Thermal drift

[JMahon]

Anyone doing tight tolerances on robodrills?

 

I'm having a heck of a time chasing these features around.

 

It's on a brand new robodrill with a koma 5th axis table.

 

I looked into the thermal comp but it has to be "taught" from what I understand...

 

Thoughts?

[Guest]

You have to teach it for it to work properly. Cheating not allowed.

[Newbeeee™]

I have read before that the Y is a problem for other folks.

We have 2x E models (2006) where they move a little (Z) but once running (warm) hold all day long.

I have never played with the thermals as I have never needed to.

[civiceg]

I had issues on parts that were 20 + inches long with a profile of .002 on features in the x. Had to split the tolerance to get good parts. The ball screw was growing .0012 from "warmed up" to end of shift. Never had time to dial in the thermal comp on the X, because parts were too involved to scrap and how close they were to being OOT. I did get the Z dialed in pretty good, but it would still drift a few tenths (if not warmed up fully prior to running). A huge problem when trying to hold a profile of .0005 in the Z over 20 inches. Job runs fine on a Matsuura and Feeler. 

 

I believe it can be done, but trial and error on prototype quantity parts, forget it. The key is keeping the thing running at the same rpms and not letting it sit idle. If I couldn't keep it running parts during lunch or when nature called, I would run an old program above my part to keep it warm. 

 

I am pretty sure I had a value of 60% in the Z, it increased the part thickness slightly but did decrease inconsistencies between one end of the part to the other. Robodrills are nice machines, but the parts we were trying to run on them are not Robodrill parts. 

[civiceg]

Also we were running at 19500 rpm pretty much all day with a 24k spindle. HTH

[JMahon]

You have to teach it for it to work properly. Cheating not allowed.

What do you mean by cheating?

Chasing offsets?

[JMahon]

I'm trying to hold a +.05mm -0.00 tolerance on a feature that's milled at c0. and c180. The feature cut at c180. is out by 2x the error in Y.

The machines are only running 1.5 shifts and i have been spending half a shift chasing my tail before I get any consistency.

Material is approx 500/ft

Pulling out my hair!!

[civiceg]

Know the feeling, best of luck

[Newbeeee™]

I had issues on parts that were 20 + inches long with a profile of .002 on features in the x. Had to split the tolerance to get good parts. The ball screw was growing .0012 from "warmed up" to end of shift. Never had time to dial in the thermal comp on the X, because parts were too involved to scrap and how close they were to being OOT. I did get the Z dialed in pretty good, but it would still drift a few tenths (if not warmed up fully prior to running). A huge problem when trying to hold a profile of .0005 in the Z over 20 inches. Job runs fine on a Matsuura and Feeler. 

 

Curious - what model Feeler?

[civiceg]

HV 1000 With a Koma 4th, its had some issues but overall its been pretty good. Not a Matsuura

[JMahon]

So methods is saying thermal comp won't help me because my dimensions are within a small area.

Isn't thermal displacement calculated from home of the ball screw?

So if "y" center of rotation is approx -8.00" from machine home. Wouldn't thermal comp help me?

[civiceg]

Can you have the machine turn on before your shift and have a warm up program running before you get there? Either via another employee or machine function. 

[JMahon]

Well we tried warm up programs and it didn't really stabilize. Methods suggested leaving out part program running all night and it did fine this morning.

But how long will our spindles last with it running max rpm all night...?

[civiceg]

Yes, running the actual program as the "warm up" would be best. I hope you don't have a 24k spindle running at 24k for along period of time. 

[Guest]

Well we tried warm up programs and it didn't really stabilize. Methods suggested leaving out part program running all night and it did fine this morning.

But how long will our spindles last with it running max rpm all night...?

If you run at like 19,000RPM... that spindle will last a LONG time.

 

Trying to hold .05mm should be doable... challenging... but doable. The warm up program is crucial and honestly, I've had the best success holding tight tolerances by looping the actual part program whith coolant flowing instead of a full/near full stroke warmup, etc...

