What's new
What's new

Parts getting progressively taller

PriddyShiddy

Cast Iron
Joined
Mar 1, 2011
Location
anaheim, ca
Last night I ran my production parts and my OM2A had cut each one progressively higher in Z moving X-

The fixture holds 6 rows of 19 parts. The code calls sub program that faces each part for a row, then calls M97 sub for all 19 parts tool 1, then same for T2 etc. When it's done I have a seperate file that takes a skim cut straight across all 19. When it took the skim cut it was hitting one side and not the other. The parts are tapered to a 0.000" line width then skim cut to the correct line weight (0.004") The actual machining cut each part a few tenths maybe a thou higher with each part right to left on the table.

**EDIT TO ADD*** The parts are 3D machined with tapered engravers so cutting higher in Z makes them wider. I have run 5,000 parts on the machine in 2 months without this issue. I tried again this morning. Same thing. Machine is a 2010 with only 300 hours when it arrived here in Feb.

I pulled the first and last part from a row after I initial facing and they are the same length. Jogged an indicator across the fixture, pushed pulled. Fixture is tight and square. Tried pulling parts out with 14" channel locks, they are tight. I pulled a file from last week that was fine (still in machine hard drive) and it did the same thing. Each part slightly (wider, cut higher in Z) from right to left.

After thinking about it all night I checked the minimum oil system and realized it is still almost full from when I got the Machine (used) 2 months ago. I have 600 hours on the machine in the last 2 months and I'm pretty sure it should be giving more oil than that. Could stiction from inadequate oil cause this? An immediate and repeatable error? Mind you I ran parts earlier in the day and they were fine.

I pulled the Z way cover and while not dripping wet the screw didn't seem completely dry neither did the rails. The oil reservoir is nearly full, pressing the solenoid valve Spike's the pressure up to 30 PSI and the sensor obviously has not detected a lack of oil pressure on the back of the machine where it splits off to each axis. This makes me think it isn't the solenoid valve or air pressure but a clogged line somewhere or everywhere.

Any ideas before I get Haas out here to poke the thing with a stick?
 
small variations as machine warms up, tool holder warms up, spindle bearings warm up, part and fixture warms up, etc is common.
.
i often just take a final .001 cut at the end. or 2 each .0005" cuts
.
i have even seen coolant temperature cool things down too much. sometimes final cuts done dry are more precise. coolant evaporation actually absorbs heat and has a chilling effect. thus even when coolant is temperature control there still can be instability.
.
if you have 12" long tool in tool holder and it gets 10 degrees warmer it actually grows longer about .0007" but whole machine warming up grows taller too
 
Very true Tom. Likely not the culprit for this particular issue. The error is .01" from one side to the other and happened on the first tool after warm up (which hasn't happened in 5,000 parts) as well as the second row which was an hour after an hour of running. The issue disappeared at the start of the second row which would also mean it isn't a growth issue since the second row was high only on the last parts.

HFO had me try checking Z repeat ability. DTI on the fixture, rapid up 3" G4P1., rapid down to .025" then G1F5.Z0. which I ran 100 times and it repeated 0.0002" on both ends of the fixture and both ends were within 0.0003" of each other.

Checked Z backlash which was dead nuts zero on the 0.0005 DTI, X and Y showed less than 0.0002

I can push the table 0.0015-0.002" in any direction but it springs right back. Mind you the Tools are primarily 0.02 or smaller tip so pretty sure it's rugged enough for that kind of hogging.

Put the mag base on the head and DTI on the tool holder and pushed up and pulled down as hard as I could on the tool. Maybe .0001 or .0002 movement on the needle, maybe, but definitely sprung back.

Checked X and Y backlash and

I'm going to try reversing the order and machine from left to right and see if the high cut moves to the right or stays on the left.
 
Did this just happen on one cycle out of the blue?

Yes. The machine got to take it's first 2 days off since I got it while I got caught up. Yesterday afternoon I ran one row of parts and they were perfect. Shut the machine down for a few hours. When I ran 2 rows of parts after warming up this happened. Then ran a third row and checked the parts after the first engraver did it's roughing and it was blatantly happening. Warmed her up this morning and ran a known good code from a week ago (only half the fixture wide) and it was visibly .004+ off by the 9th part.

**Then I tried running just the first and last part and they were .001" off from each other. NOT the 0.01" they were off when I run all 19 parts in the row
 
Interesting that the parts are getting taller not shorter. This is an interesting one. Trak part growth vs how much you use the spindle and air temperature. Are you scrapping parts? Can you keep running the machine? I would want to get a little more information on what is moving under what conditions before calling a tech. Can you do a final finish on all the parts at the end of the cycle to minimize this growth variance between parts.

Food for thought. The first thing I noticed when the spindle bearings went out on a Fadal was increased growth as it warmed up. Long before I started seeing it in the tool marks or the sound of the spindle.
 
servo oscillation or a axis moving back and forth and slowly stopping is common on many machines
.
i often use G4 P20000 or 20 second delays to give time for machine to stop moving. fast oscillation shown the waves are close together slow oscillation the waves can be a foot apart. sometimes just having a cutter start 6" away by the time it feeds over 6" the machine has stopped oscillating
.
also older machines often when direction is reversed and at a balance point a slide can tilt. for example on a horizontal mill a change in Y can cause a change in Z from a tilting effect. i explain as a 4 legged chair on 3 legs at times it shifts so one leg goes up and the other comes down.
 

