Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

WillAdams
Posts: 8625
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by WillAdams » Mon Dec 19, 2016 11:50 am

Yeah, 3D printing, while complementary to traditional CNC is very much an indoor activity.
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets

jacquard
Posts: 13
Joined: Fri Dec 09, 2016 10:48 am

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by jacquard » Thu Dec 29, 2016 10:21 am

OK, over the last week or two I've finally finished rebuilding the machine at 850mm and doing a lot of checks/calibrations/re-checks along the way, and just come back from the workshop after a couple of small test cuts.

I'll write up my build notes shortly as I found a few things that would probably be worth adding to the wiki, but the long story short is that in the rebuild and stiffening, I've removed pretty much every source of error, EXCEPT for the 0.5mm that was giving me the trouble in the first place! With the few test cuts I did I got very nicely repeatable numbers (+/- 0.03-0.05mm) and things were a lot squarer than before. Just 0.5mm short! :P

This was with the feed rates I was using before. I went down to the Easel recommended feed rates and the error went down to about 0.3mm. So the prime suspect for me is something spindle/cutter related, be it runout, spindle and/or bit flex, general QA issue with the Chinese quiet cut spindles, etc. Unfortunately Santa didn't leave any Bunnings vouchers under the Christmas tree so moving to a Makita spindle isn't on the cards just yet.

When I finish building the integrated vice, I'll do another circle-diamond-square and post it up. I'll also give some even slower feed rates a go and see if the error scales down further.

ijv
Posts: 71
Joined: Sun Apr 06, 2014 10:41 am

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by ijv » Sun Jan 01, 2017 9:38 am

Have you tried running the Shapeoko Calibration pattern using your machine as a pen plotter?

http://www.shapeoko.com/forum/viewtopic.php?f=7&t=742

That would at least isolate spindle and milling force issues from the underlying setup of the machine

WillAdams
Posts: 8625
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by WillAdams » Sun Jan 01, 2017 2:15 pm

I really suspect that it's spindle runout / endmill (deflection as hpgreyling noted). Hopefully Santa will provide a new trim router.
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets

HPGreyling
Posts: 123
Joined: Sat Nov 15, 2014 11:37 am
Location: South Africa

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by HPGreyling » Sun Jan 01, 2017 7:50 pm

Have you tried lowering your feed speed? It seems by looking at your calibration cuts of the circle and square that the bit is getting pulled forced all over the show. Especially since you are running at max speed and the cuttings are not getting evacuated fast enough and thus deflecting the bit.

I wouldn't say it is runout rather deflection .

jacquard
Posts: 13
Joined: Fri Dec 09, 2016 10:48 am

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by jacquard » Tue Mar 07, 2017 11:02 am

OK, after a long time I finally sourced a set of metal Makita mount plates, and borrowed a Makita trim router from a good friend who happened to have one (for ages... if only I knew this a long time ago!)

Stiffened the X-axis with two 3mm aluminium plates and gave the Makita a go. Now it's giving perfectly consistent inaccurate cuts (LOL!) HOWEVER - going through the Z axis spindle plate parts, I've noticed that two of the V-wheels, top-left and bottom-right, have a huge amount of play in them (despite the eccentric nuts being correctly set). So I suspect these V-wheels are cracked or otherwise broken, and this is a big source of flex in the Z assembly.

I have a heap of Delrin V-wheels left from my abandoned Ordbot Hadron project (since bought a Wanhao i3 rebadge and built an enclosure out of an old metal AV cabinet) so my project for tomorrow is to rebuild the Z spindle carriage with new V-wheels and see what happens.

Note that I haven't recalibrated the machine yet, or even done a wire test to make sure the router is sitting perfectly straight in the holder. I just used a square to make sure the mount was square to the floor and bolted the trim router up (I have been waiting sooo long to get this thing going, I wanted to get to it right away!)


Some of the notes regarding the build revisions I did (probably more to come as they come to mind):

- cutting the Makerslide. It's perfectly doable with any old mitre saw, even a cheap one, the key is to use a new (or newish) blade with lots of teeth (40+ is good). My $80 cheapie mitre saw did a better job than a friend's $500 one, simply because the blade was newer despite being an inferior brand.

- Y axis braces. I used some 110x50x3mm steel plates from the hardware store that are normally used for building. I marked out 4 holes on each end (one to go with a 3mm aluminium L-bracket rail on the machine base, the other to coincide with the paths of the T-nuts in the Makerslide) and drilled them out,

- automotive spark plug gap feelers make great shim sets, especially the 30 packs that have a shim for every 0.03mm increment. Find the correct one in the set and cut it to length once you've confirmed the right gap needed.

Final note - the Makita is INSANE compared to the 400W DC spindle I was using before. It moves through the wood as if it isn't even there.

