What's new
What's new

Effects of temp gradients and general deflection in tools and work pieces.

Demon69

Titanium
Joined
Mar 3, 2013
Location
Area 69.
Thought id start a thread on these topics as theyre hot right now boom boom :D

All kidding aside, is an interesting and important thing to consider when trying to fit or measure. Have been caught out a few times so will try and recount those when I get a mo. Am thinking to video some experiments that try and capture things as they occur, a straight edge up on blocks with an indicator set underneath with heat being applied in various ways springs to mind.

Im sure most of us aware at least in some way how temp gradients and deflection under load etc etc can move things around, but when should it be of concern to us? Where is a sensible place to draw the line in a given situation? With different objectives theres no one way or answer imo, but should make for interesting discussion, and a place to list some GOTCHAS.
 
Thought id start a thread on these topics as theyre hot right now boom boom :D

All kidding aside, is an interesting and important thing to consider when trying to fit or measure. Have been caught out a few times so will try and recount those when I get a mo. Am thinking to video some experiments that try and capture things as they occur, a straight edge up on blocks with an indicator set underneath with heat being applied in various ways springs to mind.

Im sure most of us aware at least in some way how temp gradients and deflection under load etc etc can move things around, but when should it be of concern to us? Where is a sensible place to draw the line in a given situation? With different objectives theres no one way or answer imo, but should make for interesting discussion, and a place to list some GOTCHAS.

I like the thread, but i would propose to move it to metrology, so you get more exposure. Just a side note, when we went to NIST they explained to us the temperature controls in the two main rooms, and how they are stabilized them using a large number of sensors hanging into the space at various points. Was fascinating.

dee
;-D
 
Thought id start a thread on these topics as theyre hot right now boom boom :D

All kidding aside, is an interesting and important thing to consider when trying to fit or measure. Have been caught out a few times so will try and recount those when I get a mo. Am thinking to video some experiments that try and capture things as they occur, a straight edge up on blocks with an indicator set underneath with heat being applied in various ways springs to mind.

Im sure most of us aware at least in some way how temp gradients and deflection under load etc etc can move things around, but when should it be of concern to us? Where is a sensible place to draw the line in a given situation? With different objectives theres no one way or answer imo, but should make for interesting discussion, and a place to list some GOTCHAS.
.
.
i wrap .0001 indicator bore gage with rag to protect from heat of my hand. i often see gages change more than .0002" just holding in hand too long
.
how else could you bore to +/-.0001" tolerances ??
.
deflection under load ? how else can you get .0002" flatness tolerances milled unless you take light .0005" final passes ?
 

Attachments

  • Boregage.jpg
    Boregage.jpg
    90.2 KB · Views: 239
  • WarpageRemoving.jpg
    WarpageRemoving.jpg
    68.8 KB · Views: 218
i use a 2000 gallon temperature controlled coolant system on 4 large cnc mills but just coolant evaporation can cause a localized cooling effect.
.
many a part on rechuck shifts around a bit as its size changed with a slight temperature change. part rotating in Azimuth is common. i often have to recut parts cause over a .0004" per 10" tolerance after rechuck
.
sometimes it is as simple as cutting at .001 degree on B axis
 
Thanks Tom, am aiming more at the scraping crowd tbh but interesting to hear about the bore gauge and coolant. While were talking machining, how would you go about finishing say 2' x 6" x 3" block to close tolerance? Would you get get the thing close all over then rechuck for the finish?

Im thinking stress induced or relieved when taking any meaningful cut might hurt things.
 
So heres a time I got caught. :)

Prints fine with the fresh scraped prism.
114876-a.jpg


Checks out with the cross slide.
114875-20140815-125831.jpg


Indicates well end to end on the plate. Here
114874-a.jpg


and here.
114873-20140815-132248.jpg




Job done right! :)



Until Richard suggested that the prism might well be sagging and that I should check it.

He was right, as usual lol, and Id scraped that sag into the surface :o. When checked with the camelback and back on the plate it was found to be low in the centre around 0.0003. Hadnt checked the centre before, why should I!? Fresh scraped nicely hinging prism.
115793-20140823-135108.jpg
 
Thanks Tom, am aiming more at the scraping crowd tbh but interesting to hear about the bore gauge and coolant. While were talking machining, how would you go about finishing say 2' x 6" x 3" block to close tolerance? Would you get get the thing close all over then rechuck for the finish?

