What's new
What's new

5 Axis Centres Thermal Movement

It has not yet and probably never will run for long enough... 1-3 hours 10-15 times a week. Longer cycles are zero bother if the machine stays accurate.
I just assumed that the machine builders would have a way of keeping track of the kinematic centres, the idea that this would be an issue for the operator didn't cross my mind. The idea that on a machine of any caliber above a haas they'd be variable by more than microns I'm finding cognitively dissonant.
Most of the parts it will make are half round mold inserts, it was 8 setups minimum before including milling, surface grinding, 4 jaw lathe, 3 jaw lathe, more milling etc so to 5ax them cuts about 10-20 hours off the labour depending hence the justification for the machine, otherwise it would be a bit of an over-indulgence.
This particular task is a bit of an unusual job but if we can nail it down will create some enourmous improvements for our mold shop.
I've just got to work so I'm gonna run an urgent 3 axis job on it, then break the vice down and run the probing cycle a bunch of times and see how much it moves as the machine cools down.
 
I just assumed that the machine builders would have a way of keeping track of the kinematic centres, the idea that this would be an issue for the operator didn't cross my mind.

How does the machine automatically know when your floor has settled, how much weight has been on the table the last few months, or which way things got tweaked last time somebody smacked the vise? Once it is setup, the center of rotation should be relatively constant, but even on a good and well cared for machine, it will often drift a few tenths every month. On a budget platform like that Mazak, it's going to be even less stable.

The idea that on a machine of any caliber above a haas they'd be variable by more than microns I'm finding cognitively dissonant.

#1 - I think that you really don't appreciate how many variables are at work here, and some of your questions suggest that things aren't quite right from the outset. If you go back to my list of three most frequent issues; I can almost guarantee there are problems with one if not all of those items.

#2 - Your expectation that the budget Mazak is going to be leagues above a HAAS in quality and accuracy is completely unfair. Would you buy a Chevy Malibu and expect it to perform like a Corvette, just because they are both built by the same company? Brand is not nearly as important as model.
 
Last edited:
Valid points, its been a lesson in expectations vs reality with a healthy slice of unknown-unkowns. It's also hard to think of a machine that costs over 150k as budget but I accept this is a different game and the tools are priced accordingly.
Didn't get round to probing cycles, probably pop in tomorrow and do that and will post accordingly.
 
Valid points, its been a lesson in expectations vs reality with a healthy slice of unknown-unkowns. It's also hard to think of a machine that costs over 150k as budget but I accept this is a different game and the tools are priced accordingly.
Didn't get round to probing cycles, probably pop in tomorrow and do that and will post accordingly.

No horse in this "Race" ~ interesting read from ALL contributors of course.

Theoretically the "Layout" of the machine is optimal.

So I'm wondering if something isn't bolted down properly or is moving about / not torqued correctly ?

As mentioned before (in thread) - floor settling or foundation issues.

What's your foundation like ?

Personally a half-arsed gander at the machine seem like it would be good for five axis prismatic-ish work but not a go-to for simultaneous cuts that need to be ridiculously accurate.

I'm guessing testing out cut repeatability of a quick "dumb" part in different static orientations might be one way to eliminate if anything is loose / not bolted down correctly - spindle casting / saddle or mated upper to lower castings ?

I have to admit I'd be super suspect - with all the re-monkeying around that something wasn't loose / not tightened correctly and maybe a factory "Build error". + double check orthogonality for everything.

@yowxman - Good luck I'm sure you'll figure it out.
 

Attachments

  • cv5grab_1.jpg
    cv5grab_1.jpg
    51.2 KB · Views: 26
Been running probing cycles all afternoon. Ran it 3 times slowly and stabely maybe 1.5 micron movement at most and then run a thorough warmup for 40 minutes. Everything moved, z moved 13 microns (nearly 3 metres between z motor and base so not wild) Y moved 5 and then kept moving in the same direction as everything else cooled down and went the opposite back to original conditions. Engineer is back tomorrow so will go over it with him.

