Machines are not repeating

stevo1 · Jun 25, 2009

Hey guys I need some help. I have a VMC and a 5 axis HMC both with Fanuc 15mf controls. They both have a rotary axis. I have the VMC machine 0 point at the center of the table. On the HMC I have the head rotated -90deg to represent a VMC and have the machine 0 at the center of the table. The HMC is a 93’ and the VMC is a 00’

I set 0 up by using a precision setup bar in the spindle and spin the table with an indicator mounted on it. I get this less than .0004” around, vertical is .0002”, and spindle run-out at the end of the bar (12”bar) is about .0008-.001”. Run-out could be better but this is not my concern at this point.

Problem is the 0 point is changing from day to day and sometimes not at all. It will vary +/- .006 radially. We do mainly bolt circle drilling. Tell the machine to go to the BC radius and then just keep rotating the table. As you can see this is a major problem with trying to hit my BC diameters. If it runs -.006 then this makes my BC small by .012” and my tolerance is .01”.

I have tested for backlash and found no issue. This is about the extent of my knowledge to check when it comes to machines not repeating. We had the machines laser checked when we first set them up about a year ago and the results were surprisingly accurate. The HMC is the one that needs the most attention. When I originally set this up I would go to a given X,Y, Z, A, C and zero the indicator on the bar. I would then do moves all over the machine, tool change, power cycle. Send the machine back to the coordinates and indicator would read 0,0. The machine can still do this. Is the Z,Y of the machines drifting??

Any help on what to look for is much appreciated. Please be descriptive.

Thanks guys,
Stevo

psychomill · Jun 25, 2009

If it's not something in the program and you have a solid part/fixture set up... Thermal growth? Loose encoder? Is the head rotation on the HMC repeating 90°?

What are the machine make/models?

stevo1 · Jun 25, 2009

Thermal growth of .012" total variation? It is hot here but I did not think that it could vary that much.

The head rotation on the HMC is repeating right on the money.

The VMC is a Monarch RTB. The HMC is a Forest Line' Flexiax 516.

Thanks for the input, I will check the encoder.
Stevo

psychomill · Jun 25, 2009

I know... it's far fetched but I'm just throwing it out there...

The part repeats on the set up right? Drilling into a flat surface or angled? Cutters are not running out? (more dumb questions..)

stevo1 · Jun 25, 2009

Parts repeat great on the setup. No problems drilling flat, side, or angled holes. Once the machine is setup the machine is very accurate. It repeats within .0005” all the way around the part on hole location. No cutter run-out. Hole size stays put even through 64 holes of inconel 718.

It’s just some days that you throw the test bar in to check it you have to adjust it about .002” radially. When I say it has a variation of +/- .006 I don’t mean that I have gone one day of -.006 to the next day of +.006. I have been tracking it for some time now(few months) and writing down what each change is and that is the range of change.

They are not dumb questions. Keep em comin something is bound to be right. It is the law of averages

Thanks again Psyco,
Stevo

D · Jun 25, 2009

How big of b'circle are we talking? Are the tools used for the holes walking? How much difference in length is the indicator used to find center versus the tool length? Thermal growth does seem far fetched. However a loose encoder coupling is not uncommon. Also the reason I asked about the difference of set-up tools to the cutting tools is that an out of square axis could be the cause. With machines like we are used to with x travels of 198"plus we are accustomed to location problems (even though they are all new)may be far fetched but hey. .0012 is a whole bunch of thousands :crazy:

spock · Jun 25, 2009

are you checking it at startup and finding the variation from the day before? Or after homing the machine at any time?

psychomill · Jun 25, 2009

Wisconsin - How close are you to a large body of water?
Machines on foundations?
Get vibrations from other machines around these?
Ever leave an indicator overnight or through the day? I know you're seeing change but you might be able to narrow down the axis or axes that is shifting.
Do you power down or leave the power on?

You had a problem with the Forest Line before didn't you?

stevo1 · Jun 25, 2009

D said:
How big of b'circle are we talking? Are the tools used for the holes walking? How much difference in length is the indicator used to find center versus the tool length? Thermal growth does seem far fetched. However a loose encoder coupling is not uncommon. Also the reason I asked about the difference of set-up tools to the cutting tools is that an out of square axis could be the cause. With machines like we are used to with x travels of 198"plus we are accustomed to location problems (even though they are all new)may be far fetched but hey. .0012 is a whole bunch of thousands

On average they are about 55-60” BC diameter. The holes do not appear to be walking. We check roundness and perpendicularity, it all checks out ok.

Our average tool length is about 5” from gauge line. Our setup bar is 12” long, however I only move down the bar about 5” to be in the same range as the tool length. I do also run the bar up and down and it holds true within .0002”. If I center the bar at 12” from GL I get the same numbers if I was 1” from GL.

Move it over one more decimal .012". Way too far out for me :crazy:

spock said:
are you checking it at startup and finding the variation from the day before? Or after homing the machine at any time?

