shrink fit issue between gear and shaft

[god_paul]

Hi, everyone, have a question about shrink fit process...
Here are basic dimension At room temperature:
Shaft: 42.189 - 42.2 mm
Gear bore: 42.139 - 42.15 mm
Gear outer diameter: ~ 64 mm
Gear height: 75 mm

currently, we cool the shaft to -40C, and heat up the gear to 150C
cooling and heating are done approximately at the same time, (we just let the gear keep being baked in induction coil and wait till cooling temperature reaches the set -40C)

After that, a small robot arm grabs the gear and put onto shaft.

The process works but we have 20% of failure rate, mainly happen when the gear is stuck on the shaft just a couple of millmeter away from the final installation position when the robot arm is guiding the gear down.

I am no expert in heating/cooling process but what could be the causes? is the speed of dropping the gear onto shaft too slow? Some people mentioned that the shrink fit process needs to be speed of light? is the heat transfer from the gear to shaft really fast to a point that the gear bore size starts to shrink fast during the installation process?

Thank you

 

[DDoug]

Your not ramming it down with a press and then holding it in the press ?

 

[god_paul]

Your not ramming it down with a press and then holding it in the press ?

No, and Yes,
No, we use robot arm to drive it down to the final stop position
Yes, after that, a press head comes down to hold the gear in place for a couple of seconds (basically a fake press)

 

[DDoug]

No, and Yes,
No, we use robot arm to drive it down to the final stop position
Yes, after that, a press head comes down to hold the gear in place for a couple of seconds (basically a fake press)

So either way, it's getting stuck on the way down, and you don't have enough speed & force to get it home every time.

 

[god_paul]

How fast should this process go?
basically the speed of movement of robot arms or any mechanical components to move the gear down fastest it can ?

also. The entire top of the shaft is cooled , is this necessary? Because the moment the gear starts to drop onto shaft, heat transfer occurs right away… I am thinking if we only need to cool the lower portion of the shaft only?

 

[MCritchley]

The robot doing the assembly is tough. Does the gear rub the shaft on the way down? I do shrink fits often, i go fast! My rule is .001" per inch of interference per inch of dia.

Can you heat the gear up more without changing the temper? 150C is clearly not enough.

 

[guythatbrews]

For a steel shaft and gear starting at room temperature you've got about .0013" clearance at assembly. Not much. Be sure the gear is exactly perpendicular to the shaft. I suspect the gear is cocking.

 

[god_paul]

The robot doing the assembly is tough. Does the gear rub the shaft on the way down? I do shrink fits often, i go fast! My rule is .001" per inch of interference per inch of dia.

Can you heat the gear up more without changing the temper? 150C is clearly not enough.

MCritchley, would you elaborate your rule a little bit? what is 001" per inch of interference per inch of dia. any simple example?​

also, what happens if the gear is rubbing with the shaft when it is down? does this facilitate the gear to get stick to shaft? like our tongue will stick to ice, similar concept?

Thank you

 

[MCritchley]

MCritchley, would you elaborate your rule a little bit? what is 001" per inch of interference per inch of dia. any simple example?​

also, what happens if the gear is rubbing with the shaft when it is down? does this facilitate the gear to get stick to shaft? like our tongue will stick to ice, similar concept?

Thank you

Sure, for a 3.5" dia shaft i would make the bore of lets say a gear .0035" smaller than the shaft. And a half inch shaft i would have at the maximum .0005" interference between shaft and bore.
Heat to 400C and drop the gear on the shaft.

 

[guythatbrews]

Sure, for a 3.5" dia shaft i would make the bore of lets say a gear .0035" smaller than the shaft. And a half inch shaft i would have at the maximum .0005" interference between shaft and bore.
Heat to 400C and drop the gear on the shaft.

That's about .0115" clearance at 400C. Easy.

@god_paul can you increase your temperature? Any frost at all on the shaft?

 

[Erich]

You have two parts that are 190 C degrees apart. Both are exchanging heat to air via convection. A fairly slow heat transfer mechanism. If the shaft touches the gear they exchange heat with each other via conduction. A MUCCCCCCCCCCCCCCCH faster mechanism. Shaft grows, gear shrinks, and they bind up.
You could dunk the shaft in a tank of dry ice and alcohol. That could take your shaft down to -90C. So you get another 21% temp differential.

