Shrink Fit Failure 1

[CodeRed7]

I am currently designing an assembly that requires a sleeve to be shrink fitted onto a potentiometer shaft. I am having issues with analyzing the stresses caused by the fit. I have tried using an online calculator, however no matter what I do I always seem to have stress levels that are too high. I have summarized the information below.

Material

Hub Material: 360 Brass Alloy
Shaft Material: 303 Stainless Steel​

Dimensions

Shaft Dimensions:

Maximum Shaft Diameter: 0.1248"
Minimum Shaft Diameter: 0.1245"
Length of Shaft: 0.313"​

Hub Dimensions:

Maximum Inside Diameter: 0.1244"
Minimum Inside Diameter: 0.1242"
Outside Diameter: 0.25"
Length of Hub: 0.4375"​

Fit Specifications:

Type of Fit: Light Drive Fit [FN1]
Maximum Interference: 0.0006"
Minimum Interference: 0.0001"
Engagement Length: 0.1309"​

If I need to specify any more information, let me know. Any help would be appreciated, thank you.

 

[desertfox]

Hi CodeRed7

If possible I would increase the interference provided the nylon stresses don't get to high and doing that would give you more margin for error.

desertfox

 

[CodeRed7]

Hi DesertFox,

Allright that makes sense, there are standard fits (with standard interferences) that are posted on the ANSI guidelines. Should I follow those standard fits or would I be able to design my own?

 

[desertfox]

hi CodeRed7

No reason why you can't do a fit of your own.

 

[CodeRed7]

DesertFox,

Thanks again for everything. Right now I'm looking into the shaft retaining compound that 3D Dave mentioned as it removes the need for the extra sleeve. If this does not turn out, I will go back to the shrink fit with the sleeve.

 

[desertfox]

Hi

your welcome!

 

[rb1957]

1) what's wrong with a press fit yielding one of the elements ? (it seems this is why brass was de-selected) actually, isn't this a good thing 9'cause it'll increase the torque that the joint can react (ok, torque isn't a big design element here).

2) i don't know your application, but be careful choosing materials. i would caution changing from brass to plastic to suit one element of the design. why did you pick brass to star with ??

Quando Omni Flunkus Moritati

 

[tbuelna]

Plastics behave differently than metals. You will not be able to maintain a "shrink fit" between a plastic component and a metal component over time. The plastic component, whether it's nylon or polycarbonate, will creep and the interference fit will be lost. If you want to make a reliable connection between a plastic part and a metal part, you must use some type of adhesive bond.

 

[MikeHalloran]

There are a few issues not covered so far:

If you use an adhesive to bond shaft and hub, you may have problems with the adhesive wicking into the potentiometer's internal bearing.

If you use a press fit to bond shaft and hub, you may dislodge the tiny snap ring adjacent the potentiometer's internal bearing and damage the potentiometer.

Both potential problems go away if you buy the potentiometer with a flatted shaft and push on a plastic hub with a molded D hole. That also allows you to mechanically index the plastic hub's features to the potentiometer's wiper with manufacturing controlled geometry, and avoid the need to manually angularly adjust the wiper position to that of the sensed mechanism during assembly.





Mike Halloran
Pembroke Pines, FL, USA

 

[CodeRed7]

Hi All,

Rb1957
1) It notes on the website and I have heard in other discussions to compare the stresses that result from the calculations to the yield strength. My assumption is that since the material is trying to fully return to its original shape, this causes the force required to maintain the fit. If the material is yielded, the force caused by the interference will be decreased since the material is not trying to fully return to its original shape. Either way, the stresses that I found when using brass exceeded its ultimate tensile strength, so the material would fail either way.

2) It choose brass since multiple other components in the assembly are made of brass, but these will be under larger torque requirements (100 oz-in), whereas this fit will be under much smaller (around 0.7 oz-in).


Tbuelna
Thank you, I will keep that in mind, I will most likely go with using the adhesive anyways.

MikeHallorn
Unfortunately, I have already ordered the potentiometer. Either way, I don't think I was able to find a potentiometer that met my specifications that had a flatted shaft. In addition, I couldn't find a stock gear with a D hub, this would most likely have to be a custom gear would be more expensive. Would you know of any way to avoid the adhesive wicking into the potentiometer's internal bearing?

 

[CodeRed7]

Hey guys,

I was wondering, are there any issues with using a clamp with a clamp hub gear? This seems to be a feasible option that has not been mentioned yet, and was wondering what the limitations of this option were.

Thanks.

 

[MikeHalloran]

That should work.
Be sure to leave the clamp accessible for adjustment.
... and to slot the end of the pot shaft for a screwdriver, or let it project beyond the clamp so you can grip it for adjustment.


Mike Halloran
Pembroke Pines, FL, USA

 

[CodeRed7]

I can leave the clamp accessible for adjustment, however most of the clamp hubs I can find are longer than the potentiometer shaft, and the shaft isn't slotted. Are there any other solutions? Just to clarify, I'm using a gear with a clamp hub, a potentiometer, and a separate clamp. If you want to look at the products I have listed them below.