I am checking it at no particular time at all. I check it sometimes in the morning, afternoon, or night. I use to check it when I thought of it or when the guys were getting close to max or min tolerance. Now that I have been checking it on a daily basis it seems 1 day to the next it won’t stay the same. I have had it shift or change by about .0015 in a day.

psychomill said:
Wisconsin - How close are you to a large body of water?
Machines on foundations?
Get vibrations from other machines around these?
Ever leave an indicator overnight or through the day? I know you're seeing change but you might be able to narrow down the axis or axes that is shifting.
Do you power down or leave the power on?

You had a problem with the Forest Line before didn't you?

I am about 30miles from Lake Michigan. The machines are on a new foundation built when they arrived. I have seen the press move the indicator by .0015 but moved back to 0,0 when I happen to be testing the machine. However the press has not been run in about 2 weeks and I still have variation on a daily basis. The other machines around are 4 VTL’s no hard milling.

I will try leaving an indicator on it overnight. I know it is the Z-axis moving on the HMC because the only things that are changing is the BC. I put the indicator at 12 o’clock and center it on the test bar and zero it out. I then program a rotation of the table 180deg to 6 o’clock and that is when I see the variation that needs to be adjusted. 3 o’clock and 9 o’clock are 0,0.

Good memory, you are correct I had problems in the past with this machine in the A-axis (head rotation). There was slop in the head even when it was clamped. We had fixed this problem with meshing the gears properly. Now we get less than .0005” movement when it is in a fixed position.

Sometimes we do full power downs and sometimes it’s on all night. It usually stays on all night if 2nd shift runs it. If not day shit usually powers it down.

Thanks,
Stevo

FP4NC_Finster · Jun 25, 2009

If you're not using "Positive Approach", keeping the lash all on the same side of the mechanicals, the Ballscrew Thrust Bearing packs would also be a good place to look as well. As Pysco says, look for easy stuff first, lose encoder or the coupler between the servo and the screw, servo motor mount tight? etc.. You might try doing a Circular Interpolation using the straight axises for motion and then check the circle produced, paying particular attention to the quadrants where one screw must reverse direction.

D · Jun 25, 2009

Stevo1, Sorry about the extra zero, maybe just wish-full thinking hey.
Guess I better not try any programming today.

That's a good idea psychomill brought up to leave it powered up with an indicator on it.

Ox · Jun 25, 2009

On D's that big - a slight change in angle on the 5th axis could easilly effect the B/C that much. Have you ran an indicator along the face of your 4th to check that the 5th is still sqr when this happens?

I know it is the Z-axis moving on the HMC because the only things that are changing is the BC.

Y axis maybe? :confused:

OK - you state that you co-ax or probe or whatever_some sorta locating pin. Power cycle - and go back and it's still the same place. Have you checked this point many hr later? (I will ass u me so)

But I think my biggest point may be:

How far from the anchored ends of your screw(s) is this location point?

My point is that the closer it is to the anchored end of the screw - the less that thermal growth has an effect. The X screws on my lathes will grow .001-.0015 in the first few hrs and they only have - what - 8" stroke? Take that out to 40" or whatnot travel and you git your .006". ???

Bolt pattern is bigger than X/Y travels of the machine I take it? Or maybe this is just easier to program and keeps everything on C/L better - even if it is bigger D - it is at least on C/L... ???

BTW - Thermal growth won't have much at all to doo with temps - but rather temp rise (like a motor rating) after homing.

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

Think Snow Eh!
Ox

psychomill · Jun 25, 2009

BTW - Thermal growth won't have much at all to doo with temps - but rather temp rise (like a motor rating) after homing.

That's what I meant... Stevo, did you take it that way? Not air temp.... machine temps... (motors, ballscrew, casting, coolant, oil...). And the size of the machine will have relative effects to it. Thermal growth is linear

HuFlungDung · Jun 25, 2009

Ambient air temperature differences in the shop could have a significant effect over 60 inches. 10 degrees F rise over 60 inches is .0036", and on a rotary setup, you'd be doubling the error. You'd almost need laboratory controlled conditions to verify the diameter at particular temps.

psychomill · Jun 25, 2009

How about thermal growth of the part?

CMLbend · Jun 25, 2009

Right off the top of my head I agree with psychomill's first post about the encoder. What voltage/amperage do you run off. Is it on an isolated circuit , sometimes welders can create a spike that will throw off an encoder even if it is a non-malfunctioning part.

stevo1 · Jun 25, 2009

D said:
Stevo1, Sorry about the extra zero, maybe just wish-full thinking hey.QUOTE]

No problem. You are right “wishful thinking”. I would be thrilled if it was running that good.

Ox said:

On D's that big - a slight change in angle on the 5th axis could easilly effect the B/C that much. Have you ran an indicator along the face of your 4th to check that the 5th is still sqr when this happens?

OK - you state that you co-ax or prob or whatever_some sorta locating pin. Power cycle - and go back and it's still the same place. Have you checked this point many hr later? (I will ass u me so)

How far from the anchored ends of your screw(s) is this location point?

Bolt pattern is bigger than X/Y travels of the machine I take it? Or maybe this is just easier to program and keeps everything on C/L better - even if it is bigger D - it is at least on C/L... ???
Ox

Click to expand...

Yes I have checked to make sure that the 5th axis is square with the 4th axis when this problem happens.

