Springs - coil vs. torsion

[eightballracing]

Hi, as an oval racer and FSAE alum I have been curious as to why a lot of racing teams (professional, amateur, and FSAE) are switching to torsion bar style suspensions. I noticed at FSAE-M this year that Graz is running torsion bar suspensions and schools like Western Australia have been doing it for a few years now. Of course many LeMans and F1 teams have been doing it for a while now as well.

In the short track oval racing world I have always heard of people running transverse torsion bars on sprint cars, midgets, supermodifieds and the like because it lengthens the "spring base" and is supposed to slow the reaction of the suspension. Through all the research I have done in Miliken, Tune to Win, and a few other sources that the term "spring base" has never come up. I get the feeling it is somewhat like the oval term "bite" which means nothing to most in the vehicle dynamics world.

I'm going to guess that this slowed reaction of a torsion bar compared to a coil spring has something to do with the inertia of the torsion bar and arm package. With a greater inertia about the rotating axis of the torsion bar it might react slower than a coil spring of the same rate? Will this allow you to get slower, more predictable motions without having to use stupidly stiff rebound settings on the dampers?

With our FSAE car this past year we stuck with the coil springs without giving much thought to torsion mostly because a coil spring is so much easier to package. It's also much easier to change out when you want to make spring rate changes while testing.

Any other thoughts on this matter?

-Pete

[eightballracing]

Hi, as an oval racer and FSAE alum I have been curious as to why a lot of racing teams (professional, amateur, and FSAE) are switching to torsion bar style suspensions. I noticed at FSAE-M this year that Graz is running torsion bar suspensions and schools like Western Australia have been doing it for a few years now. Of course many LeMans and F1 teams have been doing it for a while now as well.

In the short track oval racing world I have always heard of people running transverse torsion bars on sprint cars, midgets, supermodifieds and the like because it lengthens the "spring base" and is supposed to slow the reaction of the suspension. Through all the research I have done in Miliken, Tune to Win, and a few other sources that the term "spring base" has never come up. I get the feeling it is somewhat like the oval term "bite" which means nothing to most in the vehicle dynamics world.

I'm going to guess that this slowed reaction of a torsion bar compared to a coil spring has something to do with the inertia of the torsion bar and arm package. With a greater inertia about the rotating axis of the torsion bar it might react slower than a coil spring of the same rate? Will this allow you to get slower, more predictable motions without having to use stupidly stiff rebound settings on the dampers?

With our FSAE car this past year we stuck with the coil springs without giving much thought to torsion mostly because a coil spring is so much easier to package. It's also much easier to change out when you want to make spring rate changes while testing.

Any other thoughts on this matter?

-Pete

[exFSAE]

Still plenty of pro open wheel and stock car classes that use coil springs. As for UWA, they may not be a great example as their suspension setup is much different from the 'typical.'

I've never heard of "spring base" or changing suspension "reaction speed" and it honestly kinda sounds like BS to me. I could certainly be wrong. As for forward and side bite.. I wouldn't say those are unheard of in the VehDyn world. Most race engineers at least in the US would probably be familiar with it.

One nice thing about torsion bar springs is you can fabricate them fairly easily and in fine increments of rate.. as opposed to having custom springs wound.

[Adambomb]

We wanted to run torsion bars a couple years ago, but after several weeks fighting with packaging were forced to give up and go with coils. One reason to use torsion bars over coils is to eliminate bending loads on the shocks.

As for "reduced spring base," if I were to take a completely wild guess it would be that the "sprung" portion of your chassis could be a lot shorter if you package the bars longitudinally, so the "sprung" ends are close to each other. Although going through a free-body diagram that doesn't tell the whole story, as even though the "sprung" ends of the torsion bars would be closer together, your pullrod/rocker forces would still be acting at the same point. Just a guess though.

As for "reaction speed," first I'd like to consider the possibility that your torsion bars have enough inertia to affect their motion. Then you've got an additional coupled second order system wobbling in there between your sprung and unsprung mass. Not something I would consider desirable.

[Richard Pare]

Torsion bars are used in F1 etc because they can be mounted in a manner that produces no friction, whereas a coil spring over a shock body always produces a side load on the shock, in turn creating undesirable extra friction.

Torsion bars on sprinters and midgets are done for simplicity sake and cost.

The bars for F1, however, are extremely expensive and very short-lived - 2 hours use and they are thrown away.

[Adambomb]

I can imagine that, one other problem I had with packaging them was making them large enough to have some sort of reasonable fatigue life!

One other thought on this whole "reaction speed" thing: I'm sure most people here have seen the videos of engine valve springs at high speeds where you can really see inertial effects. I recollect some '60s or '70s vintage high-revving engine, thinking an early Honda F1 or motorcycle engine, that used torsion bar valve springs. Very interesting.

