[ztimbo]

Hey Guys,

I'm trying to machine some small rectangular shaped pieces. They are a bronze alloy and are 1.370"L x .583"W x .500"H with a 9/64 radius. I'm using a Hyundai Kia VX-500 and a .500" 3 Flute Kennametal Endmill. My finish pass removes .030 at the full depth of .500" with 2 spring passes. I'm having trouble with the fitment of the radii of the parts. It's cutting the 1.370 and .583 to size correctly. If this makes sense, across diagonally of the part isn't the same. I'm measuring 1.407 across one side and 1.411 across the other side.

I've also had this issue with cutting an octagon. In gibbs cam everything is symmetrical and the machine is cutting a non symmetrical part that varies anywhere from .003 to .010.

Does anyone know what this could be caused by?

[Delw]

Have you checked backlash?

[ztimbo]

Have you checked backlash?

Nope..Haven't thought about it with this machine only be 2 years old and very few hours. How can I check it?

[RICHARD ZASTROW]

Back in the old days we called that axis reversal error.LOL

Dick Z

[txcncman]

Nope..Haven't thought about it with this machine only be 2 years old and very few hours. How can I check it?

Do you know what backlash is?

[txcncman]

Maybe this will help:

What is backlash? Detail of CNC machine tool backlash in ballscrews, bearings, gibs, and servo backlash.

Or:

Milling: Operations - Backlash

[ztimbo]

Do you know what backlash is?

yes, i know what backlash is. I didn't think it was a problem with CNC's nowadays.

[txcncman]

yes, i know what backlash is. I didn't think it was a problem with CNC's nowadays.

Ok. Yes, backlash is always a problem in mechanical movement. If you know what backlash is, then you should know how to check it. You mount a dial indicator on the spindle so that it touches a solid point on the table. Then you move the axis over to load the indicator and move to zero. Then move the axis back the opposite direction at the smallest resolution and count the detents to find the amount of backlash. Then you can go into the parameters and adjust the backlash for that axis. Then repeat for the other axis.

[ztimbo]

Ok. Yes, backlash is always a problem in mechanical movement. If you know what backlash is, then you should know how to check it. You mount a dial indicator on the spindle so that it touches a solid point on the table. Then you move the axis over to load the indicator and move to zero. Then move the axis back the opposite direction at the smallest resolution and count the detents to find the amount of backlash. Then you can go into the parameters and adjust the backlash for that axis. Then repeat for the other axis.

hmmm, sounds simple enough...I've only been using a cnc for the last 2 years so im kinda new to all this. All ive ever done has been conventional mill and lathe stuff. I was just under the impression that backlash didn't occur with cnc's due to the ball screws.

I also understand that backlash occurs in all mechanical movement if you get right down to it, but I'm talking about a substantial amount here, around .010". I just didn't think cnc's got that kinda of backlash

[Delw]

hmmm, sounds simple enough...I've only been using a cnc for the last 2 years so im kinda new to all this. All ive ever done has been conventional mill and lathe stuff. I was just under the impression that backlash didn't occur with cnc's due to the ball screws.

I also understand that backlash occurs in all mechanical movement if you get right down to it, but I'm talking about a substantial amount here, around .010". I just didn't think cnc's got that kinda of backlash

acording to you first post its only .004 which is .002 perside thats a far cry from .010
how are you checking the part calipers? Mic? toss that calipers in the garbage if you are and use a more accurate way of checking the part.
also look for taper on the flats, burrs etc.

[ztimbo]

acording to you first post its only .004 which is .002 perside thats a far cry from .010
how are you checking the part calipers? Mic? toss that calipers in the garbage if you are and use a more accurate way of checking the part.
also look for taper on the flats, burrs etc.

I will not toss my calipers in the garbage. I can measure within .001 all day everyday with them.

I'm using a mic to check this tho, and please re-read my first post and you will see where I was talking about cutting the octagon.

[Delw]

can you show a picture of the part?
BTW anyone can measure .001 with calipers, all day but its not correct trust me.
I saw the part about an octagon does it measure the same across each flat on the same part?
if not more than likely backlash or drives need to be tuned

[RICHARD ZASTROW]

I'd be willing to bet a few beers that txcncman is right, backlash is probably your problem. Did you read any of those links?

Worst case I saw was a ball screw support became loose and rocked back and forth even though it was doweled in position. Another was a rotary axis "zero backlash" split gear that slipped into "lotsa backlash".LOL

Dick Z

[txcncman]

can you show a picture of the part?
BTW anyone can measure .001 with calipers, all day but its not correct trust me.
I saw the part about an octagon does it measure the same across each flat on the same part?
if not more than likely backlash or drives need to be tuned

I am actually going to disagree with this, with some caveats. A well trained and experienced machinist can measure very accurately and consistently with dial calipers when they use gage blocks or other standards to verify the caliper readings. I have frequently measured parts with calipers while on the machine and then verified within one or two ten thousandths on a CMM off the machine. Just because a micrometer is more precise does not mean it is always more accurate.

This is not something I would put any faith in a novice to do.

[Delw]

I am actually going to disagree with this, with some caveats. A well trained and experienced machinist can measure very accurately and consistently with dial calipers when they use gage blocks or other standards to verify the caliper readings. I have frequently measured parts with calipers while on the machine and then verified within one or two ten thousandths on a CMM off the machine. Just because a micrometer is more precise does not mean it is always more accurate.

This is not something I would put any faith in a novice to do.

so your wasteing time by takeing 3-4 steps to do one step with a micrometer?

why go through the hassle of doubting yourself and double checking with mics later on do it correct the first time with a mic. saves time which is money.

there is a reason for quality control. theres also a reason why all parts should be check with 2 different pcs of equipment shop flor and inspection room. 1st articles etc etc.
the problem is alot 1-5 man shops dont Follow this practice/procedure and thats why there is quality problem.
I have 4 people working here me myself and I and my wife.
I always check all parts then have here double check NOT with the same tools but actual inspection tools that get calibrated.
unless we both end up with a dylexic moment theres never a problem to print.

one of my biggest pet peeves on this forum and outside processing is "well isnt close enough" or people suggesting useing different materials thats NOT called out on the print because it machines easier.
sorry if its not to print its garbage there is NO close enough except in china.

most if not all parts should run exactly at the mean size. granted we have tolorances but theres really no reason in most parts needed to use all of it. with some exceptions (dont get me wrong there are parts that will eat lots of tols but usually in nickle alloys and big rings with lots of contours.)



Delw

[txcncman]

so your wasteing time by takeing 3-4 steps to do one step with a micrometer?

Delw

Ok. I will sit here and watch you use your 48" micrometer to measure the center to center distance between 2 holes and hold + / - 0.001 on the location on a compressor mounting frame while is is still on the boring mill.

Then right after that I will watch you use your 72" micrometer to measure the male register of a lift pump housing and hold +0.000 / -0.003 on size and circularity while it sits on that boring mill over there.

Don't get too bent out of shape over how other people run their quality systems. If yours is so much better, soon you will be the only shop left in business.

LOL

[Delw]

WOW you have incredible skills if your using calipers for such a thing you got lucky or let close ones go by.

I would generaly use a muller bar and a set of starett gage blocks.

your completely missing the point, and NO I would never use calipers to measure a +-.001 tolorance especially on something that long.

[ztimbo]

I'm having a service tech come out and check everything on the machine. So we shall see what he finds wrong, if anything.