What's new
What's new
By:
Advanced search…

How is accuracy of modern CNC machines affected by temperature changes?

Finegrain

Finegrain

Diamond
Joined
Sep 6, 2007
Location
Seattle, Washington
So I have this part that is ~12" long and has a couple reamed holes ~11.5" apart, +/-.002" on the hole spacing. The shop is at 55F this morning. I do the math, 24 millionths per *K per inch, and I figure the thermal change will be .00199", perilously close to the tolerance. Then I got to thinking, wouldn't the machine itself also be affected? Granted it is iron and steel, so would not change as much, but still, it seems like the machine components would also be smaller on a cold day. This is a 2016 Brother Speedio S700X1.

Not talking about thermal comp settings in the control BTW. I get that as the machine heats up the control can compensate for expansion.

What do you think?

Regards.

Mike
 
Finegrain said:
So I have this part that is ~12" long and has a couple reamed holes ~11.5" apart, +/-.002" on the hole spacing. The shop is at 55F this morning. I do the math, 24 millionths per *K per inch, and I figure the thermal change will be .00199", perilously close to the tolerance. Then I got to thinking, wouldn't the machine itself also be affected? Granted it is iron and steel, so would not change as much, but still, it seems like the machine components would also be smaller on a cold day. This is a 2016 Brother Speedio S700X1.

Not talking about thermal comp settings in the control BTW. I get that as the machine heats up the control can compensate for expansion.

What do you think?

Regards.

Mike
Click to expand...

Probably about as fast to run one and check it (providing you have a "good" way to measure and repeat, etc). Run one at nominal and check right off the machine and check again at the end of the day. To answer you question about the machine moving, I assume it would take longer to see the changes because of the mass, but that is pure speculation. :o
 
I think your shop is warmer than mine this morning. :D

I came up with .0022" over 11.5" if it's .000024/inch per *K. That's assuming 55F to 70F (shift of 8*K). I'd be surprised if it really did that.

Does the machine have scales or is it going by the encoders?

You could always fudge the numbers closer together by .001"...
 
NO THINKING INVOLVED
This is the oldest story in the book. If you want to hold tolerance,the part and the machine and the thing that is measuring the part must all be the same temperature. The larger the part is ,the more important it is. This is why precision shops have been temperature controlled since 1946 or so.I guessI am getting too old.When I read something like this ,it is like someome telling me that they have discovered fire .Edwin Dirnbeck
 
edwin dirnbeck said:
NO THINKING INVOLVED
This is the oldest story in the book. If you want to hold tolerance,the part and the machine and the thing that is measuring the part must all be the same temperature. The larger the part is ,the more important it is. This is why precision shops have been temperature controlled since 1946 or so.I guessI am getting too old.When I read something like this ,it is like someome telling me that they have discovered fire .Edwin Dirnbeck
Click to expand...


:popcorn::popcorn:
 
First it is beyond my comprehension why one would not heat the shop

Second, cast iron moves about half of aluminum, IIRC, so the machine is not that far off.

Third, if you heat the shop sporadically, the machine is probably not at 55 degrees

Forth, you can bring the temp of the part up before machining.
 
If you regularly deal with tight tolerances, no matter how big your parts are, conditioning the shop is mandatory.

The "HVAC is too expensive" excuse is much like thinking spindle probes are too expensive. Once you see how huge an impact it has on things you never even considered (faster and more consistent setups in a conditioned shop, and even simple parts becoming somehow easier to run, for example), the money you spent becomes a distant and hazy memory. I keep my shop at 68ºF year round. That's 24/7. It's essentially pointless to turn the HVAC off when you're not there and might even make the situation worse, not better.

And then there is the employee side of things. A conditioned shop is considered a large perk to the guys versus working in a cold / hot shop.
 
Can you put the material in a warm office for a few hours, and run a warm up program on the Brother?
 
Yes, an HVAC-equipped shop is on the list, but not likely due to a variety of reasons, cost not being one of them.

The 12" + .002" tolerance + 55F combo is an anomaly. Nearly everything else I do is "less" of one of those three. Smaller, looser tolerance, or temp closer to 68F, or all three.

Temperature of the coolant is all that matters in my Speedio. Regardless of the material temp, once the coolant hits it, the part will be at the coolant temp.

BTW, where are you he-men that were bragging about working in your icy-cold or roasting-hot shops ;). Now all of a sudden we have to have 68F temp control :bawling:.

Regards.

Mike
 
Keeping the shop at a "normalized" temp should be considered overhead, a business expense, not a variable. I help a friend out in his shop a few times a month. This month he wanted me to setup his slant bed lathe to run a .0002 window on a journal 1.000 inches long. It was 4degrees outside when he wanted me to do the job. His heat was set at 77' just to keep up, and you could still feel the chill coming through the cinder block walls. The heat, overhead forced hot air, just blares on the lathe. I told him I'm not making the parts until next week. Next week came, it was 55' outside, and the parts came out fine...until you leave them on the ice cold granite surface plate that is. What you're asking shouldn't even be a thought. If it is, perhaps raise the shop rate to cover such expenses.
 
