Maximum Torque for drive shaft design.

[Phil1988]

Hey guys,

Got some question due tothe driveshafts.

We calculated the weight transfer and looked into the tire data. With the Hossier 10" LCO we came up with about 650 Nm per Wheel witch can be transferred at about 80°C.

So my question is:

Are these Values real? Has anyone meassured the transferred torque at the driveshafts? How to regard on peak values like strokes? With with safety factor do you calculate?

Last season we designed with 450 Nm because we had no tire data.. The driveshafts are working pretty well....

Regards Phil

[onemaniac]

I doubt...
A typical, FSAE restricted, 600cc engine will put about 50Nm of torque at the crank. At 1st gear and with typical 3:1 final drive the torque at the differential, with 0 efficiency loss, will be around 800-900 Nm range (or 400-450 per drive shaft)
So, it should definitely be lower in real life.

Even with that kind of torque we see most FSAE cars manage to make some decent wheel spins at competition without launch control.
Take a note that the testing surface at Calspan has considerably higher mu than most parking lots or FSAE tracks.

If your team already has proven (for last year's car at least) that 450 Nm as target load was good enough, you should probably go lower this time unless your new car is considerably heavier than the last. Test, break, and change.

[Warpspeed]

If you paint a long stripe down each driveshaft, that can give you some indication if the shafts are gradually taking up a permanent twist. And be on the constant lookout for the start of fatigue cracks.

You are unlikely to just snap a brand new shaft outright. But if you continually give an under designed shaft a really hard time, it can deteriorate and often provide obvious clues in ways you can see, before it breaks.

[Racer-X]

I would not be concerned with how much power the motor puts out unless you are severely power limited. You need to consider how much "grip" the tires have and how much torque it takes to break them free. Most FSAE cars are already traction limited through all of first and most of second.

Now I think your values for tire "torque" are off. I know our car will easily spin both tires and it doesn't even make 900nm in first at peak torque. Or maybe you have some awesome tires...

[Z]

Hmmm... So is maximum driveshaft torque determined by maximum engine torque, or maximum tyre grip?

Thinking ... thinking ... when have I seen drivelines blow up ... ???

Errr ... well imagine your car with the back wheels OFF the ground (so tyre grip not a factor), with the clutch disengaged (so no engine torque going through driveline), and the engine revving at, say, 10,000 rpm. Now engage a gear, any gear, ... and ... "dump" the clutch.

Yep, quite often a sharp crack, and then relative silence, and smoothness (= no forward progress). Or maybe that rattling sound that always sounds expen$$$ive. http://fsae.com/groupee_common/emoticons/icon_frown.gif

Or try teaching your teenager to drive a manual, without doing "kangaroo" starts (so far not much difference between boys and girls http://fsae.com/groupee_common/emoticons/icon_smile.gif).

(Hint: Think accelerating frames and "inertia", NOT "steady-state"...)

Z

[Canuck Racing]

Originally posted by Z:
Hmmm... So is maximum driveshaft torque determined by maximum engine torque, or maximum tyre grip?

Thinking ... thinking ... when have I seen drivelines blow up ... ???

Errr ... well imagine your car with the back wheels OFF the ground (so tyre grip not a factor), with the clutch disengaged (so no engine torque going through driveline), and the engine revving at, say, 10,000 rpm. Now engage a gear, any gear, ... and ... "dump" the clutch.

Yep, quite often a sharp crack, and then relative silence, and smoothness (= no forward progress). Or maybe that rattling sound that always sounds expen$$$ive. http://fsae.com/groupee_common/emoticons/icon_frown.gif

Or try teaching your teenager to drive a manual, without doing "kangaroo" starts (so far not much difference between boys and girls http://fsae.com/groupee_common/emoticons/icon_smile.gif).

(Hint: Think accelerating frames and "inertia", NOT "steady-state"...)

Z

Z, once again your mindless ramblings have carried the topic off track just so you could grace us with your infinite wisdom. I am eternally grateful for this, but his concern is with the design of axles - nothing else.

We typically design axles with a factor of safety of 3 or 4. It's a little high but when you get into custom axles we've always found that at the FSAE level you run into manufacturing limitations before you can actually get axles that might fail from torque delivery.

You'll also find (typically way too late into the season, http://fsae.com/groupee_common/emoticons/icon_smile.gif ) that due to these manufacturing problems, stress risers become an issue of concern since most shops aren't really set up to do something of this size and there may be some trial and error in their machining.

