PDA

View Full Version : spindle stretching



darryl
02-06-2003, 05:23 AM
I've been machining something on the mill, then taken a break, come back to it, and found the cutter height reading has changed. Typically, this will be around 3 to 4 thou, and is temperature related. When the spindle is warm, the cutter is lower. Vice versa. How does one go about machining a surface, required to be accurate, knowing that the cutter position is going to change as you progress through the work? I can understand preheating the machine by letting it run for a while, or continual checking and readjusting, but that seems inefficient. How do production shops handle this, and how does a cnc machine handle this? I noticed this today while working on my folding guitar. The piece I was working was a flat aluminum strip clamped to the table, and stayed at the temperature of the table during the milling operation, so I'm ruling out the workpiece expanding or contracting, or warping for that matter, since it is clamped closely to where the cutter was working. Am I correct in thinking the spindle is expanding, or is it more likely to be the chuck holding the endmill, or the endmill itself? So many questions---

SGW
02-06-2003, 08:30 AM
Welcome to the rubbery world of precision metalworking!

The expansion rate for steel is (about) 6 millionths of an inch per inch per degree F. If you have a 1" length, a three thou expansion would require a temperature increase of (about) 500 degrees. (.003 / .000006)
If the part being heated is 10" long (the machine spindle, perhaps), then 50 degrees would do it.
(Maybe somebody can check my math....)




[This message has been edited by SGW (edited 02-06-2003).]

ShavingMaker
02-06-2003, 09:40 AM
The effect of temp change on metal can be significant in many areas where you wouldn't think of it. I worked in real estate for many years and more than once came across problems with surveys matching up.
When the US Govt surveyed the lands west of the original thirteen colonies to get ready for homesteading in about 1858-1862 they used some pretty primative tools. Measuring was done by a "chain" of 66 links, one foot to the link, 80 chains to the mile.
The chain was measured at a standard temp of about 75F. If the day was hot and got up to about 90F the expansion of the steel chain would result in an error of about 8 inches in a mile! The survey crew had to check the temp regularly and adjust their distances to keep accurate. Some did, some didn't. I have seen some "adjustments" of up to 100 feet where two crews came together from opposite directions, so obviously one of those crews was not adjusting along the way.
Expansion due to temp does matter!

lynnl
02-06-2003, 11:41 AM
I'd think that's an erroneous assumption to rule out the workpiece expanding/contracting, even tho it is clamped directly to the table. Also, do you have the knee and stationary axis locked thruout the 'before', 'during', and 'after' operations when the measurements are taken?

docsteve66
02-06-2003, 01:17 PM
ShavingMaker: Those places where the surveys did not match up were sometimes deliberate (so I understand). They were sold.

One of my favourite cartoons is the one where where the Transcontinental rail roads met end to end. Big Irish man pointing back and little China man pointing back the other way down his stretch of rail. The rails join perfectly on one rail, the other is "off" exactly the width (gauge) ofthe track.
TEMP CHanges?
I know some of the old CNC machines compensated for "tool growth" in some manner.

You can bore at noon time an engine block,Mike it and its round, at five oclock remike, there will be small changes. come in next morning and there other changes. All due to Delta T. Horrible thing is you can see where the changes ocurr coresponding to webs, coolant holes etc. So you put a re-inforcing block as a dummy head, torque it down tight, and bore. When you take the bolts out there will be groooves where the bolts forced the cylinder into the boring bar. Support the engine block where the motor mounts are and the holes are no longer round. I suspect that it all changes when the engine heats in use. So what is a round bore? Beats me- depends on when you measure it.

Some very good advice is in SGW's "Welcome to the rubbery world of precision metalworking!". Think of the machine as rubber, the tools and work piece are also rubbery. The more precise you want it, the more you gotta compensate forthe rubber effct. Precision, safety, pollution, noise all follow a general rule: to move the decimal place over one place cost just as much as to move it the first place. In other words, to make something to one inch may be cheap, but it will cost probably ten times more to get .1 inch and then ten times that to make it .o1 etc. As we learn to measure (pollution is good eaxample) smaller amounts the paper pusher and theory boys hold us to tightertolerances. Today we call pollution in parts per billion and a few years back we could not measure those amounts. "Pure","round", "flat" "quiet" all change to meet needs, be those needs real or imagined.

Endmill
02-06-2003, 01:58 PM
There is another possible explaination, internal stress. Aluminum plate is severely stressed internally during the rolling process. To remove this, sheets are "drawn" out, which is an uncertain process. Finding a method of measuring internal stress is currently an area of very active research, but as yet, there has been no solution.

I once split some brass bar stock with a slitting saw, and it peeled apart and made a good tuning fork. Machining also induces stresses which cause later changes in dimension which are not necessarily related to temperature.

lynnl
02-06-2003, 02:37 PM
... Strange! normally the double post immediately follows the original. This one (herewith deleted) wasn't posted til after other intervening posts by other members. Oh well, another of life's mysteries I guess.

[This message has been edited by lynnl (edited 02-06-2003).]

Spin Doctor
02-06-2003, 06:10 PM
Machine tool spindles are designed to allow for growth due to temperature increases. However the growth is usually minimal at the tool holding end of the spindle. Normally the rear bearing set is mot locked against a shoulder in the bearing bore and the fit tolerences are a little looser allowing the bearing set to float back and forth in the bore. If both the front and rear sets were loked against a shoulder heat would force the arbor to spring in the center between the bearings. This may sound outlandish but next time your in the mill and you raise the knee to the position you want for size set-up a dial indicator from the column to the knee and see if the table drops at all. I've seen it happen

Thrud
02-08-2003, 12:14 AM
darryl:
CNC machines are supposed to be warmed up. Some manufacturers like Haas now have a spindle warming feature built in that will autostart before the shop workers arrive thus allowing them to start working immediately on arrival.

Spin Doctor
02-08-2003, 11:11 AM
Thrud; we run transfer lines for automotive componets, cyl hds, blocks etc. The transfer lines have to be run befor production starts to warm up the machine, spindles and stretch the transfer bar to proper length.