Equal Length/Torque Matched Driveshafts

[drivetrainUW-Platt]

Between team members, we were having some discussion about the importance of matching up driveshafts. We are running the Honda atv differential which is very narrow, and with the transmisson output gear ofset to one side, our driveshafts are definitely not equal lenght, I believe they were 6 or 12" different. I had them made to the same diameter as the stock shafts, and they held up thru competition, but I would like to hear what others have to say about emphasizing equal lenght or torque matched shafts.

[Conor]

Building your shafts to the correct cross sectional thickness to counteract the effects of torque steer is a pretty good idea. At competition, one of the design judges (didn't catch his name) talked with me for quite some time about the importance of getting the thickness of each shaft correct so you don't get any torque steer. He was very keen on this and told me that my present from him for next year was to make sure I got that one right. So, next year, I'll be making sure our drivetrain doesn't get any torque steer. By the way, can you tell me a little more about the differential you're using. I've been shopping around and was looking into ATV differentials. What kind of custom housing did you have to make for it, how much did you pay for the diff, did you modify the factory housing, etc, etc... Thanks.

[Greg 08]

I had this discussion with a member of my team as well. We decided this was just fine for based only upon simple mechanics of materials. We calculated the difference in linear travel at the wheel with the same torque (max produced from engine converted up from gearing etc) going to each shaft to be somewhere in the neighborhood of 1/4". My question to you is if that creates significant steering input? We didn't think so, but then again it may. Regardless, how much effort is it to center the diff? For our car, it meant offsetting the motor, or finding a new source and paying quadruple for non-stock driveshafts. IMHO, it doens't seem worth it, compared to where else we could spend our time and money. Duwe, where do you plan on getting custom hardened driveshafts? Conor, we made our own diff housing out of 1 piece of billet aluminum. It was supposed to be 7075, stuff happens and it came out 6061 but appears to be holding up with the limited driving we have done on it. It was an extensive machining project by a very talented and ambitious member of our team. It took him nearly 50 hours of machine time, and his second would probably be half that, but unless you have professionals doing it I wouldn't estimate it taking any less. He made it during the summer at our sponsors machine shop when he wasn't working there. We are reasonably happy running our torsen diff.

Sorry about the length.

Greg Ehlert
Michigan Tech

[Chris Clarke]

I think that torque steer due to deflection in different length driveshafts is probably one of the most misunderstood concepts around. If you actually think about it, it makes absolutely no sense whatsoever. You are telling me that because one side deflects 0.001 radians more than the other, it is going to transmit significantly more force to the other wheel? I dont really think so. Even if that was the case, it would only be in transient until the driveshafts reach their loading, which lasts about 1/100th of a second.

Also, trucks have had unequal length driveshafts in the front for how long? The difference in their lengths is around 3 feet. This was never an issue with them.

Our car had driveshafts that were about 6" different in lengths, and we never had any noticable torque steer at all in the car.

I went through this a while ago and these are the reasons that I came up with for torque steer occuring, feel free to disagree with them.

I do think that torque steer in FWD, independant susp cars (maybe cars with rear steering also) is due to different length driveshafts, but I think the reasoning is different from what most people say.

With independant suspensions come CV joints, or something similar. The design of these joints is so that you can allow for rotation in 2 planes, one for steering and one for vertical tire movement. Part of the consequence of CV joints is that you get secondary forces out-of-plane with the axis of the driveshaft. These out of plane forces increase as the angle of the CV joint increase. SO, when you have unequal length driveshafts in a independant suspension car, you will get different angles in each side as the front end lifts/drops, like when you are accelerating. These different angles result in a net "out-of-plane" force on one side. This force will act about the kingpin axis to create a moment about the steering axis, which gives you torque steer.

Having said that, since everyone in FSAE is RWD, and for this discussion, I am assuming not using rear wheel steering, all the cars will have some form of toe-link in the rear. This toe-link will resist the force about the steering axis, so the only "torque-steer" that anyone in FSAE should be seeing is due to the deflection in their toe-links, which should be none.

