then go down, or is it okay to do one pocket to full depth, lift out and move on to the next?

I'd love to see your G-code.

Still moot until I work out a better way to tension the belts --- guess I'll try some square aluminum tubing, and I've yet to replace the spindle, and I guess I should check the stepper current.
Shapeoko 3XL #0006 w/Makita RT0701 Router w/0.125″ and ¼″ Elaire precision collets
Nomad 883 Pro #596 (bamboo)

twforeman
Posts: 1347
Joined: Tue Jan 29, 2013 4:51 pm
Location: Minneapolis, MN
Contact:

Re: Ran the Circle Diamond Square test

Post by twforeman » Mon Mar 11, 2013 12:57 pm

When I cut out my router mounts from HDPE I ran at a pretty high feed rate and plunged .030" per pass. This was with a 1/8" diameter 4 flute end mill, not a Dremel bit. Seemed to work out pretty well. I got a lot of melting and build up on the cutter at slower feeds.

As a matter of fact, I was just reading that if your cutter is chattering, you should slow down the RPMs and speed up the feed.

As for step over, 40% of the cutter diameter is pretty standard for flat bottomed end mills.

Finish cuts are generally .010"-.015" (.254mm - .381mm) but for smaller cutters you could go as low as .005" - .001" (.127mm - .025mm). But it depends on the accuracy and repeatability of the machine. When I did my circle diamond square I cut the bottoms of the pockets to full depth on the roughing pass and left .015" on the sidewalls for the finish path.

Sprialing into the part is optimal if your end mill doesn't cut in the center. Plunging straight down is usually okay in softer materials. Best (if possible) is to plunge in air and then do an arc so the cutter comes in tangent to the side wall. CamBam has this option, but I didn't use it for this part. Coming in tangent means fewer cutter marks in the finished wall.

For machines like the ShapeOko (not super stiff and rigid) conventional milling is recommended. This means clockwise. If you climb cut (ccw) the tool gets pushed away from the part and accuracy suffers. Of course with conventional milling the tool can get sucked into the part, so... Here is a pretty good explanation of the two methods.

As far as whether you should do the machining by levels, or pockets, that's your choice. Most machinists get paid by the job, so they optimise their code to minimise the time the cutter is milling air. I try and do the same, and I think that doing each pocket to full depth saves a little time. YMMV.

I'll upload my files when I get home this evening. I'm going to regenerate the G code to tweak it a little bit before I run it again.

I loosened up the rails on my machine last night and tried to square it up some more. I think I might have a way to mount my indicator and a square to do some better checking tonight. Assuming that my square is really square. I have a small machinist square, but it's older and I've dropped it a few times. I should buy a new, larger one. Just what I need, an excuse to buy more tools. :)
Ender 3 3D Printer
ShapeOko v3 serial #0004 - upgrade thread
All of my ShapeOko related blog posts

cvoinescu
Posts: 4442
Joined: Thu Jul 19, 2012 6:50 pm
Location: Camberley, UK
Contact:

Re: Ran the Circle Diamond Square test

Post by cvoinescu » Mon Mar 11, 2013 1:58 pm

The gantry is flexible enough that it can bend in and out of square. The frame is also flexible enough to bend slightly, which you can use to your advantage. Pick one end to square (front or back -- if your rails are precisely equal length, both ends will end up square, but if not, focus on one end only), and use the hold-down brackets to screw it in position. Then move the other end sideways until the Y rails are square to the first end plate, and screw it down. Once that's done, you can just move the gantry by hand until both sides rest firmly against the squared endplate, and the gantry will be square too. Fancy firmware will be able to do this automatically (I think LinuxCNC can do it, if you have homing switches on both sides of the gantry, and separate drivers with separate inputs for the two Y motors).
Proud owner of ShapeOko #709, eShapeOko #0, and of store.amberspyglass.co.uk

WillAdams
Posts: 8489
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Ran the Circle Diamond Square test

Post by WillAdams » Mon Mar 11, 2013 2:52 pm

Trying to simplify this:

- for conventional milling, if enlarging a pocket go clockwise, if reducing the size of an element from the outside, go counterclockwise
- for climb milling, if enlarging a pocket go counter-clockwise, if reducing the size of an element, go clockwise

Is that right?
Shapeoko 3XL #0006 w/Makita RT0701 Router w/0.125″ and ¼″ Elaire precision collets
Nomad 883 Pro #596 (bamboo)

Post Reply