[billywight]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">One other thought on this whole "reaction speed" thing: I'm sure most people here have seen the videos of engine valve springs at high speeds where you can really see inertial effects. </div></BLOCKQUOTE>

The difference in shaft speed between valve springs and suspension is huge. You would be suprised at how slow the max shaft speed is on a typical shock if you don't already know it.

[Richard Pare]

There actually is a difference in inertial considerations between a torsion bar and a coil spring. In both cases, you are twisting the wire, but only with the coil spring are you moving the (almost) entire wire length longitudinally. That extra inertial force, while fairly small, acts as extra dampening of the suspension system.

On valve springs, it is the inertia of the coil mass that causes the "surge" you see in the high speed videos. I just received a patent on a means to dampen that surge, and am starting to play with the pieces to get a video of just how effective it is. It will be interesting to see what that opens up by way of changing how racing valve springs (and the valve stems, since it also decreases the torque load seen at the keeper) are designed.

[Peter7307]

"I recollect some '60s or '70s vintage high-revving engine, thinking an early Honda F1 or motorcycle engine, that used torsion bar valve springs. Very interesting."

Honda used torsion bar valve springs in their F1 cars of around 1964. The engine was a liquid cooled V12 and was mounted transversely.

The motorcycle engine was an aircooled vertical twin of 450 cc and also used torsion bars.

Honda's logic was the torsion bar moved the load path of the valve spring and allowed the inlet and exhaust tracts to be better placed for breathing.

Peter.

[Neil_Roberts]

A coil spring is actually a torsion spring that is wound into a helix so that it doesn't need torque transfer features on its ends. If the cross sections of both spring types are either solid or hollow, there is very little difference in energy storage efficiency between them.

For suspension use, the major advantage of torsion bars compared to coil springs is that torsion bars are usually hollow, and are therefore more structurally efficient than coil springs. That is because in a hollow cross section spring, no material on or near the section neutral axis. The secondary advantage that torsion springs have, particularly for off road use, is that the volume required to package a long-stroke spring can be much smaller than coil springs.

The downside to torsion springs is that each spring requires two stiff and strong support structures in weird locations on the car, and one of them has to be stiff and strong in the torsion direction. That is usually a real challenge for the designer. Some production cars have two concentric torsion springs so that only one chassis mount is required.

There is a direct analogy between torsion or coil suspension springs and solid vs. hollow halfshafts. Hollow ones are always more expensive, and are always lighter than solid ones.

I assume that the reason coil springs are always wound from solid wire is so that wall buckling is never an issue during the forming process.

[Richard Pare]

The main reason for solid wire springs is the availability of proper material. Currently, the material for the best racing springs is H11, while expensive, is at least available in the proper form and heat treat.

Hyperco did make some hollow tube springs on a custom basis for a while, but the material available did not allow matching the H11 in weight, stroke length efficiency, fatigue life, etc.

[brettd]

A couple more reasons for torsion bars:

While the hollow ones are even more structurally efficient than solid (however difficult to machine they may be), a solid torsion bar is still more efficient than a coil spring because it is not loaded in combined bending/torsion like a coil spring is.

Another major reason is the packaging allows for mono-tube dampers, which depending on what you want (and if you're making them) can pretty much make the decision for you itself.

Regarding the downside of a second stiff mount (the first one is part of the rocker/bell crank which you have to have anyway), remember that it only has to be stiff in torsion, which if you think about it is very easy to mount as stiff or soft as you want. Even then, because it is only a spring mount (no damper), as long as you know the stiffness there, the mount can effectively become part of the spring, so there usually still is a weight advantage to coilovers in FSAE.

[eightballracing]

Some interesting information guys, and a lot of things I never really thought of. I guess it all goes back to the idea that there is no "right" design for an FSAE car. I guess I never thought much about the bending load on the shocks introduced by coil-over springs but now that I think of it, it can be significant.

I still prefer coils for the ease of packaging, adjustability at the track, and fatigue life over a torsion bar. On the other hand it is easier to change just the dampers on a torsion bar setup without changing ride height settings and spring pre-load.

More than anything though was how I always hear the supermodified and midget guys talking about how a torsion suspension "reacts slower" than coil over suspensions. I don't really get what sort of laws of vehicle dynamics would govern this if the wheel rate was the same for both types since the reaction rate is controlled by the dampers (ignoring the very important tire effects here, also assumed the same for both coil and torsion). I've always thought it was a bunch of BS.

[Richard Pare]

The "react slower" that they speak about is most likely the difference in suspension friction - that friction does indeed transfer weight more quickly.

The frictions produced by the springs on coilovers can be reduced substantially by using the Hyperco/ICP hydraulic spring platforms (disclosure: I invented them, hold the patents, and manufacture all of them), which are now standard equipment on just about every top level (and most mid to low level) cars out there.