There are other reasons why torque steer may occur under cornering (or different surface conditions), but that isnt generally what people talk about.

[Chris Clarke]

Oh, another point, if you have a have a shaft driven solid-axle RWD car, you will get torque steer due to the difference in vertical loading on the tires from the reactions on the crown gear, but this isnt the case with chain driven, or independant susp RWD cars.

[drivetrainUW-Platt]

For the Honda atv diff, do a search, I said to much already in another forum.

I really didnt think that torque matching was that critical, I never really did the math on deflection, I will have to run some numbers to put the other members at bay.

For our shafts we had a gear company is Wisconsin cut the splines then we had them heat-treated at another facility in Iowa.

[BStoney]

Originally posted by Conor:
Building your shafts to the correct cross sectional thickness to counteract the effects of torque steer is a pretty good idea. At competition, one of the design judges (didn't catch his name) talked with me for quite some time about the importance of getting the thickness of each shaft correct so you don't get any torque steer. He was very keen on this and told me that my present from him for next year was to make sure I got that one right. So, next year, I'll be making sure our drivetrain doesn't get any torque steer. By the way, can you tell me a little more about the differential you're using. I've been shopping around and was looking into ATV differentials. What kind of custom housing did you have to make for it, how much did you pay for the diff, did you modify the factory housing, etc, etc... Thanks.

I am going to venture to say that your design judge was Steve Fox from PowerTrain Technology outside of Chicago. We got this question two years ago in 2005.

[Conor]

Regardless of what you guys feel like arguing about this topic, if you don't have torque steer covered and it comes up at next year's design competition, you'll want to shoot yourself in the foot. You either say screw it, go the easy route, and don't worry about it, or you try for the extra design points and get it right. Those are the only two choices to be made. I have no idea which I'm shooting for yet, but the easy way sounds much better.

[Underthefloor]

If you are running a torsen with 2.6 TBR unequal drive lengths could be a problem. The half shafts act like torsional springs with, I assume a, linear spring rate. The shorter side will twist less and so the diff will bias to that side the full TBR.

I don't know how fast the tires can absorb the difference in the drive angles but I assume that the higher loaded tire will have a higher slip ratio.

[Wright D]

Equal length driveshafts are easier to do then different cross section drive shafts. You can get the driveshafts the same length by building a custom diff housing that centers the diff in the car. If you are building a custom diff case, why not also solve the driveline problems at the same time. By centering the diff there is no need for custom shafts to be made; so getting replacements, or spares is much fast/easier. Of course there are many ways to get the engine connected to the tries, this way happened to fit with our needs.

Here is a picture of what we came up with:

http://formula.engr.arizona.edu/pics/4-march-2005/05_img%20(55).jpg

[Chris Clarke]

Regardless of what you guys feel like arguing about this topic, if you don't have torque steer covered and it comes up at next year's design competition, you'll want to shoot yourself in the foot. You either say screw it, go the easy route, and don't worry about it, or you try for the extra design points and get it right. Those are the only two choices to be made. I have no idea which I'm shooting for yet, but the easy way sounds much better.


Implementing design changes based strictly on the reasoning "so and so told us to do it" will get you nowhere with the design judges.

[RacingManiac]

our drivetrain person did the calc and concluded that the torque steer effect on vehicle yaw under power is about 1 degree in our particular setup. This is the worst case scenario applies to situation like launching the car or exiting a slow corner in 1st gear, therefore we found it to be not worth while. We tell Judge Steve Fox this and his point is that even with only 1 degree of induced yaw it is something the driver have to correct for and thereby incresing his workload.

Valid point, how important you want to take it is your call.

