How Much Heat ?

If your intent is perfect concentricity, isn't a flat the last thing you want to do?

TANGENT ALERT:

Response to tangent: A small flat, not the full wall length of the engagement. The flat is to provide purchase for a set screw to prevent any "reasonable" chance of twisting like the tiny set screw size flat on the side of an end mill. There would be full circular engagement above and below the tiny little flat.

A quick search shows that generically the coefficient of expansion "steel" is about 0.0000072 inches per inch per degree F. So figuring worst case at .500 and figuring I need .002-.003 clearance to be able to finish the project before it cools and seizes up I should figure on needing .004 expansion per .5 inches.

.0000072*.5=.0000036.

.004/.0000036= 1111.11 degrees. This doesn't sound practical for the tools I have on hand for my application.

.003/.0000036=833.33 degrees. Maybe doable, but uncomfortabley close. Time may be an issue.

.002/.0000036=555.55 degrees. Certainly within the ability to reach temperature, but tolerance will be close during fit up, and time may be to short to complete fit up. Because of the close tolerance the shaft may rapidly cool the part.

TANGENT THOUGHT. The amount of energy required to get the growth to make the fit up suggests that the set screw may be totally superfluous.

Back on topic... a sample part and shaft to test with may be a wise investment of time. Something cut with hexes so it can be clamped up and torque of the joint / joint slip tested afterwards.

IIRC, shrink fits are generally .003" for the first inch of diameter, .001" per inch larger.

So, you might bore your 1-inch part's hole .4965" for your .4995" shafting.

Slowly and evenly heat the part with the hole to a dullish red. Don't let it develop any scale.

Insert the cool shaft in one swift motion, you won't have time to mess with your set screw alignment. The shaft instantly starts to expand with the heat, further motion will not likely be an option.

Then cool it all off with water.

Your set screw will probably be redundant.

Good luck.

Paul, Thanks. I did a little research and a little math in one of the other posts. I did not know the guideline for size difference, but I did calculate that it takes quite a lot of energy to expand the metal sufficiently to get a fit up, and guessed that the screw might not be needed.

My only real issue then is getting a good bore of .4965 with the tools I have. Reaming to .499 would be easy, but I have to bore to any off size.

Let's see. Steel coefficient of expansion is .0000072" per inch per degree Fahrenheit above ambient. Later post says .003 needed. .003/.0000072=416 degrees. Bore is .5 inch, so double that, 832 degrees. Plus 70 degrees ambient, about 900 degrees. Clean part so no oil scales up in bore. Heat slowly, probably on a fire brick, so it doesn't lose heat so fast when removing flame. Squeeze bottle with water will cool it down fast, without shocking it.

You are working with almost the same dimensions and materials as a connecting rod with a pressed pin.

Watch your heating method, you want the heat even and slow to prevent thermal distortion. Make sure the hole is perfectly clean and the shaft is perfectly clean. Anything that causes a snag will screw you up. As a twist inhibitor you could put striations on the area of the shaft where the part will stop. Just make sure they are below the surface so they don't inhibit the assembly. This is done on some pins that have oiling provisions to prevent rotation.

In the case of the pins (which are hollow) you have around 5-6 seconds before the parts are locked. With the solid shaft you might get a bit more but I wouldn't count on it.

Do NOT use any forced cooling, let them air cool slowly. Any type of quench will cause thermal distortion and probably thermal shock enough to damage the part.

"Steve W."