All I did was take the test bar and call a position of X,Y,Z,A,C then clamped(magnet) the indicator to the machine table and set it to the test bar. I then ran the machine everywhere and powercycled the machine leaving the indicator in the same spot. When I called the same X,Y,Z,A,C positions the indicator read 0,0 like I set it. I was just trying to see if the machine was repeating it’s position.

The X-axis has 78” of travel. The position in which I am centering and checking this is EXACTLY in the center of the screws of the axis.

Mainly all we do is BC drilling. We keep the tool on CL of the table and rotate the degrees per hole. The tool can’t reach over the center of the table without OT.

I take the test bar and run it to X0,Z0 center of the table. I mount the indicator to the table and move Z to find the exact center of the test bar. I then zero the indicator and spin it 180deg, make adjustments and set this as my X0 machine home. I do the same thing for my Z axis. So at any given time I can program X0,Z0 put an indicator on the bar and spin it 360deg and it never moves.

Now when doing this the X axis is always 0 (never moves) but the Y axis is drifting/always changing.

psychomill said:

That's what I meant... Stevo, did you take it that way? Not air temp.... machine temps... (motors, ballscrew, casting, coolant, oil...). And the size of the machine will have relative effects to it. Thermal growth is linear

Click to expand...

I understood a bit of what you were saying...Ox just described it more.

psychomill said:

How about thermal growth of the part?

Click to expand...

I don’t think it is expansion of the part. These parts are always running true on the machine when indicating them in. Once the guys measure there BC diameters and find them to be off I then proceed to check Z center and it is off the exact amount the BC is off.

Ok one thing I am starting to realize after explaining this more is I want to clarify the type of machining center. I said HMC…the spindle is horizontal. The table is horizontal but the spindle will rotate from 0deg(horizontal) to -90deg(vertical). This is acting exactly as a VMC except the Z axis is Y on this machine. I don’t know if anyone was confused on that or not but I just wanted to clarify.

Thanks for all of the responses so far. I will be trying some things in the morning.
Stevo

stevo1 · Jun 25, 2009

I know I missed a few people on answering their questions. You were posting while I was typing.

Sorry I will get to ya when I get a bit of time

. I got a few things to button up at the shop then I gotta run....my daughter says if I am late 1 more time shes gonna string me up :willy_nilly:

.

stevo1 · Jun 26, 2009

barbter said:
stevo
On reading again, would it be a major issue in a tempory re-program?
Make the machine pitch (XY) the holes which will illiminate the rotary?

Given the machine can’t reach past the center of the table I would have to program this in quadrants. I have macros written for every configuration these holes come in. They go from equally spaced with 1 or 2 offset holes. They go up to 12 spaced as shown (not equal), equally spaced within 90deg. To write these into macros would be very time consuming, it is just easier to spin the table to the degree needed. I would rather try and fix the issue at hand and leave that as a last resort. We are still up and operating. I just don’t like to leave checks and adjustments to the off shifts if I got a 75k part sitting on the table. So in the mean time I do it until we can fix the problem.

CMLbend said:
Right off the top of my head I agree with psychomill's first post about the encoder. What voltage/amperage do you run off. Is it on an isolated circuit , sometimes welders can create a spike that will throw off an encoder even if it is a non-malfunctioning part.

I am not sure exactly so I will have the maint guy get me the answers. We do have a lot of welders in the shop but they are in a different cell across the shop. Don’t know if that matters or not.

All,
Ok I am also doing a test today. I came in this morning and the machine had been powered down all night. I did a power up and home. I also set my X0,Z0 around the test bar. I got it within .0002”. I then wrote a looping program to run the machine on all axis to the end of their travel at the same time. This I did for about 1/2hr and checked the machine. X stayed 0 and Y is running out .0025 (+.0015 to -.001) around the bar.

Hmmmm maybe you guys were correct in thinking thermal. Or it could be the encoder. I have not checked it yet, it could still be the problem. I would think that I would have variation in the X if it was thermal??No??

Anyway I wrote down the data of adjustment and started the looping program again this time to run for 1hr. I will then check the numbers. I would like to continue this all day and collect the data. Then power down and let it sit all night and check it first thing in the morning (cold) to see if it returned to 0,0.

What do you guys think? Is this a sound way to check? Should I be collecting any other data?

Stevo

Tonytn36 · Jun 26, 2009

Stevo,
Have you got a hand held electronic thermometer?
You are on the right track with your testing. I'd check it every hour and record the results. This definitely sounds like a thermal issue. BUT, with enough data, it can be compensated for.
I'd be checking the temperature of critical components while I was doing this, and recording those along with the other data. This allows you to get a Time/temp/change plot where you can build some compensation into the programs. Check temp of Y casting, Y screw, motor, etc and record the run times and stop times also.

Machines are not repeating

Cast Iron

Titanium

Cast Iron

Titanium

Cast Iron

Hot Rolled

Stainless

Titanium

Cast Iron

Cast Iron

Hot Rolled

Diamond

Titanium

Diamond

Titanium

Plastic

Cast Iron

Cast Iron

Cast Iron

Diamond

Similar threads