[Chuck Maddocks]

dustin,
looking at the picture of your diff carrier, it appears that you will have a very small housing/sleeve for holding the oil, and therefore very little oil. i noticed a few other teams at competition also had this design. 1.) did you run this diff at competition? and 2.) if so, was there any issue with the tiny amount of oil (especially during endurance)?. other teams with similar design feel free to comment. We tried (unsucessfully) to move to a torsen based driveline last year, but we have a very large housing/seal so that the diff would have an oil bath that it would dip into each rotation. big moment of inertia.

as far as torque steer due to unequal shafts, search the forums, there are other posts on this topic. another way around this is to use inboard stub axles with unequal lengths/cross sections. taylor-race also has two different size tripod housings.

also, when cornering, a torque biasing diff will send unequal torque to each wheel, which can amplify or cancel out the torque steer, depending on which way you turn. therefore i feel that having equal angle of twist at the wheels is important because your torque biasing diff will amplify the torque steer in the worst case scenario. in a straight line (ie acceleration test) torque steer, no matter how small, will require some steering correction. the tires will scrub and you will not go as fast. however, if correcting this problem means axles with higher moments of inertia, you will loose horsepower everywhere. so it's a trade off. just try to get them as close as possible, with the lowest possible moment of inertia (hahah!)

[John_Burford]

Given the large slip ratios of these high power FSAE cars, torque steer as a result of shaft deflections won't be noticed. On the other hand, different angles left and right from the wheel rotional axis could cause torque steer. This can happen when you combine an offset between the diff axial and the wheel axis with a short shaft on one side and a long shaft on the other.

The bigger issue is with the strength of the shorter shaft. The shorter shaft being stiffer will result is higher impact loading and is usually the first to fail.


Dustin

Have you had problems with the diff. bolts loosening?

John Burford

[Greg 08]

our drivetrain person did the calc and concluded that the torque steer effect on vehicle yaw under power is about 1 degree in our particular setup.

RacingManiac: How did you come up with that number? What assumptions and calculation methods did you use? Do you run a torsen with unequal halfshafts? What is 1 degree measuring, the difference in yaw of the car due to the unequal rotations of the shafts? Did you account for sidewall deflection?

We figured that the minimal difference in deflection at the tire would be compensated for in sidewall deflection, but I could buy into Under the Floors comment:


If you are running a torsen with 2.6 TBR unequal drive lengths could be a problem. The half shafts act like torsional springs with, I assume a, linear spring rate. The shorter side will twist less and so the diff will bias to that side the full TBR.

I don't know how fast the tires can absorb the difference in the drive angles but I assume that the higher loaded tire will have a higher slip ratio.


Beaver Racing

Thanks Maniac, this is an interesting discussion and if Steve Fox is reading any of this, his input, line of thought, assumptions would be greatly appreciated.

Greg Ehlert
Michigan Tech

[Conor]

Originally posted by Wright D:
Equal length driveshafts are easier to do then different cross section drive shafts. You can get the driveshafts the same length by building a custom diff housing that centers the diff in the car. If you are building a custom diff case, why not also solve the driveline problems at the same time. By centering the diff there is no need for custom shafts to be made; so getting replacements, or spares is much fast/easier. Of course there are many ways to get the engine connected to the tries, this way happened to fit with our needs.

Here is a picture of what we came up with:

http://formula.engr.arizona.edu/pics/4-march-2005/05_img%20(55).jpg

Sprockets on most engines are offest from the center of the engine. Even if you center the differential in the frame, the sprocket itself isn't centered if your engine is centered. How does the offset sprocket effect this? Also, thanks a million to your team. You guys really helped us out at competition and it was a pleasure sharing a paddock next to your team. Our entire team is grateful for the new starter clutch you guys so gracially provided.

[murpia]

Originally posted by Underthefloor:
If you are running a torsen with 2.6 TBR unequal drive lengths could be a problem. The half shafts act like torsional springs with, I assume a, linear spring rate. The shorter side will twist less and so the diff will bias to that side the full TBR.