Im thinking stress induced or relieved when taking any meaningful cut might hurt things.
.
.
i do not know of any close tolerance parts that are not rechucked at lower torque and let spring to shape it wants to be before final cuts which are often .001" or less. i often machine far larger parts in the 10 to 80" range that need to be within .0005" tolerance flatness, parallelism, perpendicularity and a lap tolerance between passes of under .0003"
.
supporting parts on their Airy points is fairly standard. sometimes cause shape is not a simple rectangle the weight distribution needs points where part is supported adjusted a few inches
.
very few parts i make are simple rectangles. some parts need stress relieving multiple times, before milling, before grinding, etc. i often see some parts when machined and unchucked distort.
.
for measuring i often use a dial indicator and the CNC digital readout. i can zero indicator, zero digital readout, move to next location til indicator zero and read digital readout to .0001, sometimes have to clamp gage block on part bottom, zero to top of gage block and move and move to part top and see what DRO says. some CNC need to move only in one direction as they have backlash and some machine sections rock like 4 legged chair with one leg short. i see backlash and rocking shifts of over .0002" quite often. compensating for rotary table center of index error where 2 bores on opposite side need to be the same distance from a perpendicular side to less than .0003" difference especially challenging.
.
i use ring gages and gage blocks with indicating gages and compensating for errors often measure to .0001", i often see differences in say a inside mic to outside mic calibration of over .0003" especially when over 8"
.
larger parts being milled i often use trim blocks. these are blocks bolted to table and a .001" cut is taken and then indicated to confirm they are all within .0003" , this is done cause most milling tables are not flat to below .001". then when part is put on trim blocks the part can be milled parallel to .0005" per 40" tolerances
 

Attachments

  • 600HC_Yslide_CoolantSplashHorzHead_smaller.jpg
    600HC_Yslide_CoolantSplashHorzHead_smaller.jpg
    92.8 KB · Views: 172
So heres the first video with a full grip in the centre of the camelback for 2 minutes. The measured change was 0.00035 (it went concave). The temp of where I had my hands rose around 2°c. temp of hands around 30°c

Was set up on blocks like so and the camera and indy were set behind.
attachment.php


 
And the same test using a finger tip pinch grip at the ends of the camelback.
No measurable indicated change and no measurable temp change on the casting.

 
So heres the first video with a full grip in the centre of the camelback for 2 minutes. The measured change was 0.00035 (it went concave). The temp of where I had my hands rose around 2°c. temp of hands around 30°c

Was set up on blocks like so and the camera and indy were set behind.
attachment.php


.
.
thats why high precision levels have plastic "insulating" grip surfaces. even inside mics often have plastic "insulating" grip spots. me wrapping a rag around a .0001" indicating gage is cause it will change size if i held with my bare hands.
.
your straight edge would be better if you put on plastic "insulating" grip material. i find even wearing gloves is not enough to insulate my hand.
 
I wear tig welding gloves to handle my large camelbacks. The smaller one I can pick up with my fingers but I gotta get a good grip on the 36/48".
 
I know where this thread originated from, and it looks like I created a monster. Ha Ha! :D A good monster however and I think many people will learn from it.

In my opinion, when should temperature and gradients be considered? Always! As I said in another thread, I think if you're always conscious and aware of it, you're likely to be ok. Just get used to that being in the back of your mind the whole time. But it doesn't mean it will be a factor in reality.

When will it actually cause an issue? That depends. In short however I'd be super cautious with work that has a high height to thermal mass ratio. In other words is high but not very heavy. Camelback straight edges must be one of the worst, so it doesn't take much heat energy to increase the back by (in this case) 2 degrees. If that camelback was the same approximate dimensions but more solid, say as a machine part, it would have a much higher thermal mass and could be handled far more without affecting the shape.

However, if you were to take 2 similar shaped parts, one with a low thermal mass (eg a camelback) and one with high thermal mass (eg a machine saddle) and put the same temperature differential into them (eg letting sun shine on the top of them through a window), they will both move approximately the same amount.
 
Well I will put in my $1.56.

Take what I post here with a pinch of sugar to keep yourself sweet.

I have seen plenty of old photos of US manufacturing plants and repairers and to be honest they do not seem to be taking too many precautions against thermal imbalances creating problems. I think just be sensible about it and you will have no problems.

Do not place your workpiece where the sun will be blazing in on it. Do not manhandle your straight edge for minutes at a time. Carefully whack it on, rub it, hinge it, then carefully whack it off.

I would be just as concerned that you have the workpiece properly set up so it is not sagging.

Only time I take extra precautions for heat is when I am scraping in the straight edge to start with. I want it a known good and I know when you get down to the finalising it, at least with mine, they tend to wander all over the place a bit. But we are talking tiny amounts. Amounts that placing the larger straight edges on the surface plate they hinge differently as the plate is not perfectly flat either. When it gets like that, I say it is good enough.

No need to reinvent the wheel or start to overthink things. Do what they did in the past and it will be good enough.
 