 

JM2CFWIW

[Guest]

Yes, running the actual program as the "warm up" would be best. I hope you don't have a 24k spindle running at 24k for along period of time. 

Those spindles are rated for a certain period of time @ 24k with a finite lifespan. Keep it at 19k(or lower) and that spindle will live a long life as long as you don't abuse it.

[Newbeeee™]

What I would say, is it sounds like your fighting linear (Y) dims only?

If that's the case, I'd run the loop program overnight with coolant on etc but the spindle override waay down, so it's only running at say 3 or 4krpm.

Don't have it off, as coolant may/will get into the bearings.

[djstedman]

If this were me (and often times it is.. since I deal with this a lot) I would put some money into a renishaw touch probe, then probe for location.. this way your probe 'follows' your part around as the machine drifts around in x y and z.. not sure if probes are able to be put in robo drills but I really cant see why they shouldn't be able to be.

 

Not only will this solve movement in Y causing error, if you also probe in Z you can eliminate issues from runout of the part as well.

 

Also If your stuck going the warm up route, I personally wouldn't lower the rpm's on the spindle way down, since a lot of the thermal growth you are seeing is coming from the heat generated from the spindle running at rpm.. lowering the spindle rpm too much could have a significant impact on the amount of movement you see.

[Newbeeee™]

 

Also If your stuck going the warm up route, I personally wouldn't lower the rpm's on the spindle way down, since a lot of the thermal growth you are seeing is coming from the heat generated from the spindle running at rpm.. lowering the spindle rpm too much could have a significant impact on the amount of movement you see.

On a Robodrill, I'd assume that the Y expansion is because it has the weight of the X and rotary as well, and that 60m/min rapids and high acc/dec = heat = growth?

I really cant see that the spindle/head could affect the Y.

However, it would be easy to do some over night testing to eliminate.

24k spindles are expensive, I just wouldn't want to wear one out prematurely.

 

What I would also add is coolant tank size and temperature?

But I would only think this may affect if the factory heating is turned off overnight etc.

Coolant shouldn't go up/down too much if you're only light milling???

[Matthew Hajicek - Singularity]

I know!  You need to replace your ball screws with belts and pulleys!  

[djstedman]

Not having really worked on Robodrills I couldn't say for sure, I was just going off of experience I had on one of our Mazak's. We mounted an indicator on the table and never moving the column in Y and only running the spindle at 18K saw growth of .0017 in Y after 2.5 hours of runtime, the tech's at the time attributed the movement to the heat created by the spindle, and with a infrared thermometer we were able to measure the casting surrounding the spindle bearings at about 150 degrees F.

 

I just thought it worth mentioning that its a possibility that the spindle itself could be contributing to some of the movement.

[civiceg]

Especially if he is maxing out that poor spindle

[JMahon]

It's only a 10k

[JMahon]

Not having really worked on Robodrills I couldn't say for sure, I was just going off of experience I had on one of our Mazak's. We mounted an indicator on the table and never moving the column in Y and only running the spindle at 18K saw growth of .0017 in Y after 2.5 hours of runtime, the tech's at the time attributed the movement to the heat created by the spindle, and with a infrared thermometer we were able to measure the casting surrounding the spindle bearings at about 150 degrees F.

 

I just thought it worth mentioning that its a possibility that the spindle itself could be contributing to some of the movement.

That's what I was saying! But my boss and methods engineers were saying that wasn't possible!!!! Arggg!!!

 

Also were not running coolant. These machines are dedicated plastic machines in a very controlled environment.

 

So I actually got parts to repeat well today. That was after running part program all night long with thermal comp on. So now my next test is how long it takes to get it stable enough to run real parts...

[Newbeeee™]

^^^our Robos are 10k and they can run all day and the spindle/head feels stone cold after a hard days work.

If you had a spindle or motor going bad I can see heat being generated and it has to go somewhere but???

Million dollar question - you say plastics. Sure it's not the jobs moving...