This makes no sense. Diff torque bias is not influenced by the amount of shaft twist.

Regards, Ian

[RacingManiac]

Originally posted by Greg 08:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> our drivetrain person did the calc and concluded that the torque steer effect on vehicle yaw under power is about 1 degree in our particular setup.

RacingManiac: How did you come up with that number? What assumptions and calculation methods did you use? Do you run a torsen with unequal halfshafts? What is 1 degree measuring, the difference in yaw of the car due to the unequal rotations of the shafts? Did you account for sidewall deflection?

We figured that the minimal difference in deflection at the tire would be compensated for in sidewall deflection, but I could buy into Under the Floors comment:


If you are running a torsen with 2.6 TBR unequal drive lengths could be a problem. The half shafts act like torsional springs with, I assume a, linear spring rate. The shorter side will twist less and so the diff will bias to that side the full TBR.

I don't know how fast the tires can absorb the difference in the drive angles but I assume that the higher loaded tire will have a higher slip ratio.


Beaver Racing

Thanks Maniac, this is an interesting discussion and if Steve Fox is reading any of this, his input, line of thought, assumptions would be greatly appreciated.

Greg Ehlert
Michigan Tech </div></BLOCKQUOTE>

I am not the D-train man on the team, so I can't tell you for sure. Go to our website in my sig and I am sure you can find his e-mail and he can probably elaborate. We did run a Torsen in comp and I believe that was the calc based on.

[Wright D]

Originally posted by Conor:

Sprockets on most engines are offest from the center of the engine. Even if you center the differential in the frame, the sprocket itself isn't centered if your engine is centered. How does the offset sprocket effect this?

Check out the diff Taylor Race Engineering sells, it has an offset sprocket and the diff is centered. Also, if you check out more of the pictures on our site, http://formula.engr.arizona.edu, you will notice that the diff is driven from one end of the housing, allowing the sprocket to be offset.


Our ass has been saved at competition by some other nice teams help too, just pass on the favor http://fsae.com/groupee_common/emoticons/icon_wink.gif. The more teams that do well; the better fsae competition is. The better fsae competition is; the better we all do.

[Underthefloor]

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Underthefloor:
If you are running a torsen with 2.6 TBR unequal drive lengths could be a problem. The half shafts act like torsional springs with, I assume a, linear spring rate. The shorter side will twist less and so the diff will bias to that side the full TBR.

This makes no sense. Diff torque bias is not influenced by the amount of shaft twist.

Regards, Ian </div></BLOCKQUOTE>

Hey Ian, I misspoke. I don't think that the shorter shaft would always get the max torque allowed by the tbr but that would be an upper limit. The difference in torque would be related to the difference in half shaft stiffness.

[murpia]

Originally posted by Underthefloor:
Hey Ian, I misspoke.
Hey, no problem.
This whole thread has got me thinking the past few days. My 'intuition' tells me that within sensible margins (say a driveshaft stiffness ratio of less than 2:1) then it has a negligible effect. But we shouldn't just rely on intuition we need to back it up with calcs...
Hopefully I'll post something next week.
Note that there are 2 separate issues here. One is a difference in driveshaft torsional stiffnesses left to right. The other is a difference in joint angularity left to right resulting in unequal torque reactions. This angularity issue predominates in FWD road cars which is why they spend extra cash to have identical shaft lengths (which results in similar joint angularity) to reduce torque steer feedback through the steering wheel. Of course even with identical length shafts the angularity will never be exactly equal due to steering geometry, ackermann and suspension roll. Also the steering trail and scrub radius and pneumatic trail/scrub will play a part, as will split mu surfaces and torque biasing diffs. So a FWD road car applying torque with steering lock will probably always suffer torque steer to some extent.
For the independant suspension RWD case, a difference in shaft angularity will be reacted through the toe compliance of the rear suspension. If this is sufficent then you have no problem. Split mu surfaces, torque biasing diffs and/or unequal contact patch forces due to cornering forces can and will result in a yaw moment being applied at the rear axle, which the driver could interpret as 'torque steer', but that is a whole different area of analysis...
Regards, Ian

[kwancho]

Great post. Definitely makes it all more clear for me. It makes sense that you really don't want, and really shouldn't have, enough toe compliance to give you real torque steer.