Lastly, as everybody already said, your limiting factor will be torque from the tires (assuming you are able to spin them on take-off.)

I would recommend you start making your design and running some calcs to figure out what you could get away with, then figure out if you could actually make it. Your journey may just end right there.

[Canuck Racing]

Oh, I should mention that all that assumes steel (of some alloy) axles. I have yet to see carbon axles that work as well as steel as far as weight, packaging, cost and reliability - but that's for you to decide Phil.

[NickFavazzo]

Canuck, what Z was saying is look at inertia of your driveline parts and look at how that will effect the peak loads in your driveline. No need for personal attacks when they aren't warranted.

[MCoach]

It's important to look at what Z is saying rather than dismissing it in this point.

The force that the wheel is able to put out in Fx is dependent on the Fz. When a car lifts a wheel and then slams back down it it can actually impart a much higher load that what's capable than when static. So, it needs to be understood that a higher load can exist than what your static weight or even cornering load may be. A way to go about this would be to estimate how high you can lift a tire or if your car is capable of doing so during a corner and then find what the load is when that wheel is (assumed) accelerated back into the ground. Take your own measurements, look at a few videos from other cars and do some calcs. The tire spinning up brings inertial properties into this. The highest loads will be seen when the engine spins up the wheels and puts them back down violently. So, the velocity at the bottom of the tire and the road will be quite different and trying to get back to an equal speed.

A note on this, I've seen more suspension parts fail first including in professional motorsports. If you have calculated another part to fail first in almost anyway you look at it, then you could probably bring the part down into the 0.9 safety factor range. Yes, yes, yes, I know, that sounds bonkers, a 0.9 SF! It all matters on what you define your safety factor off of anyways as an expected load. However, it was mentioned in the suspension thread to keep in mind which parts will fail first as 'sacrificial' parts so that you don't destroy a damper or other expensive component. Will your suspension fail before you achieve the load condition you are calculating for? Will something else?

I know very well that one OEM in particular designs certain parts to 0.9 SF, but that is all relative to the load conditions they are calculating to.

Does the car lift like this?
http://chivethethrottle.files.wordpress.com/2011/10/tripod-500-work-6082211-1-flat550x550075f-lifting-wheel-through-the-dipperwtmk.jpg?w=500&h=333

or like this?
http://stwp.selftitledmagazi.netdna-cdn.com/wp-content/uploads/2011/03/lowrider.jpg

Never mind both of these are lifting front wheels, they were convenient for pictures.
Axles don't fail often. If they do, you're trying too hard.

[Warpspeed]

Hard core 10,000 rpm clutch dumps in first gear will definitely generate a much higher torque transient than the engine's normal rated torque.
The initial static breakaway friction of the tires will be higher too, especially with the extra sudden rearward load transfer.
All that plus inertia... where the irresistible force suddenly meets the immovable object.

Best to allow for some axle windup to hopefully absorb some of the impact energy, which (should) be the case anyway.
If you are running super stiff ultra thin wall axles to save weight, good luck with that.

A more practical solution might be springy axles that will wind up, and absorb part of those vicious torque transits. But that in itself creates a highly stressed potentially unreliable part. Not really a problem if you keep an eye on those axles for fatigue cracking or yielding.

How you formally design it I really don't know, because the loads cannot easily be quantified.
Going too far in either direction spells trouble.
Testing and experience, plus some idea of cycles to failure might create some acceptable rules of thumb for different classes of vehicle.

One thing is for sure, the serious off road guys seem to be able to break just about anything. Making the axles much stiffer and stronger usually just moves the failure point to some other part of the transmission.

[Canuck Racing]

True, I did forget about axle hop, though if you follow my advice of figuring out what can actually be manufactured you'll still find you have plenty of safety factor.

On a personal note, I don't hate Z, he's one of the few here that realize how over-designed most FSAE cars are and how simple they can really be. His way of simplifying things... well that can be debated...