Having the part at room temperature is one thing, the machine is another. However, it's not the cast iron you need watch, but the steel ball screws. The ball screws getting shorter would be what makes the part off (no scales). Steel coefficient of expansion is 1/2 of AL, so it should be just over a thou. Probably less as it will take longer to get cold the the air in the room. I've thought of running a electric heater pointed at the frame during cold nights. The cast iron is a good heat sink. Like a cast iron tub, they suck a lot of heat getting up to temperature, but they stay there for a long time too.
 
Finegrain said:
Yes, an HVAC-equipped shop is on the list, but not likely due to a variety of reasons, cost not being one of them.
Click to expand...
uhh, then what is it?

A torpedo heater under conditions where it might get below a decent temp would be effective, so would enough nickel and dime electric heaters

You are having a problem, you know the solution, but money is not the issue.....

You aint makin sense


BTW my shop goes from 67 in the winter to 79 in the summer, and I would have better AC if it mattered to me
 
Mud said:
Would the scales change things more or less than on machines w/o scales? Aluminum spars, glass scales inside.
Click to expand...

Depends of how nice your scales are I suppose: link

per crappy rental conditions. I build foam shrouds over my roll up doors because they were very leaky. A few hundred bucks and a day or two and it helped a lot. Blow foam on ceiling which is a gift to my landlord and portable/window heat and A/C. It was 25 this morning and came into 55 in the shop with one small heater running. I could probably stay pretty close to 68 if I wanted to spend 200/mo.
 
If the cost is not the reason for climate control, then WHAT is? I also do not understand. If you are going to pay all of the cost for CNC machines then proper climate control (heat and AC) is just a part of that investment.

If you really can not have proper climate control, then you can work with the situation. Do some test cuts with long dimensions at various temperatures, with different coolants, and with various materials. See just how they come out and RECORD the variations. Then, when you need to make a large part you can apply those correction factors. Of course, you are going to have to do the measurements on those test pieces in a temperature controlled area and only after those pieces have had time to stabilize at that standard temperature.

Personally I vote for the climate control.
 
Right now it is unusually cold in our area. I generally have no problems keeping my shop at 68 in the winter with a 300 watt space heater. Last night I set the temp higher and switched on the 600 watt turbo boost. It was 55 this morning in my shop as well.

As far as does the temp matter, well yes it does. Try to get everything as warmed up as you can and fudge position a little to be safe. +/- .002" isn't that tight. I would test the coolant temp as well. Just warming the part up won't work, once it is in the vise it will quickly stabilize to that temp even if you don't use flood coolant.
 
Back to the technical question - as noted above all of the parts of the machine have thermal mass, and in fact having the whole shop at exactly 68F doesn't assure than any particular part of the machine will be.

One trick to get around this (I've read about but not tested) is to have some sort of calibrated artifact (a gage block being the high cost example) which could be kept in the same or nearly the same environment - and compared with probe, indicator, etc., to the part. I'm not up for having a spendy 12" gage block stack be washed down with coolant, but that would be a way. (And, of course, you need to account for differences in CTE between the artifact and the work material.... Are we having fun yet?)

In a post some years ago Tonytn36 pointed out that you can't actually have the machine at a uniform temp, so what they had to do was run parts, measure them in the (stable) temps of the process, compute the expected variance, wait for them to temp soak to spec temp, and check them in the lab. Feed back adjustments - repeat. Point being that you learn the particular process will make a part that measures say 0.001" oversize in the machine(s), but when temp soaked to 68F it will be dead on. Probably not something you would get right on the first try.
 
You must log in or register to reply here.

Similar threads

P
Deep Holes with a Difference
Replies
0
Views
5K
PracticalMan
P
P
Balanced equations: Toolholder vibration makes for miserable machining, but balancing
Replies
0
Views
6K
PracticalMan
P
P
Big Machining for Big Machinery
Replies
0
Views
7K
PracticalMan
P
P
Inspection System Brings Programmability to Shop floor Gaging
Replies
0
Views
3K
PracticalMan
P








 

ABOUT PRACTICAL MACHINIST

With more than 10.6 million unique visitors over the last year, Practical Machinist is the most visited site for metalworking professionals. Practical Machinist is the easiest way to learn new techniques, get answers quickly and discuss common challenges with your peers. Register for the world’s largest manufacturing technology forum for free today to stay in the know.

Learn more about us.

STAY CONNECTED

Facebook Instagram YouTube LinkedIn Pinterest Twitter Tiktok

Register Become a forum member. Register today.

Sign Up Sign up for our e-newsletter.

© 2024 Copyright Practicalmachinist.com. All Rights reserved.

Back
Top