[Buckingham]

Driveshaft stiffness is analogous to suspension spring stiffness. How much does it matter if you have a different wheel rate on the left side of the car than on the right side of the car? Design judges ask you why you design a system with unequal stiffness shafts for they same reason they would ask you why you would knowingly design a suspension with different left/right wheel rates. The difference is that most judges wouldn't notice if you designed your wheel rates to be 2% stiffer on the left side than the right, but every judge can SEE the OD and length of your driveshafts, so you'd better be able to explain WHY you DESIGNED them to be different.

[Underthefloor]

I think that a 2:1 stiffness ratio would result in a 2:1 torque ratio with a torsen untill tire slip equalizes the loading. I don't know how fast the reaction time for this would be but every time you let of the gas or shift, the diff will be biasing torque.

I don't understand how having different drive line angles would result in different torques. Is there a good explananation of this available?

[ACXY]

Hi,

From what I have read in this thread so far, it seems like torque steer due to unequal drive shaft lengths is rather negligible, especially since it is momentary.

On the other hand, I subscribe to the idea that different operating angles of the drive shafts constantly results in a torque difference. At this point, am I right to say that there will be 'loss' of torque when it is transmitted through an angle? Afterall, torque is a vector and has direction...

Finally, I am puzzled about the role of the differential in the event of torque steer. Does it even play a role in the first place? Does anybody care to share/clarify?

Cheers,
Anthony

[Buckingham]

When considering the idea that a 'loss' of torque can occur, don't forget that (assuming no energy storage): Power in (torque*speed) = Power out (torque*speed) + heat/sound/etc

[Bill Kunst]

Donovan and all,

This is a very interesting topic. There are a few areas that are pretty over emphasized. I will list the areas in discussion and their importance.

1. Torque steer (driveshaft length)= negligible degree difference as determined by many teams.

2. Torque steer (drive shaft angle)= Really, how much angular difference is seen. These are cv's, right?! Therefore they should not have a difference in speed, so no difference in torque,right? (minus that lost to heat and motion from the <5 degrees of difference in angle)

3. Torque steer as a result of the diff accomodating #1&2. Does anyone really know this? I would guess that this could be the most significant difference (between 1,2,3) in some applications- cam and pawl vs. clutch packs vs. torsen type.

4. Design judging = The most important portion of this discussion. When it comes down to it, you have to be able to convince the judges that everything you calculated makes it okay, or great, or better than the rest. Really, you want to convince the judges that you have the best case scenario from your design. If you dont have equal lenght/equal angles you will have you work cut out for you.

5. Having common parts= refer back to #4

This is just how I see it, but my emphasis for doing this would be to gain points, not for "reduction in torque steer"(even though this would be the reasoning in the design tent along with other excellent reasons).

Keep going with the research, I want to know the conlusive evidence if anyone has it (for the sake or knowing).

Bill Kunst

[murpia]

Originally posted by Underthefloor:
I think that a 2:1 stiffness ratio would result in a 2:1 torque ratio with a torsen untill tire slip equalizes the loading. I don't know how fast the reaction time for this would be but every time you let of the gas or shift, the diff will be biasing torque.

I don't understand how having different drive line angles would result in different torques. Is there a good explananation of this available?
Yes to the first part, this is the crux of the whole 'different stiffness shafts = rear axle yaw moment' problem. 1) it is transient and 2) it depends on TBR & tyre slip ratio.

As to the 2nd part different joint angles = different torque reactions, not different torques. There is a big difference and engineers need to understand it...

