Printed Spools

Printed Spools

Postby Nicholas Seward » Sat Jan 11, 2014 3:50 am

Using my puny 6" dial calipers I was able to map the linearity of half the arm's range.

ARM POSTIONING ERROR.png
ARM POSTIONING ERROR.png (18.24 KiB) Viewed 14600 times


So the position is +/-80 microns on the arm I happened to pick. That is actually not horrible. I think some of my other arms may be worse. I ordered these 12" calipers. I plan to make a simple little fixture so I can bolt it onto an arm to make a full range mapping easier and more consistent.

Before the test I was wondering if the printed gears and/or the printed spool were the main problems. It seems clear to me that the spool is the number 1 problem based on the periodicity of the error. I have ordered these spools to fix this problem. However, I really want to get this tuned with printed spools to help me with the BOB design in the future. When When I get a better full range error map for all the arms I will throw that data at my preprocessor and have it do error correction. My hopes are to get the error under 20 microns. That doesn't seem unattainable. Thoughts?
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Printed Spools

Postby brandonh » Sat Jan 11, 2014 4:10 am

Nice! This stuff is great. I think there's no reason this couldn't be lower, unless the gear fit is inducing more error than expected. I can repeat this test with the spools I'm using now, which were 3d-printed, but then "poor-man's lathed" ("chucked" to a stepper manually turned, with a milling bit removing material). I also have the Graber spools with threads that I can mod to try here. Those have tolerances below 0.001".

If you do design an adapter that mates to a dial caliper, I'd print it and try it.
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: Printed Spools

Postby owens-bill » Mon Jan 13, 2014 2:00 pm

One interesting possibility; apparently you can adapt the cheap Chinese calipers to an Arduino pretty easily. The protocol details are described here (archive.org copy because the original seems to be down):
https://web.archive.org/web/20131103115619/http://www.shumatech.com/support/chinese_scales.htm

And someone's already hacked it into a baby Arduino board:
http://www.instructables.com/id/Reading-Digital-Callipers-with-an-Arduino-USB/

*And* the six-inch versions are on sale at Harbor Freight for just $12 ;) I'll stop on my way home tonight and grab a pair to verify that they're the same ones in the documents. It doesn't seem as though it would be too difficult to bring together a motor shield and the connection to the calipers and make an automated, high-resolution arm mapper.

EDIT: The Harbor Freight calipers are the right kind, at least if you get the ones with a black plastic body; as with many things, they seem to have multiple sources for what they consider to be the same product. They're very cheaply made, but seem to have reasonable repeatability.
owens-bill
 
Posts: 35
Joined: Tue Nov 26, 2013 3:17 am

Re: Printed Spools

Postby brandonh » Mon Jan 20, 2014 4:30 am

Nicholas Seward wrote:Using my puny 6" dial calipers I was able to map the linearity of half the arm's range.

ARM POSTIONING ERROR.png


So the position is +/-80 microns on the arm I happened to pick. That is actually not horrible. I think some of my other arms may be worse. I ordered these 12" calipers. I plan to make a simple little fixture so I can bolt it onto an arm to make a full range mapping easier and more consistent.

Before the test I was wondering if the printed gears and/or the printed spool were the main problems. It seems clear to me that the spool is the number 1 problem based on the periodicity of the error. I have ordered these spools to fix this problem. However, I really want to get this tuned with printed spools to help me with the BOB design in the future. When When I get a better full range error map for all the arms I will throw that data at my preprocessor and have it do error correction. My hopes are to get the error under 20 microns. That doesn't seem unattainable. Thoughts?


I picked up the same 12" calipers for testing my Simpson arms; I don't see any other practical method of measuring linearity. Nicholas, did you design a mount? I'm thinking an 8mm to caliper-end mount, attaching to the rectangular portion of the caliper jaws. If the spool is off, I wanna know before attempting printing.
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: Printed Spools

Postby Nicholas Seward » Mon Jan 20, 2014 4:51 am

Here is my setup.

CALIPER MOUNT.jpg
CALIPER MOUNT.jpg (311.64 KiB) Viewed 14418 times


You need 4 608's and 4 of the stl I attached. (I know you don't have to have bearings on each side but just to future proof it I made them all the same.)
Attachments
caliper holder.stl
(141.88 KiB) Downloaded 625 times
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Printed Spools

Postby brandonh » Mon Jan 20, 2014 7:24 am

Nice! I'll definitely print a few of these. Did you learn anything more about arm linearity after making these?
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: Printed Spools

Postby Nicholas Seward » Mon Jan 20, 2014 7:58 am

I literally just slapped it on for the pic. No testing yet. I am trying to give Wally all my attention for a few days.

BTW, this is designed so the calipers can touch the bolts. If you have long enough M8s, you can make this contact on the smooth part of the bolt.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: Printed Spools

Postby brandonh » Wed Jan 22, 2014 7:40 am

I printed 4 of the calibration pieces and mounted them up. The parts will rock back and forth slightly where they contact the bolt, but as Nicholas mentioned, this is by design.

The gap was a bit thick for my 6" calipers, but they should be easily fixed with some shims. 12" should come later this week.
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: Printed Spools

Postby epicepee » Sun Jan 26, 2014 3:09 am

How much error does this cause in the effector's position?
epicepee
 
Posts: 20
Joined: Tue Dec 17, 2013 4:52 pm

Re: Printed Spools

Postby Nicholas Seward » Sun Jan 26, 2014 3:32 am

@epicepee: With out rigorous analysis, I am coming up with +/-300 microns. That agrees with the prints I have made. In case that number doesn't mean anything to you, it means that you can't make mechanical parts and it is hard to get a good first layer. I would like to get under +/-50 microns.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Next

Return to GUS Simpson

Who is online

Users browsing this forum: No registered users and 1 guest

cron