I agree with Steve. Don't shock it with a quench! It's not needed... pdk

If you are willing to experiment you could compare the hot fit to pressing the parts together cold. jsw

If I was doing your assembly I would heat the one part and freeze the other just to get a little more clearance. Then I would carefully run a drill inside the tapped hole to dimple the shaft. Finally, I would install the setscrew. I think the screw would be redundant but it won't hurt to put one in. And a cup point setscrew, or the type with the cup point that's knurled, will locate well in the little dimple. And the dimple will be perfectly aligned with the screw hole. Eric

Freeze the shaft and heat the part. The coefficient of thermal expansion for steel is 0.00000645in/in/deg F. You have 1/2 inch to work with, so .000003225" per F degree temp difference.. If you measure at 70F, and freeze the shaft to 0f and heat the part to

For a small part like this I'd consider heating the collar on a steel or copper block with a hole drilled in it for the shaft. Mabee a recess to hold the collar, - hole oversized and drilled to required depth to precision locate the shaft on insertion. How much load will this part be under? I very much doubt you will require a 3 thou interference fit for this installation. The block will hold heat to keep the temperature of the collar up. The collar has less mass than the shaft. Heat the collar and block - freeze the shaft, insert the shaft and rap it into the hole with one good smack of a mallet and "Bob's your mother's brother"

Don't know if this will be considered a tangent or not . Is a heated shrink fit a requirement?? Compleat the job as per the first three sentences above, then apply some LockTite bearing retainer. That plus your set screw should provide all the retention needed. rgentryatozdotnet

Sounds reasonable.

Easiest to avoid if you can heat it in an inert atmosphere -- Nitrogen or Argon for example.

If you want a bit more clearance to work with, get some liquid nitrogen, and *cool* the shaft while you heat the collar (or whatever kind of part -- I'll use collar for simplicity). You can get the Liquid Nitrogen from a welding gas supply place usually. Take a big thermos if you don't have a Dewar made for the task. Ideally, one of the fragile glass-lined thermos jars. If not that, several large styrofoam cups nested, and handle with good gloves. Drive with the windows open so you don't pass out from the nitrogen which boils off displacing the oxygen in the car.

The Liquid Nitrogen is at 77 K (-231.07 F), so that will add a bit more difference in size.

Can you place the collar on a support with a hole sized to pass the shaft? If so, mount another ring above it a loose fit on the shaft at room temperature, and a few inches above the collar. This will let you drop the shaft in with a pretty good alignment, and the support will be at the same temperature as the collar, so it will keep that warm a few seconds longer. Set a stop below the support so the shaft goes the right distance into the collar. Then heat the one, cool the other, drop in and stand back.

Agreed. That will be a serious interference fit. If it isn't, the assembly can always be drilled for a taper pin to lock it.

Enjoy, DoN.

Possibilities:

However, all three require a way to *precisely* measure the bore. Nicest is one of the the tri-mikes of the right range. Lacking that, a shaft with a minimal taper (e.g. 0.001/inch) and slide it in until it stops. Mark at the edge of the hole with a felt-tip pen, pull it out, and measure with a tenths-reading micrometer. (Likely more repeatable than a small-hole gauge, which requires a really good feel.

Or -- perhaps set up a toolpost grinder on the lathe (with proper protection for all sliding surfaces). Set the compound as close as you can to 5.71 degrees (nearly parallel to the workpiece axis) and this will allow you to move in 0.0001" for every 0.001" you turn on the compound feed dial (assuming that the dial is direct reading -- *beware* some compounds read half of actual motion so you know how much diameter you will take off instead of how much radius). It makes more sense for the cross-feed to be so calibrated, but sometimes it is also done on the compound where it only makes sense when parallel to the cross-feed. :-)

Good Luck, DoN.

On Wednesday, January 15, 2014 12:21:42 PM UTC-5, Bob La Londe wrote: Then I was wondering if I might get a better

You might consider this as going off on a tangent, but to get the best concentricity I would join the rough machined parts and then machine the part.

Dan

It's all a matter of degree. If your primary objective is concentricity, shouldn't any deviations from circular be radially symmetrical both dimensionally and for applied force/setscrews?

I suggest that setscrews work by deforming something elastic.

One man's tangent is another man's failure to consider the consequences. Sometimes, the cure is worse than the disease. It's all a matter of degree.

Regardless, it did not have any relevance to the actual question asked.