Regards, Ian

[Underthefloor]

Ian,

That makes sense to me about the different reaction torques. I went back and read the posts more carefully and I think that you are right about the angles and toe compliance.

The transient time for unequal length halfshafts is probably small because the wheels only have to slip about an inch or two in the most extreem cases. Is this enough to cause problems?

[murpia]

Consider the four following limit cases, all are technically impossible yet illuminate the physics. Assume non-equal stiffness driveshafts in all cases, with no other compliances in the system:

1) Zero tyre slip ratio (tyres 'geared' to the road) / Unity Torque Bias Ratio (an open diff)
As torque is applied each driveshaft twists a different amount according to its stiffness. This differential twist is absorbed by the open diff and no yaw moment results. The time taken to reach a steady-state condition depends on the torque applied and the component inertias.

2) Zero tyre slip ratio (tyres 'geared' to the road) / Infinite Torque Bias Ratio (a spool)
As torque is applied the spool biases the torque in the ratio of driveshaft stiffnesses. As no slip can occur anywhere in this system this torque bias is maintained indefinitely and results in a constant yaw moment at the axle of Torque * Driveshaft Stiffness Ratio * Axle Track Width.

3) Infinite tyre slip ratio (tyres just spin and the car never moves) / Unity Torque Bias Ratio (an open diff)
As 1) except that depending on the various inertias in the system some driveshaft twist will be absorbed by the tyre slip. No yaw moment results.

4) Infinite tyre slip ratio (tyres just spin and the car never moves) / Infinite Torque Bias Ratio (a spool)
As torque is applied the spool biases the torque in the ratio of driveshaft stiffnesses. Once the driveshafts are twisted by their full amount the tyres are transmitting equal torque and are displaced rotationally by an amount equal to the difference in driveshaft twist. A transient yaw moment is applied which decays from Torque * Driveshaft Stiffness Ratio * Axle Track Width to Zero over the time taken to twist the driveshafts.

Any real car is an example of 4) where the Infinite Tyre Slip Ratio and Infinite Torque Bias Ratio are replaced by real values. To calculate the duration of the yaw moment transient requires real numbers for all the component inertias and stiffnesses (not just the driveshafts), the diff TBR and the tyre slip ratio curves.

Hope this helps, Ian

[Steve Fox]

"...this is an interesting discussion and if Steve Fox is reading any of this, his input, line of thought, assumptions would be greatly appreciated."� Posted by Greg Ehlert, Michigan Tech

I knew my ears were burning for a reason, I just couldn't figure out it was THIS thread in FSAE.com. causing the problem. I am sorry it took me so long to chime in here, but this has been a very busy summer racing season for us. Greg, you asked, so here is my input:

First I guess I should say that I am flattered that one question from a design judge at a competition would generate this much virtual ink. This question is really just a simple way of getting you, the designers, to think about drivetrain efficiency, and avoid the easily corrected problem of torque steer induced by unequal torsional spring rates in your half shafts. http://fsae.com/groupee_common/emoticons/icon_wink.gif

This problem is caused by unequal length (or unequal diameter) half shafts. It is very easy to solve for the torsional spring rate of a shaft, and all you have to do if you do end up with unequal length half shafts on your car is match the torsional spring rates of each shaft. This is most easily accomplished by increasing the diameter of the longer shaft.

I should go on record right now and say that it does not bother me if your diff is centrally located, and you have equal length half shafts, or if it is not. As long as you can explain to the design judges why your design / widget / concept works, that is perfectly acceptable. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Giving consideration to commonizing the parts on your car for manufacturing simplification is admirable. As good design engineers we always try to reduce the number of different parts on our race cars. Equal length half shafts that are interchangeable from left to right would be considered an advantage in the manufacturability, and serviceability sections of the design judging criteria.

Solving for the difference in linear travel at the wheel is just an over complicated way of looking at the problem. You will get accurate results with this method, but you will be just as accurate solving for torsional spring rate of the half shafts.