Was getting the same error in simultaneous as in 3+2 test pieces.
It's a good design I think but if Mazak don't expect it to move about theres something up with the thermal comp and maybe Y ballscrew. If they do it's probably alright but either way I'll keep digging till the bottom is reached.

Thankyou for the luck, it may be needed 😂
 
View attachment 386963
Makes a great finish though.
^^^ Beautiful finish ~ hard to find hard specs on this machine - machine is really new and homegrown to Europe / UK -

Buuuuuut I thought most MAZAK machines have or push "Core Cooled" ball screws + pretty decent (large) diameter

and fine enough pitch (ballscrews) + chiller ? I'm sure they will do their best to make it right as there's some "Pride at stake"...

for said homegrown machine

:cheers:
 
Glad your running tests!
Know your tools and know their behaviors.

This might sound funny but you might try to put a few light bulbs in the machine to keep the thing warm. The machine sounds like it’s fighting hard to reach its thermal equilibrium. Why not keep it there with a few light bulbs.
 
^^^ Beautiful finish ~ hard to find hard specs on this machine - machine is really new and homegrown to Europe / UK -

Buuuuuut I thought most MAZAK machines have or push "Core Cooled" ball screws + pretty decent (large) diameter

and fine enough pitch (ballscrews) + chiller ? I'm sure they will do their best to make it right as there's some "Pride at stake"...

for said homegrown machine

:cheers:
They're pretty chunky, but core cooled is only on the 18k spindle, and we went 12k because we cut mostly steel.
They've been stellar so far, so hopefully we reach a conclusion and resolution soon.
 
Glad your running tests!
Know your tools and know their behaviors.

This might sound funny but you might try to put a few light bulbs in the machine to keep the thing warm. The machine sounds like it’s fighting hard to reach its thermal equilibrium. Why not keep it there with a few light bulbs.
If thermal equilibrium is what it needs I will buy foamcore panelling to build a room and an hvac/heatpump thermal control unit. Obviously I would prefer the machine just comp the temperature if it can cos it doesn't appear to be set correctly unless I'm asking too much of it, but we will see.
Also I assume you're in the states because we can only buy LED light bulbs in the UK now and they are efficient.... that's where the good points end.
 
No horse in this "Race" ~ interesting read from ALL contributors of course.

Theoretically the "Layout" of the machine is optimal.

So I'm wondering if something isn't bolted down properly or is moving about / not torqued correctly ?

As mentioned before (in thread) - floor settling or foundation issues.

What's your foundation like ?

Personally a half-arsed gander at the machine seem like it would be good for five axis prismatic-ish work but not a go-to for simultaneous cuts that need to be ridiculously accurate.

I'm guessing testing out cut repeatability of a quick "dumb" part in different static orientations might be one way to eliminate if anything is loose / not bolted down correctly - spindle casting / saddle or mated upper to lower castings ?

I have to admit I'd be super suspect - with all the re-monkeying around that something wasn't loose / not tightened correctly and maybe a factory "Build error". + double check orthogonality for everything.

@yowxman - Good luck I'm sure you'll figure it out.
Also two things I forgot to answer:
It's bolted down to a 10 inch thick reinforced concrete floor rated for 35kN (i assume per square meter).
There is a weird oscillation thing in the x axis when going from rapid to stationary that's on the engineers list, about 10 microns in the spindle head and shows up in the base about 4 microns although the feet have all been checked and it's got weight all round and "pulled down" onto the feet.
 
Valid points, its been a lesson in expectations vs reality with a healthy slice of unknown-unkowns. It's also hard to think of a machine that costs over 150k as budget but I accept this is a different game and the tools are priced accordingly.
Didn't get round to probing cycles, probably pop in tomorrow and do that and will post accordingly.
I think some climate control in the shop will help a lot.

But yes, 150K is about 1/3 of what the higher end German machines cost in the same size class.
 
What kind of kinematic calibration is available on a Mazak?
I calibrated a machine Kinematic last week. probed a point all over A180° and flipping C around. Most measurements were sub-micron deviation. Cut the part dry and things heated up. Tolerances= out the window.
 