Attachments

  • ServoOscillationBad.jpg
    ServoOscillationBad.jpg
    81.1 KB · Views: 119
I have already talked to 3 HFO techs/applications and waiting on a call back from the 4th. I have to get these set up on the other machine (takes 3x longer to run but at least it is progress) then I am going to try reversing the order so it runs left to right and see which way the error goes

This is HUGE movement. The tools are only .005 - .02" tips and the parts are .015"+ different in size. I tried taking some pics. I'll try again with a different loupe and see if I can show you.

I did check the tools just now and all three touched off within 0.0001" of their touch off yesterday which is dumb luck since they change by .0005" with spindle temp change.
 
servo oscillation or a axis moving back and forth and slowly stopping is common on many machines
.
i often use G4 P20000 or 20 second delays to give time for machine to stop moving. fast oscillation shown the waves are close together slow oscillation the waves can be a foot apart. sometimes just having a cutter start 6" away by the time it feeds over 6" the machine has stopped oscillating
.
also older machines often when direction is reversed and at a balance point a slide can tilt. for example on a horizontal mill a change in Y can cause a change in Z from a tilting effect. i explain as a 4 legged chair on 3 legs at times it shifts so one leg goes up and the other comes down.

That is a VERY interesting example of what can happen on a huge machine tool. It has been many years since I operated something big enough to oscillate for several seconds, and I wouldn't have had any way to compensate for it other than take it out of gear. Probably not the problem on the OP's Haas, though.
 
Is there any part of the actual table that you can run an indicator across a fair amount in X? Then you can see if the table is dropping as it moves across.
 
tool moving

tool could be moving or slipping in tool holder. i have seen tools pushed in and pull themselves out before
.
other is tool comp in program might be a programming issue
 
Late yesterday I ran the same parts reversed order left to right to see if the error stayed on the left or transfered to the right. Parts were perfect. I ran the EXACT same code from the night before which was still in the conteol (first ones that had the error) and they were fine. I've run another 250 parts wince then all fine.

I didn't adjust anything or change any settings. Same code never left the machine. I ran 4 testes all bad yesterday and never shut the machine down. I'm lost on why it happened and what solved it and when it will be back.

FYI the tools were checked and were within a tenth so they hadn't moved. They wouldn't have moved back for the beginning of the next row them started moving all over again anyways.

There is only 3" on each side of the 100lb fixture where I could axes the table. I will pull the fixture next week and check tram and level again.
 
Could it be a coordinate rotation(skew)in an axis? I could see a multiple point probing cycle possibly doing that. Do you have probing? Other than that, I would say loose Z axis thrust bearing or nut, maybe?
 
No probing, scaling or macros :( may end up buying the macros... But $3K for that just effing kills me. It's free with the $5400 probe, but $3K is just plane NUTS. I called with CC in hand expecting $1700 or $2000 like I paid for the expanded memory but just couldn't stomach it.

Loose bearing I've never experienced. I know I got no movement pushing and pulling on tool and minimal on the head but don't know how for sure to test that. Of course it could have been all the pushing and pulling on the head and access that put a bandaid on whatever is wrong
 
I would think if it was something loose you would see it in the finish or be breaking tools.

That's a damn good point. Honestly I was shocked when I checked the tools and they were dead nuts on. The finishing tool is a 60° 2 flute 0.005" tip and it was perfect. Bearings, nuts, too strong of a breeze would kill the tip off that tool.

I am more and more leaning towards an issue in the control rather than mechanical. I can't fathom anything mechanical that could actually cause (even in theory) the error. If it was one row or if the error was random, it could be a hundred things, but it started the next row out an hour later perfect. It repeated the exact same error with three different tools in the same way on each row. It is exactly as if it started adding a progressive Z+.0005 to every G52 offset call so each part was .0005 taller than the previous then it reset at the start again.
 
usually it is program issue especially with restarts and tool comp or G55 changing to G54 when reset pressed and defaults kick in. or G17, G18, G19 issue when doing 5 axis work
.
some machines have mechanical or hydraulic problem. sometimes probe tip is loose causing varying readings.
.
grid shift check is a program that probes a target (cup) and i record numbers from machine zero return positions. i can detect and adjust by parameters if center or any spot on table has changed distance from zero return position. i check weekly or more often. my fixtures locate to center of table. i need to know the distance from zero return to table slots is correct.
.
and i check 2 or 3 times to see if repeating measurements +/-.0004", i have seen machines jump .004 to .006" cause of a mechanical or hydraulic issue. for example some machines use hydraulics to take weight off slide sections so the move easier but if hydraulic pressure shifts then the Z can change cause it acts like it is on hydraulic springs. i have had machine problems before caused by a weak $1 spring in the hydraulic pressure relief valve causing unstable hydraulic pressure. spring replaced and problem went away.
 

Attachments

  • GridShift.jpg
    GridShift.jpg
    85.6 KB · Views: 79
some machines use stepper motors with no feedback or open loop
.
if stepper motor given command to turn and resistance is high it will not move or skip steps, loose steps
.
closed loop is machine monitors actual movement and if it did not move it keeps trying to move until it does.
.
close loop if at X0. Y0. and i turn handles hard enough and move position and let go it comes back to X0. Y0.
.
some machine do not. if it moved position nothing to cause to to come back to programmed position
 








 
Back
Top