We design judges realize that all teams have varying degrees of machining and fabricating skills, so if it is easier for your team to build an offset diff carrier and equal length half shafts then that is a good solution for you... and a perfectly good solution in the eyes of the design judges. If your team is turning your own solid steel, unequal length, half shafts, and you see no difficulty in varying the O.D.'s, then that solution is perfectly fine as well. If you DO NOT have torsionally matched half shafts, you should be prepared with a mighty good explanation for why your car works better without this feature.

"Our drivetrain person did the calc and concluded that the torque steer effect on vehicle yaw under power is about 1 degree in our particular setup. This is the worst case scenario applies to situation like launching the car or exiting a slow corner in 1st gear, therefore we found it to be not worth while. We tell Judge Steve Fox this and his point is that even with only 1 degree of induced yaw it is something the driver have to correct for and thereby increasing his workload. Valid point, how important you want to take it is your call."� Posted by RacingManiac, 4th Year @ UofT Racing. My point to Racing Maniac when I discussed this with him at the competition was that IF you can do something to remove workload from your driver, wouldn't you? Your job as racecar engineers is to make your driver's job as easy as you possibly can. If your driver is having to deal with torque steer and consciously (or unconsciously) make steering input corrections because of it, you are adding to his (her) workload, and keeping him (her) from doing something more important... like winning!

"as far as torque steer due to unequal shafts, search the forums, there are other posts on this topic. another way around this is to use inboard stub axles with unequal lengths/cross sections. taylor-race also has two different size tripod housings."� Posted by Chuck Maddocks, Burlington, VT. Chuck has it figured out. It is important to look at the entire half shaft solution, and solve for the total spring rate of all pieces transmitting torque to the wheel.

"The bigger issue is with the strength of the shorter shaft. The shorter shaft being stiffer will result is higher impact loading and is usually the first to fail."� Posted by John_Burford. Another potentially correct observation. IF you build the half shaft so thick as to be extremely stiff in spring rate, you might potentially run the risk of fracturing it. The trick is to select a half shaft material that is strong AND flexible, AND design it to the correct torsional spring rate for the load you are transmitting through it, AND heat treat it properly.

The comments that I am seeing about the limited slip torque biasing ratios being affected by the different torsional spring rates of unequal length half shafts is really not even worth speculating on. The simple answer is make your half shafts equal in torsional spring rate. Then you have one less item to worry about negatively affecting the performance of your car. http://fsae.com/groupee_common/emoticons/icon_biggrin.gif

Joint (CV, U-joints, Tripods, etc.) angularity is a relatively small parasitic drag in a racecar drivetrain. It is my experience that the drag from angular deflection of these parts is magnitudes smaller than the forces of unequal half shafts we are discussing here. For the record: If any of you come up with the explanation that you are getting torque steer strictly from the different angles that the joints are subjected to due to different length half shafts with a resulting ˜out of plane' force that acts on the king pin axis, you had better have lots of engineering calcs, post design testing, and maybe a doctoral thesis to back up your argument. That one will be mighty hard to sell me on. (Sorry, Chris!) http://fsae.com/groupee_common/emoticons/icon_frown.gif

The amount of time duration of the uneven rear tire loading due to unequal half shaft torsional spring rate is directly related to the number of times the driver gets on and off the throttle. Every time he (she) goes to WOT, the half shafts wind up at unequal rates. One tire drives forward sooner than the other. This should (does) result in torque steer. Can your team detect this torque steer simply by listening to driver feedback? The answer is maybe. Maybe it is bad enough to be really noticeable. Maybe your driver is not experienced enough to feel the subtle difference. Maybe your driver is making unconscious corrections for the torque steer because he (she) has never had the opportunity to drive the car WITHOUT any torque steer. http://fsae.com/groupee_common/emoticons/icon_eek.gif

Here is another facet of this topic for you all to consider. When the rear wheels are driving forward at unequal rates (unequal length half shafts that are NOT torsionally matched) does this not put a side loading (due to yaw change) on the front tires? Isn't this side loading or ˜scrubbing' slowing the car's forward acceleration? http://fsae.com/groupee_common/emoticons/icon_confused.gif (Yes, of course IT IS!)

