Bushing compliance vs. kinematic wheel recession for reducing impact harshness

[NoahLKatz]

Is it possible to equate a given amount of kinematic wheel recession to control arm bushing horizontal compliance?

Assuming double wishbones and LCA canted down at the back, is it as simple as taking the horizontal component of the suspension spring compliance?

I'm guessing not, since the ratio of horizontal to vertical force would vary depending on the shape of the bump.

Does it help any if additionally the UCA is parallel to the LCA (which would eliminate some of the bump castor)?

This is for a light (~900 lb) reverse trike similar to a Morgan 3 Wheeler.

I'm thinking this would work OK since it's acceptable (or at accepted) on motorcycles, whose front forks are pro-dive, and bikes have much greater weight transfer with their high CG's and short wheelbases.

 

[BrianPetersen]

It isn't that simple - it's normally done via compliance, not kinematics. On automotive suspensions, the LCA chassis-end pivot axis is normally pretty close to horizontal, and that ball joint is normally reasonably close to the centerline of the front wheel, which means kinematic fore-aft movement of the wheel with suspension movement is pretty small (although not zero). The UCA chassis-end pivot axis can be sloped down towards the rear, to produce an anti-dive effect. Yes, it means the caster changes some with suspension movement, and that means the trail changes, etc etc.

What's actually done, is that on many if not most modern suspension designs, doesn't matter MacPherson or wishbones, the lower control arm is roughly L-shaped, with one chassis-end bushing of that L being pretty close to directly cross-car with the lower ball joint and with a really stiff "handling" bushing, and the other branch of that L being significantly rearward and with a fairly soft "ride" bushing. The steering tie-rod is very close to parallel to the cross-car part of that L, so that even though the soft bushing on the rearward part of the L allows the entire "L" to pivot around the stiff front bushing, it doesn't affect the toe (much) because the stiff-bushing cross-car part of the L is parallelogram with the steering tie-rod. This allows the whole spindle and front wheel to move fore-and-aft a little (based on deflection of that soft "ride" bushing on the rear branch of the L) without screwing up the steering as this happens.

Examples abound. VW has done this for decades. My Fiat is like this. I know there are lots of others.

The concept isn't new. Take a look at the traditional Chrysler torsion-bar front suspension that has been around since the 1950s - 1960s. The main lower control arm is just a lateral link (not an "A") and the "handling" bushing is on the chassis end of that. There is a trailing link diagonally from the chassis to near the ball-joint end of the lower control arm which locates the ball-joint end of the arm fore-aft, but it has fairly soft bushings.

 

[NoahLKatz]

Very interesting, thanks, I'll see if I can google that or make a sketch that fits the description.

 

[GregLocock]

The rear suspension of some production cars often has kinematic recession in bounce. For instance, rear twist beam suspensions . Therefore the only hope for acceptable impact harshness is compliance, so that as the wheel moves up and forward on the kinematics, it can also move rearwards on the compliance. I'm sure there are studies done on this trade-off, a rule of thumb from many decades ago is that a wheel centre recession rate of 300 N/mm was an upper limit.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[NoahLKatz]

> The rear suspension of some production cars often has kinematic recession in bounce... For instance, rear twist beam suspensions . Therefore the only hope for acceptable impact harshness is compliance

Do I have the terminology wrong?

I thought recession meant the wheel moves backward as the wheel moves up in bounce, so that it is in effect compliance.

 

[GregLocock]

No, I had my spelling wrong. Precession. The measurement is called recession in bounce, when negative, which is what twist beams have, is precession.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[BrianPetersen]

Here's the best image I can find. It's from under the car looking towards the left (i.e. right on photo is front of car).

https://bparts-eu.s3-eu-west-1.amazonaws.com/ftp/azor_2/000/000/751/000000751981.jpeg

Note how the steering arm is quite close to parallel with the cross-car portion of the L-shaped lower arm. The bushing on that leg of the arm is stiff so that the steering arm and that leg of the lower arm are a parallelogram. The bushing inside that aluminium bracket is squishy to let that parallelogram work fore-and-aft a little when there's a bump impact.

 

[NoahLKatz]

Thanks for digging up the nice pic.

