Half-shaft twist

[Dunk Mckay]

When working out diameters for half shafts, they can be calculated for a maximum shear stress in the shaft or for a maximum allowable angle of twist. Whichever one is likely to be reached first is the one that defines the diameters of the half-shafts.

What I'm wondering is if anyone has any info/data as to what a reasonable/recommended amount of twist is? A lecturer recommended something on the scale of 0.001rad (0.57deg) but that gives a diameter of 102mm which seems a little excessive.

[Dunk Mckay]

When working out diameters for half shafts, they can be calculated for a maximum shear stress in the shaft or for a maximum allowable angle of twist. Whichever one is likely to be reached first is the one that defines the diameters of the half-shafts.

What I'm wondering is if anyone has any info/data as to what a reasonable/recommended amount of twist is? A lecturer recommended something on the scale of 0.001rad (0.57deg) but that gives a diameter of 102mm which seems a little excessive.

[HenningO]

I am going to answer this with 2 questions.

1. What will happen (to the car and/or halfshaft) when the twist angle exceeds 0.57 deg?

2. What will happen (to the car and/or halfshaft) when the shear stress exceeds the material's ultimate shear stress?

You might also want to ask your lecturer how he came to the number 0.57 deg.

[Dunk Mckay]

Was 0.057deg sorry, but he gave me 0.001rad as a rough idea, as he didn't know any exact numbers (we have no drivetrain specialists here).

1. The maximum twist angle is used to limit the twist and deformation in the shaft, throwing of the centre of inertia, which can lead to vibrations, and through that failure of the vehicle. But I have no idea at what angle this starts to become a problem.

2. When the stress exceeds the material's ultimate shear stress the half-shaft is going to experience catastrophic failure.

[Tom W]

The third variable to consider is the loads applied back to the gearbox and engine. A shaft with that high a torsional stiffness will apply enormous shock loads to the gearbox and engine.

[BillCobb]

It probably won't be the shaft that fails, anyway, it will be the joint. (Maybe that's 420 already happened)....

[Bemo]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Tom W:
The third variable to consider is the loads applied back to the gearbox and engine. A shaft with that high a torsional stiffness will apply enormous shock loads to the gearbox and engine. </div></BLOCKQUOTE>

I absolutly agree. If your half shafts aren't too stiff, they work like a tortional spring and help to smoothen shock loads.
If you have a look at other cars you will see that diameters of steel half shafts aren't very big.
I never heard of anyone getting a problem with vibrations or something because of too much twist in the half shafts. Care about shear stress.
There are other parts on the car for which the max. displacement much more critical like suspension parts, differential mounting, frame...

And by the way, the twist angle of your half shafts is linear dependend on their length. So if you get high twist angles think about what will happen if the length of left and right half shaft is too different.

[Dunk Mckay]

Yeah, I've been looking at that. Trying to get more or less the same angle of twist on either side shouldn't be hard though. At less than 400mm each with maybe 20-30mm variance between the two, even for equal diameters the variance shouldn't be too great. We've got quite a narrow wheel base this year, with the widest Drexler model. Our main issue is probably gonna be joint angles. Having said that the rear of the chassis is being redesigned because they didn't really think about driveshaft paths. Oh well, hopefully we'll freeze the chassis soon so i can get some work done.

[JasperC]

I think that indeed the difference in twist between left and right will be of negligible importance (who cares if the left wheel rotates 1° further than the right wheel when you go from 0 to maximum torque?).

Something you might want to check, though, is how the different stiffnesses would affect the load distribution in a bump case (full throttle, hence locked differential, off a small bump). The shorter (stiffer) halfshaft would carry more torque than the longer one so you are more likely to get failures on that side.

This is especially significant if you're using a relatively stiff carbon halfshaft in combination with a relatively flexible metal stub axle, where a difference in stub axle length will have a large impact on the overall stiffness difference. Last year in our original halfshaft design we had a difference of 30mm in stub axle length left to right. One of our alumni spotted this, I calculated the bump case and found that the load distribution would be something like 62/38 so we had to modify the design slightly to stiffen the longer stub axle.

Maybe for full metal halfshafts this isn't really important, I haven't calculated that.

Best regards,
Jasper Coosemans

Chief Drivetrain 2009-2010
DUT Racing Team
Delft University of Technology

[Bemo]

Have a look at cars which break out to one side in the acceleration event. In most cases you will find steel half shafts of very different length. It's no problem if the difference is only about 20-30 mm, but if one is double the length of the other it's becoming a problem.

The angle of the shafts is a serious problem. If it is too high you will get a big problem with your joints. I think the tripods from Taylor are made for up to 7 deg, but if you try to use them in this max angle they won't last very long. Also design judges won't like a design like that.
In standstill the half shafts should go more or less strait and the joints should only have an angle while bumping.

[Dunk Mckay]

Yup, that's what I keep telling my chassis manager, I should know by Monday, fingers crossed eh? Think the GKN tripodes we used last year run up to 26 deg, but optimum is only up to 8. Looking into GKN countertracks instead of standard CV's, they look interesting, no plunging models as far as i can tell but lighter and more efficient would be a bonus.

[nowhere fast]

I agree with the point about shaft twist reducing shock loads on the gearbox, this is part of the reason motorcycles have a rubber cush drive between the rear sprocket and the wheel. Shock loadings can arise from the slack in the chain since the chain will not transmit drive until the slack is taken up.

If there is a difference in driveline stiffness of each wheel then initially higher loads will be applied to the stiffer side when torque is applied and the car will steer to one side slightly when torque is applied although this may be negligible (I haven’t noticed it when driving our car).

Large differences in the driveline stiffness of each wheel are common on cars that use a spool even with equal length half shafts. This is because drive to the right wheel passes through the spool and the half shaft whilst the drive to left wheel is passed through the half shaft only.
On our car the left side was approximately twice as stiff as the right, I did design a stiffer spool and different half shafts for left and right so that the stiffness was equal but they were never manufactured.

[JasperC]

That's interesting, nowhere fast. Would be interesting to hear if you have actually suffered more wear (or even failure) in the left part of your driveline.

We also have the same experience with a slacking chain causing problems (a nice shock load for the entire driveline every time you push the throttle after cornering). Which is why you really need a chain tensioner which is easy and quick to adjust. But that's an entirely different topic. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Cheers,
Jasper Coosemans
Chief Drivetrain 2009-2010
DUT Racing Team

[Mikey Antonakakis]

One of these days I'm going to actually try and calculate how big of an effect unequal-length halfshafts have compared to, say, a less than perfect alignment, unequal L/R weight distribution, steering wheel angle, etc...