The bottom line is this: Torsionaly matched half shafts are a superior design due to the fact that they are more efficient at transmitting torque, they offer less rolling resistance under acceleration, AND they make it easier for the driver to control the racecar. http://fsae.com/groupee_common/emoticons/icon_cool.gif I cannot think of any performmance drawbacks, so that makes torsionaly matched half shafts a feature that is too good to not include on your racecar.

The one thought that you should carry away from my post is this: It is YOUR race car. All race cars, by their very nature, are a blend of compromises in design AND performance. The truly great aspect of this competition IMHO is that it has a very thin rule book. This leads to diversity in design. I personally very much enjoy seeing all the differences in design that you all come up with. No one design is going to always be the fastest way around the track. As long as you pay attention to the basics (torsionaly matched half shafts being just one of the details you have to address) your car will have an excellent chance of proving itself on the battlefield of competition.

I am pleased to see such a lively open discussion of this topic. I sincerely hope that I have cleared up some of this topic's muddiness. What you all should be contemplating now though, is what innocent little question am I going to ask at FSAE 2007? http://fsae.com/groupee_common/emoticons/icon_razz.gif

FYI Most of the motorsports design judges look in on FSAE.com regularly. For the most part we try not to chime in and give away too many of the answers, but sometimes it is hard to resist. If any of you should want to ask me questions directly, feel free to contact me at Sales@PowerTrainTech.com I will not ˜give you the answers to the test', but will certainly try and guide you to a higher level of linear acceleration.
http://fsae.com/groupee_common/emoticons/icon_cool.gif

[drivetrainUW-Platt]

From a parts/machining/cost standpoint, it is easier to add 2 inches to a driveshaft then to make a diff housing that has a big extrusion to ofsent the mount of the drive gear, with bigger bearings and more complixity.

If we are so worried about things being symetrical, then we need to build our own transmissions that allow for the output gear to be ofset in the ideal location for equal length halfshafts.

Our Honda ATV diff is probably only 4" wide so equal length shafts would require a complete remanufacture of the housing, which is run basically in its stock form right now.

[Conor]

Originally posted by drivetrainUW-Platt:
From a parts/machining/cost standpoint, it is easier to add 2 inches to a driveshaft then to make a diff housing that has a big extrusion to ofsent the mount of the drive gear, with bigger bearings and more complixity.

If we are so worried about things being symetrical, then we need to build our own transmissions that allow for the output gear to be ofset in the ideal location for equal length halfshafts.

Our Honda ATV diff is probably only 4" wide so equal length shafts would require a complete remanufacture of the housing, which is run basically in its stock form right now.

Mike,

I don't mean to contradict or prove you wrong in any way, but we will be running a centered ATV diff next year. I'm not going to give away the exact details, but a sprocket hat will be involved. The reason I've chosen to do this is because I like the better mass distribution in the rear, and I quite simply don't like the fact of having two different sized axles. Symmetry just tends to work a little better when your correcting torque "equally" from side to side. To center the diff, the stock housing will remain intact. The only machining performed on it will be cutting off the side gear and machining an o-ring to seal it.

[Will M]

Take a look at the linked picture.
They discuss how they intentionally made unequal twist drive shafts to reduce wheel hop.

I'm not a drive train guy and a Cadillac is no race car but I thought is was an interesting idea.

http://media.caranddriver.com/images/media/51/dissected-2013-cadillac-ats-photo-441386-s-original.jpg

http://media.caranddriver.com/...41386-s-original.jpg (http://media.caranddriver.com/images/media/51/dissected-2013-cadillac-ats-photo-441386-s-original.jpg)

-William