So if I'm understanding this correctly, when a bump tries to push the lower ball joint back, the arm pivots around the stiff bushing at the upper right.

Given that the angle is roughly 45 deg from the BJ to the upper left bushing, the latter would have to be allow as much radial as axial deflection, so presumably this car is aimed more at ride comfort than handling precision.

Still wondering about my original question - compared to the above strategy, how effective is tilting the wishbones back, say, 20 deg?

 

[GregLocock]

In theory it makes no odds if your wishbones are symmetrical about the axle in plan view, ^ arms, or if it is an L arm. In practice once you consider frame stiffness and so on then for a lightly built car where refinement is a secondary priority, symmetrical ^ arms make more sense usually. Small change like your 20 degrees will be lost in the weeds of practicality.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[SwinnyGG]

Given that the angle is roughly 45 deg from the BJ to the upper left bushing, the latter would have to be allow as much radial as axial deflection, so presumably this car is aimed more at ride comfort than handling precision.

Little bit of confusion here I think. The bushing on the left in that photo is under basically zero axial load; all load on this bushing is radial (up or down in the photo; near-zero force left/right). BMW used this arrangement in the front for decades, specifically through the 3-series until the E92, and that series of cars were known as some of the best handling cars ever produced in their class.

In the BMW version (pic below) the horizontal bushing in the photo Brian shared is not a bushing- it's a rather large ball joint, which provides articulation of the arm in all directions but zero longitudinal movement of the arm, which means the bushing to the rear (in the BMW system, the only bushing in the system) sees radial load only.

 

[NoahLKatz]

Greg,

> In practice once you consider frame stiffness and so on then for a lightly built car where refinement is a secondary priority, symmetrical ^ arms make more sense usually.

Refinement is *not* a secondary priority.

Using a motorcycle analogy, my aim is to build a comfortable but agile cruiser, not a sportbike.

I'm not looking for a pillowy ride, but I have a thing about impact harshness; it makes it seem like the car crashes into bumps (not so much the feeling but the sound transmitted throughout the body*) instead of rolling over them.

The latter is what I experienced in 70's VW Bugs and my brother's Porsche 356, which as you know have trailing arm front suspension that give wheel regression.

As far as frame stiffness, not sure what you're driving at, but I have the luxury and intention of designing the chassis structure with much less compromise than car mfgr's.


> Small change like your 20 degrees will be lost in the weeds of practicality.

I don't believe Bugs' trailing arm angle is much more than that, or am I mistaken to think that wishbones with the same regression rate are as effective as trailing arms?


* I don't know if I'm unusual in this respect, but here's an example of how this affects my perception of suspension operation:

My second car was a 1959 Austin Healey 100-6.

I initially thought it had a horrible rock-hard suspension because of all of the crashing sounds on bumps.

I then discovered that the gas tank was loose, and when I tightened the straps I was shocked at the difference.

No crashing sounds and I could actually feel the suspension working, which I actually enjoy, but that perception had been totally drowned out by the auditory sensory overload.

 

[GregLocock]

If refinement is important and you want to use a double wishbone suspension then L arms are the way to go. The problem with ^ arms is that both the bodyside bushes get involved in everything, for instance to get a low recession rate they need to be soft in vehicle X and Y, yet for handling they must be stiff in Y. For braking they must be able to take large loads in X and Y, needing snubbing.

Whereas with an L arm all the NVH bush has to do is to control the compliant response, again it'll need snubbing in Y for braking, but steering, lateral loads, and a good bit of braking, are handled in the body side bush in line with the axle.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[NoahLKatz]

Speaking of Bug's and 356's, anyone know their ride frequencies? That's what I'd like to target.

 

[NoahLKatz]

Swinny,

Thanks for straightening that out.

I remember having a short ride on city streets in an in-law's BMW, was new about a dozen years ago so likely what we're talking about.

Dunno about the handling, but I remember greatly disliking all the noise on bumps and dips.


I wonder if I'm spinning my wheels unnecessarily here, as there's a big thing in my favor as far as the noise aspect goes - the trike will have no roof or windows, so a lot less acoustic radiation area and less solid angle to contain the sound.

Does anyone have numbers for how much accelerations felt by occupants are actually reduced by such as the L-arm discussed above?