Some pictures -

https://1drv.ms/i/s!Ary7vfrO4iaBmjROAoGsYuXJ4CCp
https://1drv.ms/i/s!Ary7vfrO4iaBmjbgSIU9HdQuf_3F
https://1drv.ms/i/s!Ary7vfrO4iaBmjWxQu9an6lsxZig

jacquard
Posts: 13
Joined: Fri Dec 09, 2016 10:48 am

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by jacquard » Wed Mar 08, 2017 9:29 am

Inspected and replaced all the V-wheels on the Z spindle carriage plate, and modded it to run 6 V-wheels because I have the spares and why not. The "Antonov mega-freighter landing gear" look of the spindle carriage plate at least looks the business! Now I'm getting absolute error of -0.25mm and variance of +/- 0.05mm, with feed rate of 800mm and depth of 1.5mm. Reading what others can realistically get out of their Shapeoko class machines I think I'm doing quite well indeed with the variance now, and the only thing gnawing at me is that absolute error. The OCD part of me would really rather the error be 0 +/- 0.05mm, rather than -0.25 +/- 0.5mm :lol:

Tightening the V-wheels (thinking it might be Delrin wheel flex) hasn't appreciably improved the situation so I'll be relaxing them back a bit in the interest of longevity.

My theories now to try and minimise what I can of the absolute error are the following. Any comments welcomed!

1. reduce the acceleration rate. If it is bit deflection causing the absolute error, reducing the amount of acceleration should (in my mind at least) reduce this "dynamic force" (Physics classes were way too long ago for me to remember the correct term for this) as the bit is accelerating while cutting fresh wood at the same time. (Not sure what the current settings are). What time I might lose with lower acceleration I'm sure I'll more than make up for with the power of the Makita router - cutting deeper and (ultimately) faster than what's possible with the DC spindle.

2. tighten the X/Y belts a bit more. I am measuring belt tension with some fish scales, and right now they are at about 1.6kg (3.5lb) of force needed to deflect 25mm (1in). Perhaps there's a teeny bit of backlash in the belts still there, which tensioning further might resolve.

3. exactly centre the spindle in all dimensions with the bendy wire. A bit too lazy/hesitant to do this as I know it's "close to" perfect, I'm using a borrowed router, and "my" router won't arrive till the end of the week at least :-) I really should, though...

I have promised my wife that after this set of calibrations I might actually make something with the machine, rather than endlessly work on the machine to its own end. :geek:

PS. it hasn't escaped my attention that an absolute error of 0.25mm is uncannily similar to the distance travelled with one step of the stepper motor. I'm not sure what to theorise from this possible source of error though.

WillAdams
Posts: 8625
Joined: Mon Apr 09, 2012 6:11 pm
Location: Pennsylvania --- south of the Turnpike, East of US-15
Contact:

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by WillAdams » Wed Mar 08, 2017 12:50 pm

If it's always falling short, then I'd just calibrate for belt stretch and see if that works.

Otherwise, try the following techniques:

- leave a roughing clearance and take a finishing pass (ideally w/ a sharp new HSS bit)
- see if climb vs. conventional cutting changes anything --- possibly in conjunction w/ the above
Shapeoko 3XL #0006 w/ Carbide Compact Router w/0.125″ and ¼″ Carbide 3D precision collets

jacquard
Posts: 13
Joined: Fri Dec 09, 2016 10:48 am

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by jacquard » Wed Mar 08, 2017 11:26 pm

Oh that's a nifty idea, tweaking the steps/mm to optimise for the middle of the travel. So it's -0.25mm absolute at 0, bang on at 325, and +0.25 at 650. It may not be actually accurate, but at least the absolute error will be always less than 0.25mm.

Do you happen to work at Volkswagen's diesel engine department? That's a very clever way to mask the error in the machine :D

I think the ultimate solution is what you suggest regarding separate roughing and finishing cuts. But nothing beats fundamental accuracy (even if enhanced with an "acoustic function" :) )

When ordering my router I also got a couple of decent quality bits, one of which was an upcut bit. This would effectively let me attempt "climb" cutting using a conventional toolpath yes? Or have I got things conflated, and are conventional/climb toolpaths and up/downcut bits fundamentally different somehow?

HPGreyling
Posts: 123
Joined: Sat Nov 15, 2014 11:37 am
Location: South Africa

Re: Shapeoko 2 - cuts ~0.5mm short, tried almost everything!

Post by HPGreyling » Thu Mar 09, 2017 6:30 am

I found that with upcut bits the bit is trying to pull the material towards the router thereby flexing the whole assembly to a degree where the deflection is noticeable when feed rate en plunge is not correctly set.

just my experience

Post Reply