 

[god_paul]

Sure, for a 3.5" dia shaft i would make the bore of lets say a gear .0035" smaller than the shaft. And a half inch shaft i would have at the maximum .0005" interference between shaft and bore.
Heat to 400C and drop the gear on the shaft.

Sadly. This is blue print dimension so I cannot do that much …
but the heating temperature is also called out on the print 155C max , i don’t know how designer comes up this number but I can attempt to argue to get a higher temperasure zone

 

[god_paul]

That's about .0115" clearance at 400C. Easy.

@god_paul can you increase your temperature? Any frost at all on the shaft?

I can request to increase the temperature and yes the shaft has defrost process

 

[god_paul]

You have two parts that are 190 C degrees apart. Both are exchanging heat to air via convection. A fairly slow heat transfer mechanism. If the shaft touches the gear they exchange heat with each other via conduction. A MUCCCCCCCCCCCCCCCH faster mechanism. Shaft grows, gear shrinks, and they bind up.
You could dunk the shaft in a tank of dry ice and alcohol. That could take your shaft down to -90C. So you get another 21% temp differential.

For gear up to what temperature will the interior crystal property to change ? Will a too high temperature distort gear teeth?

same question for shaft , will shaft become brittle / permanently damage shaft at so low temperature?

 

[guythatbrews]

Have you checked perpendicularity gear to shaft? Pretty much any tilt at all and it's sure to jam up.

 

[Rickyb]

Sadly. This is blue print dimension so I cannot do that much …
but the heating temperature is also called out on the print 155C max , i don’t know how designer comes up this number but I can attempt to argue to get a higher temperature zone

The 155°C spec indirectly says that the original tempering temp of the gear was ~175°C. You need to be 10°C(50°F) below the original tempering temp so you don’t do any further tempering. Your option to heat the gear more does not exist so your efforts can only be directed to the shaft and pressing options

 

[guythatbrews]

The 155°C spec indirectly says that the original tempering temp of the gear was ~175°C. You need to be 10°C(50°F) below the original tempering temp so you don’t do any further tempering. Your option to heat the gear more does not exist so your efforts can only be directed to the shaft and pressing options

How do you know the tempering temp was 175? Just a SWAG?

 

[john.k]

Done lots of high interference shrink fits ......you need a second robot with a sledge hammer and a drift...........Can you use liquid nitrogen for the shaft?............The Ford truck gearboxes were fitted like this ......shrink fit ,no keys .

 

[EmGo]

Sounds like the gear is a ways down the shaft. A drawing is nice, you know. Picture=10,000 words and all that.

Normally, if the fit is way down on a shaft, you'd like to have the length before it fits as a reduced area, so it doesn't touch until it gets to the shrink fit area. If they can't reduce the long part then an option is to increase the fitted area to maybe ten or twenty bigger. Then there's no problem with heat transfer before you get to the fit.

A nice taper on the entry is good too, makes sure everything is not cocked as it goes on.

If it's right at the beginning, then you need more temp differential. Better to cool the shaft more than overheat the part.

will shaft become brittle / permanently damage shaft at so low temperature?

Not if it's anything normal. 'Cryogenic freezing' is an accepted method of stress relief that you get to pay extra for, if you want it. 9310 needs to be very coldfrozen after heat treat, in fact.

 

[MCritchley]

At a glance I think your process is not optimal if you are not allowed to heat the gear up to a much higher temp.

For starters the robot arm should be some type of fast acting press that holds the parts concentric. I would want to keep the heated gear stationary and ram the shaft in fast. Perhaps a pneumatic press, that’s just a random thought as pneumatic is fast and can hold the thing down during cooling.

Guythatbrew’s calculation of .0013 hot clearance doesn’t lend a lot of hope to this working.

Something that I would try is to make a gear with exactly .0013 clearance and attempt to slide the room temp. assembly together with the robot like 1000 times. If it works without fault 1000 times you might have a chance introducing the hot and cold parts.

Good luck mate!