Or just a general idea of how much of the impact harshness NVH problem is due to N and how much to V?

 

[NoahLKatz]

As for kinematic recession, I guess no one has anything to say because it just isn't used anymore...right?

 

[BrianPetersen]

"How much NVH gets soaked up" is going to depend (and vary widely) on the factors in individual applications. A luxury car could very well have a softer NVH bushing than a sporty version of the same vehicle platform. Don't neglect the possibility that BMW intentionally chose to make the ride firm on the vehicle that you had a ride in, to make it feel sporty (the common customer associates a sporty feeling with a firm ride). I know of one current-production vehicle in which the run-of-the-mill versions have rubber isolators between the rear subframe and the body, but a certain high-performance variation of the same vehicle has aluminum spacers in those spots. I do not think the aluminum spacers make the car any faster, but they sure make it rougher.

And ... all this calibration tends to be tightly-held proprietary information. I just see how the bits and pieces go together. I don't see what went into designing the various bits and pieces in certain ways. It's quite evident that a lot of those bushings have a ton of calibration work put into their designs - they're often asymmetrical (even though they're fitting into a round hole and have a circle in the middle for a bolt to go through - but they have to be installed in a certain orientation).

It's true that motorcycle suspension designs do allow the wheel to move back together with moving upward (and this is both front and rear, due to the swingarm down-angle in the rear) but NVH control isn't the objective there, on any bike that I have paid attention to. NVH is the last thing on my mind for my roadrace bike. Soft bushings are no bueno in motorcycle suspensions.

No modern automotive suspension design that I know of, is designed for significant kinematic aft motion with suspension compression. I know about the VW air-cooled double-trailing-arm design (which was designed in the 1930s). The bad side effects of that suspension design outweigh whatever advantage it might have in some areas. Your idea of angling the body-side pivot axes significantly down to the rear is interesting but packaging has to be accounted for. I suppose you could use the old Chrysler design with a single-lateral-link lower "arm" restrained by a diagonal trailing link, and just move the chassis (front) end of that trailing link upward. With the more modern layouts, with the L-shaped lower arm, the L is almost always towards the back, and angling it down would point it towards the ground ... not good. Having it point forward and up might work if you don't have powertrain or steering in the way, and if the powertrain is transverse, you probably will have powertrain in the way.

MacPherson + angling the lower arm that way = "pro-dive" in braking (with practical steering geometries). With upper and lower A-arms, the upper arm would need its pivot axis even further steeply angled if you want any anti-dive.

And for all this ... the damper calibration is going to matter 10 times more for NVH than any of this.

If you really want good NVH then study various Citroen, Renault, and British Leyland designs of the past. Hydrolastic was a remarkable invention.

 

[NoahLKatz]

OK, I need to spend some time on laying out components and see what packaging issues emerge.

Very interesting about bushings; are there any that have a reasonable high ratio of axial to radial compliance? Though hard to think how to do that without sliding surfaces.

> If you really want good NVH then study various Citroen, Renault, and British Leyland designs of the past. Hydrolastic was a remarkable invention.

Back in the 70's my cousin had a DS21.

It did float along like a well-damped marshmallow as far as larger disturbances were concerned, but I was quite disappointed that it was no different in noise or feel when it came to roadway texture/surface roughness.

 

[GregLocock]

I spent a fair number of years looking at the Opel Omega (80s era) chassis. I had a set of aluminum pucks to replace the rear subframe bushes. Although harshness was worse, the improvement in steering feel, and shake, was marked.

So far as whether kinematic recession is used, it comes fairly low in the priorities , ie nice to have but not worth dying on a hill for. Since, demonstrably, rear twistbeams deliver adequate NVH, with precession, then yes, it may drive more pain into the rest of the subsystem, but not to an unreasonable extent. I must confess my most recent (and probably last) clean sheet of paper rear suspension ignored recession holus bolus, because reasons.


Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[NoahLKatz]

No worries there, single rear wheel on a trailing arm.

 

[3DDave]

In a thought experiment - a suspension where there is only aft movement allowed, seems like that characteristic isn't so good.
A suspension with only vertical movement allowed, seems like that would be fine.

Has anyone made a suspension that has only linear movement but the movement is not normal to the nominal ground?