[Z]

Originally posted by Canuck Racing:
On a personal note, I don't hate Z...
Aw, Canuck, you're making me feel all... err,... unusual... http://fsae.com/groupee_common/emoticons/icon_smile.gif
~~~o0o~~~

To clarify what I was getting at above (since this is often overlooked).

The car is off the ground (perhaps on your build stand). The clutch is disengaged, a gear selected, the engine reved to 10,000 rpm, and the rear wheels are stationary. The clutch is then "dumped" (engaged). A very short while later both the engine and wheels are rotating (at speeds, relative to each other, = the selected overall gear ratio).

The question is, what is the torque in the driveshafts during the above process?

IMPORTANTLY, note that gas pressure forces from the engine (ie. "torque") is not involved here, nor is there any resisting force from road-to-tyre grip at the wheels. The driveshafts are stressed purely by the inertia of the rapidly spinning engine trying to overcome the inertia of the initially stationary wheels.

To get some numbers let's assume that the rotational kinetic energy of the engine, before clutch dump, equals rotational KE of engine plus wheels, after clutch dump. This is not accurate because of energy lost during clutch slip, but it gives an upper limit. Based on "back-of-TV-times" calcs, the final speed of the rear wheels is,

Wwf = sqrt((Ie*Wei*Wei)/(2*Iw+GR*GR*Ie))

where,
Wwf = final wheel speed (rad/sec),
Wei = initial engine speed (rad/s) (after engagement Wef = GR*Wwf),
GR = overall gear ratio,
Ie = MoI of engine (= Me*Rge*Rge),
Iw = MoI of each wheel (= Mw*Rgw*Rgw).

The driveshaft torque (for each wheel) is then,

T = Iw*Wwf/dt,

where,
dt = time taken for the clutch to "lock" engine and wheels together...
~~~o0o~~~

Example 1.
=========
Each wheel has mass = 10 kg, radius of gyration = 0.2 m, so Iw = 0.4 kg.m.m (upper end of FSAE range).

Engine rotating parts (crank, clutch, etc.) have mass = 10 kg, Rge = 0.05 m, so Ie = 0.025 kg.m.m (lower end of range?). Initial engine speed Wei = 10,000 rpm = ~1,000 rad/s (because Pi = ~3).

Overall gear ratio = 10:1, so final wheel speed,

Wwf = sqrt((0.025*1000*1000)/(2*0.4+10*10*0.025)) = 87 rad/s,

(ie. wheels at ~870 rpm, and engine at ~8,700 rpm).

Now, BIG ASSUMPTION, let's say all this happens in dt = 0.1 seconds. So,

Torque = 0.4*87/0.1 = 348 Nm!
~~~o0o~~~

Example 2.
=========
Wheels as above, but big flywheel engine (like Jawa single) with mass = 10 kg, Rg = 0.1 m, so Ie = 0.1 kg.m.m (upper end of range).

Gear ratio = 4:1 = "higher" gear, which is worse in this case (think about it...). So final wheel speed,

Wwf = sqrt((0.1*1000*1000)/(2*0.4+4*4*0.1)) = 204 rad/s,

(ie. wheels at ~2,000 rpm, engine at ~8,000 rpm).

Again, assume dt = 0.1s. So,

Torque = 0.4*204/0.1 = 816 Nm!!!
~~~o0o~~~

Bottom Line.
==========
Ok, the 10,000 rpm initial engine speed is a bit much, and the 0.1 sec clutch engagement maybe a bit quick. But, whichever way you look at it, the above "impact" loads are significant (and they are why you swing a hammer, rather than just trying to push the nail in with it). And keep in mind that when the wheels are on the road, and the engine is at WOT, then the torque from the engine working against tyre grip is ADDED to the above inertial loads.

So, as suggested by Tony and others above, a "springy" driveline is good (it increases dt), a generous SF is good, lots of testing with constant checking for twist or cracks in the driveshafts is good, and a "Plan B", such as quickly replaceable spare shafts, is also good. Oh, and dumping a heavy duty (no slip) clutch from 10,000 rpm is NOT good. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Z

[AxelRipper]

Good off the wall point there Z, as that should really be the bare minimum of what to design for (aka, what your driveshaft force would be if you were running ceramic tires on ice).

However, your tire patch forces are still quite a bit higher than even your supposed worst case scenario realistic (non-Jawa since we like transients around here) example, siting the OP's 650 NM per wheel.

I can also tell you that a lot of drivetrain design can also depend heavily on the competency of your drivers. If your drivers drive the car normally, even doing a full throttle clutch dump, your engine will likely stall before destroying anything else in the drivetrain (unless of course your safety factors are really miniscule/calculations are garbage). However, the real problem comes when your driver does something stupid like try to put the car into first gear without the clutch while holding it at redline. In our case in Cali 2011 our diff mounting tabs and .5" aluminum diff plates took the beating, and not our (really quite overbuilt) axles. That is your worst case scenario that hopefully your drivers don't do, so you *shouldn't* have to plan for in a racecar. Road car, maybe, but not a race car.

