Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Assemble shrink fit with a big shaft ∅203 mm?

  1. Dec 8, 2013 #1
    Assemble "shrink fit" with a big shaft ∅203 mm?

    Dear all,

    Just a short question that is it possible to contract the shaft in liquid nitrogen and heat up the hub around 300°C so that we can assemble them together quickly? We are delivering a prototype to our customer with a tight requirement on interference fit from drawing of them as in attachments.

    At the first time we calculated the contraction of steel per degree of linear expansion/contraction and think that it's ok and we just freezed the shaft and then got stuck about 1/3 (100 mm fit), after trying to disassemble but couldn't then we cut out the hub to keep the shaft for reason of material in stock :(

    Now is the last chance and we intend to freeze shaft and heat up hub as above but do we really need to freeze the shaft like this, or the contraction of steel will not work in this type of shaft because we did not have suitable tools for measuring accurately in a very quick time..

    Thanks very much for any advices

    Attached Files:

  2. jcsd
  3. Dec 9, 2013 #2


    User Avatar
    Science Advisor

    Once you know the crush required after thermal equilibration you can calculate the minimum temperature difference needed at the last moment of sliding assembly. The temperature difference at the start will need to be significantly greater.

    If the components are not thermally soaked evenly they can take on distortions or extreme axial curvatures that will prevent assembly by gravity or by hand. This is especially true of long bores and shafts such as gun barrels.

    You do not need to measure diameters when heated or cooled. It should be enough to know the temperature of the parts. That is what a non-contact IR thermometer is for.

    Alignment during assembly is critical. There is no going back. It is good to have an assembly jig that aligns the components during assembly to avoid skew and so speed the assembly. It is also good to have a fast acting light press to bring them together once they are aligned in the jig.

    I would make a dummy component with a minimum clearance to test the assembly jig prior to operation with production parts. That gives repeated room temperature practice at the assembly without risk of destroying components when it goes wrong.

    Heating and cooling are both applicable to metals. You must however check the heat treatment profile that is acceptable for your particular materials. With some materials, cold can be as much of a liability as heat. How you achieve the differential temperature required is material dependent.
  4. Dec 12, 2013 #3


    User Avatar
    Science Advisor
    Gold Member

    This is an unusual design. I am curious why a keyway was not used.
  5. Dec 12, 2013 #4


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    The maximum (steel?) shaft diameter, 203.309 mm, should shrink to roughly 202.95 mm after coming to equilibrium with liquid nitrogen. It may have to soak for a while though because it's rather large. I suspect that was your initial problem.

    Edit: When I've done these kinds of things in the past, I have the parts ready to be pressed together using a hydraulic press just in case it begins to hang up as it goes together. It could be the parts simply get cocked a bit or there's some ice or dirt on the surfaces that make assembly problematic even when you would think they should easily slide together. Having a hydraulic press ready to shove the parts together is good insurance.
    Last edited: Dec 12, 2013
  6. Dec 12, 2013 #5


    User Avatar
    Science Advisor
    Gold Member

    I would be very concerned with runout tolerances over such a large fit surface. I didn't notice any in your diagram.
  7. Dec 13, 2013 #6
    Hi everyone,

    @Chronos: this is the free end of a winch drum then a keyway maybe not used, and the cylindricity as well as perpendicularity tolerances are 0.1 mm, both material grades are equivalent to EH36.

    @Q_Goest: You're absolutely right, at the first time we did not carry out sufficient cooling down the whole shaft with nitrogen then this was the problem.

    This time the shaft was fully boiled in liquid nitrogen nearly 1 hour and it gone through a gauge easily (we made an C-shaped from 5mm stainless steel) with upper limit tolerance 203.05 mm.
    And simultaneously carried out resistance heating the hub approximately 450 degrees that allow the T-shaped gauge with lower limit tolerance 203.65 mm gone through smoothly.
    Finally the shaft was assembled into the hub easily at the end of the day.
    But, tomorrow morning we found a region with gap 0.5 mm as illustration in attached pictures (not whole the circle), we not sure that heating affected or radius edge 1.5 mm on shaft and chamfer edge 3x3 mm caused the problem. By the way, the inspector passed it already.

    Next time maybe just cooling down the shaft is enough :)

    Again, really thanks for all of your knowledgeable advices

    Attached Files:

Share this great discussion with others via Reddit, Google+, Twitter, or Facebook