Better climate control will help but it also sounds like you're expecting $600k-1mil performance from a $150k machine. For Mold and Die I'd be looking at a Mikron, Hermle, Röder, or Kern before anything else if you're looking for <1micron full 5 axis blends.

If you're try to realistically hit blends closer than .0002" (5micron) you better be prepared to shell out some serious coin on every aspect of the process. Machine, toolholder, endmills, workholding, foundation, HVAC, inspection capabilites, etc....
 
Last edited:

yowzman:​

You mention the foundation thickness is 10" - this seems a bit thin. The rule of thumb for foundations for reciprocating machinery is that the foundation block should weigh 3x the machine weight. If the foundation contractor didn't carefully compact the soil under the block the foundation could be rocking as the machine moves. I would think you could put an indicator on the block and cycle the machine to see if there's any movement there.

Regards,

DB
 
What kind of kinematic calibration is available on a Mazak?
I calibrated a machine Kinematic last week. probed a point all over A180° and flipping C around. Most measurements were sub-micron deviation. Cut the part dry and things heated up. Tolerances= out the window.
Renishaw C and B probing cycle can have as many positions as I want pretty much. That's where I was yesterday when I ran a warmup. Problem is if I run a warmup, the process doesn't keep the machine warm enough and it starts cooling down again. Keeps coming back to that thermal equilibrium. Gonna take some finding.
 
  • Like
Reactions: mkd

yowzman:​

You mention the foundation thickness is 10" - this seems a bit thin. The rule of thumb for foundations for reciprocating machinery is that the foundation block should weigh 3x the machine weight. If the foundation contractor didn't carefully compact the soil under the block the foundation could be rocking as the machine moves. I would think you could put an indicator on the block and cycle the machine to see if there's any movement there.

Regards,

DB
It's through the whole building, and the machine is right in the middle of 2 stress relief cuts so essentially the foundation block is 6m x 6m. Thicker would have been nicer but no one from Mazak seemed concerned about the floor so neither were we.
One of the techs did that test and saw 4 microns if the machine rapids and stops in X. He didn't like that. I am ignorant to the significance of that.
 
Better climate control will help but it also sounds like you're expecting $600k-1mil performance from a $150k machine. For Mold and Die I'd be looking at a Mikron, Hermle, Röder, or Kern before anything else if you're looking for <1micron full 5 axis blends.

If you're try to realistically hit blends closer than .0002" (5micron) you better be prepared to shell out some serious coin on every aspect of the process. Machine, toolholder, endmills, workholding, foundation, HVAC, inspection capabilites, etc....
It's less 5 micron absolute position but more 5 micron repeatability with the coordinate system. It's finishing a sharp corner on an insert pocket. We know from experience if there's deviations in the edge surface it frets during lockover and damages the inserts, the pocket itself is 120mmx240mm, plus 20um minus 0. Inserts are half round into the straight wall pocket. Not my design, but we have over 200 tools in this system which would all have to be remade overnight if we changed it. If i can sweep it with an indicator and see 5 or under I'll be confident putting it into production.
 
Any updates on these issues Yowzman?

I'm also a UK based machinist, potentially interested in one of these Mazaks.

Thanks!
 
Hey dude,

Service went over the machine with a fine tooth comb, brought down a bunch of kit, made sure all the rails were straighter than spec etc.
What it has come down to is we should've been advised to have scales on the machine and possible the 18k spindle (this option is required for cooled ball screws) for the tolerances we're trying to hit. From what we've seen on this particular machine general stuff under 50 microns you'll be fine, anything under 20 microns needs the high accuracy kit in the specification and in-cycle probing. That is with all the axis going and getting near the limits of the work envelope. I'm very happy with the machine in every regard other than it's not hitting the accuracy we needed, and Mazak have been extremely good about the issue so far if a little slow, you might say "too thorough". I'll DM you my email address if you've got any more detailed questions, I'm happy to answer them.
 








 
Back
Top