View Full Version : Using temperature as a fastner

01-23-2005, 01:41 PM
I've heard you can use temperature differential to make a strong fastner. For example for a machine I am making I need to transition from 5/8" rod to 8mm. The machine will always be around 65 degrees ambient. Rather than machine the 5/8" down, could I drill a slightly undersize 8mm hole, freeze the 5/8 rod and heat 8mm drill rod before inserting? If so how would I figure the undersize?


01-23-2005, 04:48 PM
We have a couple of machines (high speed) that use strictly heat shrink arbors. No set screws needed. I have not heard of a single cutter pushing up or comming loose. So yes, it holds well.

But I think you have your fleezing/heating backward. The 5/8" with the slighly undersized hole will need to be heated to expand the hole. Freezing the 8mm rod should not be needed unless the undersized hole is drilled too small.

I'm not sure about how far undersized you should go. Guess it depends on the application. Go too small, and the 5/8" rod will probably deform a little around the drill (actually making it slightly larger than 5/8"), or the 8mm rod might not go in at all. Go too big, and of course the 8mm rod will be loose. My best guestimate would be just .0015-.003 undersize. Again, if it needs to be a real tight fit, you may go a bit more under 8mm.

It also depends on the type of materials being used. As different materials shink/expand at different rates when heated/cooled.

01-23-2005, 05:04 PM
It's in the machinest hand book. I don;t have my copy with me so you'll have to wait till Monday for what section it's in. Unless someone else looks it up for ya.

01-24-2005, 12:11 PM
In my experience you will be lucky to get that shaft to expand more than about .0005.
I would make the hole about 8 tenths under and heat it, you will also need a press (or something) to put it together.

01-24-2005, 01:30 PM
I guarantee, if you can get the 5/8" rod into a 8 mm hole, it will never, ever come out :D

01-24-2005, 05:34 PM
I think I'll abandion the idea, based on what I am hearing and have found, looks like it's to precise for me to produce. .0005 under is a little more than I can handle.

Thanks anyway,

01-25-2005, 02:34 AM
Unless you find an easier way, have you thought about a cold 'snug' fit, so it lines up reasonably straight, and then either silver solder or braze the joint??

If the fit is good, alignment problems should be minimal to nil..

01-25-2005, 11:04 AM
Hi pimino, It's not that precise. Blacksmiths used it before powertools where around. Give it a try, it does work. At 'home' it's usually a permanent fixing, by the way, There is high speed machine tooling that uses shrink fit to mount tool bits.

You're talking about obtaining a tight fit between two parts by expanding the 5/8" holder with heat and shrinking the mating 8mm part to be fitted with cold, then assembling the two. The expansion and shrinking allows enough clearance to assemble the part, when the parts return to room temperature the 5/8th rod with hole will have shrunk around the 8mm part.

If you drill the hole smaller than 8mm by .5 - 15 thou it would probably work. It doesn't have to be accurate, it does depend on how much you can expand or shrink the parts and that depends on how hot or cold you can get them, so its trial or error. If you drill a hole that's smaller than 8mm by .5 - 5 thou it would probably work. Note the 5/8th part will expand more than the 8mm part would shrink.

You'd need to put a lot of heat on the 5/8" part to get any movement on the 8mm hole though, and the 8mm part would not measurably shrink if you put it in the freezer. Try drilling the sub 8mm hole and freeze the 8mm and put a torch on the 5/8" until its *hot* - up to colour even and try a fit.

let us know how it goes..


01-25-2005, 03:02 PM
For steel, its strictly 6 millionths of an inch expansion per inch of material, per degree Fahrenheit. Consider also that you need at least .001 clearance to be able to slip it together by hand. So let's see..
8mm = .31496"
.625 - .31496 = .31004 expansion required.

So that means at 51,840°F, the hole should be large enough.....just put the rod in the center of the plasma where your hole used to be, and let it cool down :D

01-25-2005, 05:21 PM
???? the first post says he's fitting an 8mm rod into a hole drilled in a 5/8" rod so the expansion required is a few thou' depending on how big the drilled hole is?

A 5/16th drill might be a touch too big (having an interference fit of ~0.0025) but if he can drill it then ambient + 350f might do the trick.

01-26-2005, 10:37 AM
Hu is just being a smart ass.
I will still say the expansion and contraction you will get between these parts wont be more than 1/2thou. but if you have a .001 difference a press will still force the pin in the hole-shouldnt take more than 5 ton.

02-09-2005, 04:36 PM
What you are talking about is called an interference or shrink fit and the general rule of thumb is .001" per inch. For example, a bore diameter of 2.5" would require a shaft size of 2.5025" or .0025" interference. But for what you want to do that may be too much. It's usual for smaller diameters to go for a "size for size" fit or just a few ten thousandths interference. This does not mean that the pieces will just slide together, if they are size for size they will still require light pressing. Also, it's important you have a good surface finish. A reamed hole would be better than a drilled hole.