[onemaniac]

Thanks for clarifying Z,
Now my question is, how much inertia would the transmission, chain and the differential contribute before the energy is transferred to the shafts? as well as the dampening effect by the chassis (though probably small enough to neglect)?

So what I think is, after all factors taken account - and assuming dt is longer than 0.1 seconds - the stress on the drive shaft itself due to the inertia of only the parts is not that drastically high (something that can be covered by the design FOS)
Then again this is only an assumption of mine...

[Warpspeed]

You could probably rig up a simple test.

Enclose your driveshaft in a thin wall tube, secured to the driveshaft at one end only.

At the other end, connect the tube to the driveshaft through something that will deform and record the peak angular displacement between the loaded driveshaft, and the unloaded outer sleeve.

Get lead foot Freddy to do a few burnouts, and from that, you should be able to discover how far that shaft wound up under peak load.

It should then be fairly straightforward to design a shaft that has both adequate strength, with enough spring in it to absorb at least some of the shock energy.

[Pete Marsh]

I was once told by a Ford power train engineer that physical testing of the Mustang showed only a 10% reduction in the "clutch dump" drive train stress on ice compared to dry concrete.

Sure in FSAE, flywheels and road wheels are smaller and lighter, but if you overlook this inertia effect, and how to manage it,...well, at least you will understand 10 or 20% of the loads really well.

Rally also produces many drive shaft failures, with many "fixes" moving the issue to a different part, usually the gear box.

Pete

[BillCobb]

Stress is usually not the major underdesigned driveshaft problem, strain is. Therefore, a major braking event can be your worst case for a shaft. Then, other components (motor mounts, U-joints, and gears) will be the fuses in the driveline. Power hop is the driveline designer's nightmare. To attenuate this, major players use two different shaft torsional stiffnesses so this stick-slip situation caused by tire characteristics is moderated.

[murpia]

Originally posted by BillCobb:
Power hop is the driveline designer's nightmare. To attenuate this, major players use two different shaft torsional stiffnesses so this stick-slip situation caused by tire characteristics is moderated.
FWD, RWD or both?

Would this generally be equal length driveshafts of different diameter, or equal diameter driveshafts of different length?

Elsewhere on these forums you'll find advice to use equal stiffness driveshafts to eliminate 'torque steer' a myth I've tried to bust with reference to resonances caused by tyre stick-slip, etc.

It sounds like you have much more appropriate experience here, Bill.

Regards, Ian

[BillCobb]

High powered RWD cars usually are the most difficult for "power-hop". Its a combination of the physics and the parts (motor size, transmission capacity and tire friction functions). In production cars, the problem can be solved by electronic engine controls, although a few new cars still suffer because they forgot to evaluate it in reverse gear. Yes this defeats the fun of a high powered motor. The the shape and values of the tire mu-slip curve cause this instability. So, tire construction, tread compound, and BRAND are major players.

In production cars with offset drivelines, they may use a different length AND diameter because its a different part number. The neck-down process is used (tapered shaft). In oval track, hollowing out can be done so that other teams in the garage can not view the shapes of the axles. BTW: in oval track, this situation can be responsible for 'loose-off' response, because the stick-slip on one tire (usually the inside one) suddenly breaks off the sideforce levels at the rear axle. Its also the one with the highest peak mu. Tire pressure(s) (used to control rev/mile) can help you with this if the car is pretty close to 'good'. That's why 1/2 a lb of air pressure increment is used. Its not for cornering peak, its for differential control.

FWD torque steer is a tire and differential friction problem. Now you have an axle with a steer compliance degree of freedom AND a tractive force sensitivity. This is best solved by picking tires with a large, negative pneumatic scrub (load normalized overturning moment). This preserves the sign of the mechanical scrub radius (usually negative) on both wheels and the car drives off the line relatively straight. If one wheel breaks traction and the scrub goes positive, you are gonna have a broken wrist. I'm talking about 'torque-induced-steer' here, not the pull that occurs when Mom floors the Vibe and it drifts to the right or left while she's on the phone with the soccer coach.

Whateva..