OT/ flats in quadrants

[Bob_Looney]

Trying to get an idea of "normal" here..

I have a 1 1/2 year old cnc that has been run hard since day one, 90% aluminum. It has linear ways and mitsubishi 64 control.

I have been told that flats in the quadrants are normal and not to worry until they get too big.

I don't know what too big is at this point.

I machined a 3.475" boss from a piece of 3 1/2" round stock at 100 ipm with a 1/2" carbide e-mill. I turned on G61.1 in the control. I swept the part with a 10th indicator and the part is within .0002" of round and the flats in the quadrants are all undercut .0002 to .0003.

This doesn't seem too bad to me but like I said, I don't know what normal is for a linear guide cnc. I was hoping some of you have some experience in this area and could let me know what is to be expected of my cnc.

 

thanks

 

Bob

[mayday]

sounds like a + - .0003 machine to me

junk it !

J/K

 

let the chips fly man. thats in better shape than some of our roller coaster machines

[chris m]

I have no idea what G61.1 is, but I've never seen a machine that would interpolate a circle without leaving 'transition' lines at the quadrants. The dimensions you give sound fine to me. I'd check the backlash on the machine for the hell of it, you could put a little backlash comp in it if you find a few tenths lash but other than that I'd say don't sweat it.

 

C

[Rick McAllister]

Sounds to me like you need to have the servos tweaked. I don't know where you are but we have a guy that comes in with a laser and a ball bar and dials in our servos.

 

Just a suggestion.

[Bob_Looney]

I'm in bend, OR. 3 hours from Portland and 5/6 from good laser/ ballbar guys. it's a 1000 dollar bill to get them here for the day and to have them work on it. I talked to a "good" ballbar guy and asked about the parameters in the control to tweak the flats and to get it cutting round if it wasn't. It was like I was talking a foreign language.

I tweaked my "lost motion param" and got it down to the .0003 figure in about 30 min.

[Charles Davis]

Does the machine pause at each quadrant? If so, I'd think the flat is a dwell mark, not slop in the machine. See if it will take a full circle command.

[Phil Orenstein]

Are you climb milling or conventional? Just possible that the flats are being generated from climb milling the circle from square stock. Its possible that its pushing in more where there is material on both side of the cutter. Are you leaving a small finish amount and giving it a final climb finish pass?

 

*Sorry, I reread the question - its from round stock - so that answers the square question.

 

Phil

[Bob_Looney]

does not stop at the quadrants. less than a tenth of backlash error.

[CNCme]

Good Day,

 

You say a boss @ 100 IPM, and only

out O round .0003... and how do you measure

a .0003 flat??

 

IMHO.... It might be, and usually is, a scallop

@ each quadrant... are you looking at this

With the hubble telescope... If your backlash

comp is overcompensating it would cause such

a condition.

 

HTH

 

 

Tony G

Almost Employed Senior Programmer

N.E Massachusetts - Southern New Hampshire

_________________________________________

End mills and tooling are like The "AMMO"

And coolant and chips are like the enemy

Under your boots as you advance in the

Manufacturing Battle

--------------------------------------------------

[Hardmill]

Relatively new machine or not you won't get

away from those conditions without a ballbar

test. And not even with certain machines,

especially @ 100 ipm. Slow down until you

get it resolved. You'll spend more in down

trying to tweak, let the pros handle their

jobs while we stick to ours. In my opinion thats

the common sense way to go.

 

PEACE

[Thad]

Personally, I'd be glad if our machines could cut a circle within .0015 at 10 IPM.

 

I guess the biggest question is does this fall within your typical machining tolerances? If so, I wouldn't worry about it.

 

+1 to Charles and others who mentioned "dwell marks."

 

Thad

[Barghest]

Holding +-.0003 at 100ipm on a linear bearing machine. Sounds like your machine is in pretty good shape. I wouldnt worry about it unless your out of tolerance.

[Quickmike]

That sounds like it is already a pretty accurate machine. What is the typical tolerance that has to be machined on this machine? Just curious what kind of machine is it?

[Rick McAllister]

Bob,

 

Contact Mike Coffaro at Machine Tool Metrology.

He does ball bar and laser leveling. I'm sure he'd be more than happy to help you out. I believe he's now living in Oregon. His # is 818.317.9020. Tell him I sent you. You'll see an instant improvement in finish and accuracy.

 

Good Luck.

[Bob_Looney]

Thanks guys.

I was climb milling around the outside of round bar stock at 100 ipm, pretty fast but thats how I run this cnc.

It's an Arira-Seiki (http://www.akira-seiki.com)

The .0003" error is not the flat size but the "undercut" value in the quadrants as measured by sweeping the boss with 10th indicator.

Typical tolerance for me is... take out the calipers maybe once a week and +/- .005 is just fine. Some of the IndyCar stuff is +.001/- 0.0 and the helicopter part I'm working on now has bearing pockets that are +.0008 / - 0.0

Rick, I'll call your friend and see what he can offer.

Sounds like I'm worrying about nothing.

Thanks again for the info.

 

Rick: wrong number for Mike

 

bob

[Harryman]

Sounds like darned good accuracy to me.

 

Can the Ballbar guys check those parameters at speed? One would think that the values would have to be different at higher G's.

[Rick McAllister]

Harryman,

 

The ball bar test actually measures the machines performance and accuracy at several diameters and feed rates. After these parameters are measured the servos are tuned for a specific dia/feed. Usually he will ask where you normally run a machine so he can dial it in for your specific needs. He will also go in and modify the machines accel/decel macros to increase performance at those feed rates.