Potentiometer:

http://www.bourns.com/data/global/pdfs/3549.pdf

Gear:

http://sdp-si.com/ss/pdf/79001063.PDF

Clamp :

http://sdp-si.com/ss/pdf/81001242.pdf

(Model No. S3501Y-C065)

I was wondering, why exactly do I need to able to adjust the clamp? My application is permanent, so I would not require disassembly. Would the reason be that it can become loose after a certain amount of time?

 

[MikeHalloran]

If I haven't misunderstood something, you're using the potentiometer in a control system, to measure the position of a valve.

At some point during assembly, someone has to physically move the valve to one extreme of its motion, and move the potentiometer to the corresponding extreme of its motion, and tighten the clamp, so that from then on, a particular position of the potentiometer's wiper corresponds to the analogous position of the valve's rotating parts.

Potentiometers typically ship with a rather long shaft, that may even be annealed for easier machining. It's fairly common to clamp the distal end of the shaft in a milling machine, mill a flat on the shaft without applying any force to the potentiometer itself, then cut the shaft to a desired length, and optionally slot the shaft. In that way, the machinist's setup process defines a controlled relation between the flat and the pot wiper. Said relationship will be documented in a 'make from' drawing. Once that relationship is established, further need for adjustment can be eliminated with D holes, and you don't have to pay servo techs for simple assembly operations.

You have chosen to eliminate the shaft machining process step, so the angular relationship between valve and pot wiper must be established at assembly time. Hence the clamp. A tiny drop of green Loctite on the clamp screws after adjustment would be a good idea.


Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

If this simply connecting the input shaft of a rotary pot to a valve shaft, then you probably want some sort of coupling that will isolate any forces being transmitted other than pure torsion. No bending or thrust. There are shaft couplings designed precisely for this purpose. Most rotary pots are not designed to accommodate any radial or thrust loads, so it probably not be a good idea to use a rigid shaft coupling.

 

[CodeRed7]

MikeHalloran
Yes, the application is correct, but would I not be able to adjust both the valve and potentiometer before clamping in order to achieve the angular relationship, and then apply the clamp, while being careful to not move either the valve or potentiometer? Would I still need to slot the pot shaft or have it project beyond the clamp?

Tbuelna
I forgot to mention a motor will be attached that will drive the system. The valve shaft will take most of the load, while the potentiometer will experience very little. Would I still need to look for a different shaft coupling?

 

[MikeHalloran]

We can't properly comment on particular features of your design, because we have seen exactly none of it.

We can suggest, however, that some sort of coupling (or flexible anchor) of a potentiometer might be a good idea, because their bearings are usually, well, trivial, and the tolerances and alignments in a typical valve actuator are not precise enough to not damage a 1/8" shaft with any sort of rigid connection.

Speaking of bearings, there really aren't any in a typical valve. If you are relying on the valve shaft to resist gear separating forces, and the valve manufacturer has not blessed your application, you may face serious warranty costs.



Mike Halloran
Pembroke Pines, FL, USA

 

[CodeRed7]

Hi Mike,

I have attached the basic design of my assembly. In my assembly, a coupler connects the motor and the valve, which transfers most of the torque. Attached to the coupler is a gear, which connects to another gear and then to a potentiometer. My idea for assembly would be similar to mentioned before. The potentiometer and the motor would initially be mounted. One of the gears would be press fit onto the motor valve coupler over the knurl, and then the coupler (with the gear attached) would be press fit onto the motor shaft (also over the knurl). The potentiometer would be manually adjusted to its central angular position, and then a clamp would be placed over the hub of the gear. This gear would be moved in alignment with the other gear and the potentiometer shaft, and then clamped. Finally, the valve will be adjusted to its central angular position, and the valve shaft will be slid into the motor valve coupler and mounted. This seems like a feasible idea, am I missing anything that would require slotting the pot shaft or having it extend beyond the clamp?

Also, are the problems you mentioned still apparent in this assembly? Would the fact that the potentiometer is not directly connected to the valve shaft maybe solve some of these issues?

 

[MikeHalloran]

Knurls are for providing traction on hand contact surfaces.
They are not real good for mounting gears,
nor for transferring torque.

Your structure is commendably minimalist, but it's rather delicate. If it's possible to bend something with bare hands so that, e.g., the gears dont't mesh right, some end user or random twiddler will do that, and you'll pay for a warranty call.



Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

Looking at your sketch, the main thing I would recommend is making sure your gear mesh has adequate backlash to prevent binding. Mesh binding would produce excessive radial loads on the pot shaft. Based on what I can see from the sketch, it looks like your mounting arrangement and bracket may not provide optimum shaft alignment and stiffness.

 

[CodeRed7]

Mike
I can also use the clamp to attach the coupler to the gear, so I will look into that option. And understandable, this is just the basic design for now, I will make sure that the assembly cannot be accessed by any random user.

Tbuelna
All right, I will look into backlash in order to reduce the radial loads, thank you.