I purchased a infarred heat temp gage and will do some experiments when the Texas class gets to projects when I will have a little more time. I cn share something that happened yesterday. Gary who took last years class we had here and is a Journeyman Machinist by tradeudents and lives close to Steve, volunteered to mill and scrape a HKA-24" Camel Back straight Edge I gave Steve last year and he brought it in. He stated he thought it moved when he was scraping. I got out the new infared gun I bought and the SE was 4 degree's colder then the surrounding area as he had left it in an unheated area the night before and drove over to Steve's shop. I told him we needed to wait for it to come up to room temp before we scraped it. Precision work needs constant temp and as Tom points out holding gages with bare hands and trying to hold .0001 is impossible.

I also believe we should welcome his information as he works for a New machine builder famous for precision machine tools. Thank God they are one of the only new machine builders still based out of the USA.

In the other thread that Demon and Pete talk about is how this wonderful thread got it's start. I know that precision work requires you to insulate gages so you get accurate reading and you can't let mere sunshine hit gages or machines. I can share some info on that later as I have to scoot, teaching the class in 1 hour and have to get ready as I am still in hotel and Steve's shop is 15 min drive. Rich
 
at my old job we had a 30 ton wheel 18 foot dia and 5 foot wide all metal. it had 8 spokes and the wanted to see if still round sitting in storage a decade
.
so they had me bolt on a steel precision linear rail to concrete floor next to wheel and check it is straight with optical alignment equipment. they would next day put a electronic capacitance probe on rail and move along wheel but not touch it to measure gap changes in .001" then rotate wheel to check for roundness
.
next day i check linear rail with optics and it is bowed or curved. it was under skylight window and the warm linear rail bolted on cold concrete expands at different rate. rail got longer and held at ends had to go somewhere so it bowed. i tell them we need to come back on a cloudy day and when we do rail is straight again. we check wheel and found it was not perfectly round but 8 sided a bit matching spokes of the wheel.
.
when wheel was ground after plating the grinder was at the side not the top of the wheel. as wheel turned it would sag between the spokes. cause grinder at the side it never took sag out. when wheel in machine the top compartment was air pressurized and the magnehelic gage would show pressure variations as wheel turned showing the out of roundness matching the between the spokes sag. it would sag 8 times per revolution roughly every minute, 60 times a hour for decades as machine ran. it was used to pour liquid plastic on its top and plastic was peeled off forming plastic film used for photography
 

Attachments

  • wheel.jpg
    wheel.jpg
    89.3 KB · Views: 200
  • wheelgrinding.jpg
    wheelgrinding.jpg
    65.4 KB · Views: 188
Some of the spare wheels for another type of film base had a drive on a timer to rotate them periodically while in storage.
.
.
the bigger wheels used in bld 53 and replated in bld 43 were in storage not moving. they used to rotate i believe every 6 months but the question was was it a waste of time to do.
.
since the wheels sagged between the spokes as the wheel rotated i am not sure rotating in storage did all that much.
.
the wheel compartment in machine was air pressurized with magnelic gage you could see air pressure variation as wheel rotated cause air leaking past seals the gap would change as wheel rotated cause not perfectly round due to sag. i had help setup a capacitance gage to measure wheel roundness without touching wheel surface. the technician recorded readings on laptop computer as wheel rotated. we saw out of roundness, wheel runout and wheel slightly tapered or cone shaped.
.
all things considered a 30 ton cast iron wheel with copper, nickel and chrome plated and polished to mirror finish the roughly 36 wheels used in bld 20 and bld 53 made a lot of photographic film for over 70 years.
.
all gone. bld 53 is gone, bld 50,52, 57, 9, 60 and many other hugh buildings all gone, torn down
 
film casting wheel

picture of wheel used to make photographic film. the smaller wheels were transported from the foundry on a horse drawn wagon before railroad was used
.
i believe when it says wheel was 2.5 tons they meant 25 tons. no way the wheels were only 2.5 tons they used a 50 ton crane to pick them up as 10 ton crane would not pick up
.
of course the foundry used to make the wheels is gone too.
 

Attachments

  • wheel_horse.jpg
    wheel_horse.jpg
    81.1 KB · Views: 149
..... Do what they did in the past and it will be good enough.

Maybe, but tolerances now on many parts are 1/10th of what they were 40 years ago so maybe not if you want to be WCM.
One can now mill and turn to what used to be grind tolerances, one can grind to ....... so straight, square, and flat on a machine tool takes on a new meaning.
The thou of the past "in the days of the giants" is now replaced by microns. Tenths in inches were hard, now tenths in microns.
Bob
 








 
Back
Top