Steering, Suspension, & Brakes

tips, technology, tools and techniques related to non-driveline mechanical components

Go to Thread: PreviousNext
Go to: Forum ListMessage ListNew TopicLog In


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Hypothetical geometry question
Posted by: Dawie
Date: April 17, 2009 09:46AM

Hi

As I tend to find myself in this sort of hypothetical technical conundrum I thought it an appropriate topic with which to introduce myself.

Consider a front suspension that combines lateral and longitudinal control arms, like that of the Rover P6, which had conventional lower wishbones but upper leading arms. Suppose in the interest of (relative) simplicity that the arms' pivot axes are perpendicular in plan. There are two obvious locations for the steering gear where bump-steer can easily be configured out, those being level with either the upper or the lower control arms. In the latter case the steering linkage would be fairly conventional; in the former the steering links would have to angle forwards or back as on an early VW Beetle, as they would rotate about a lateral axis.

My question is, what is the math to determine a correct position for the steering linkages somewhere between these extremes? Or is it indeterminate? Or simply too far from any compound of arc motions to be easily arranged with linkages?

Any thoughts?

Does anyone know of affordable (or free) software that can model linkages in 3D? I've tried an AutoCAD 3D model, but the software lacks the ability to constrain points in different objects to move together.

-Dawie


Bill Young
Bill Young
Kansas City, MO
(1337 posts)

Registered:
10/23/2007 09:23AM

Main British Car:
'73 MG Midget V6 , '59 MGA I6 2.8 GM, 4.0 Jeep

authors avatar
Re: Hypothetical geometry question
Posted by: Bill Young
Date: April 17, 2009 12:57PM

I don't know the math but from what I've gathered over the years if you design the steering so that the pivots for the steering outer links are in line with the pivots for the upper and lower control arms and run parallel you will have minimal bump steer. As in drawing A. Because of the necessity for most steering arms to be either pointed in or out in relation to the kingpin because of ackermann, then you adjust the corresponding inner pivot point to match and you still should be in good shape. Drawing B.
The rack should be a straight shot at the steering arm pivot point as seen from above also. In your problem then I'd think you'd have to look at the alignment from both planes and adjust as necessary to suit the conditions of both to get minimal bump steer.
Steering alignment.JPG


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: April 20, 2009 02:08AM

Thanks Bill. I'd got that far; but that method assumes that both control arm pivots are locatable in a transverse plane through the wheel centres. If you take a look at the example I mentioned, namely the Rover P6 (http://www.britishv8.org/Articles/Rover-P6-Design.htm), you'll see that the upper control arms are leading links. Their axes don't intersect that transverse plane and may therefore be considered of infinite length.

That aspect of the P6 has always seemed a bit upside-down to me, however by all accounts the P6 front suspension worked very well in practice. I'd have thought leading lower links would be more desirable. I've done some 2D studies that reveal much better camber variations than with a wishbone set-up, and a somewhat unusual but advantageous roll centre migration pattern. But I digress.

I'd thought the P6 situation might have the correct range of steering link positions defined by a parabolic path with its apex at the lower control arm's pivot and arranged on an axis parallel to the upper control arm pivot axis: but I cannot imagine what terms might define a useable formula. I think the Rover engineers might have been faced with the same quandary, as they plumped for steering links level with the upper control arm, and acting fore-and-aft.

I played around with the abovementioned "inverted P6" idea last night. I tried simply inverting the proportion of the distance of the steering arm joint along the height of the steering knuckle and multiplying that by the upper arm length to generate a steering link length. That is, if the steering arm joint is 1/3 of the way up the knuckle, the steering link is three times as long as the upper control arm, etc. That seemed to work, as though not geometrically spot-on it defined a workable approximation. Set up correct at ride height the deviation ranges from 0.04% off at 3" of bump to 0.6% at 3" of droop. I submit that that amount of bump-steer won't be detectable through the steering wheel.

I think the same principle was involved in the Alfasud steering gear. That had a rack and pinion over the longitudinal transaxle, and MacPherson struts. The steering arms were just below the springs' bottom cups. The steering links were consequently so long that the rack housing was blind-ended, and the links attached to bosses working through a slot at the middle of the rack.

Anyhow, I'm feeling a lot better about the prospect of getting the Minor's steering to work if the suspension is to interlink via tension rods with the rear. And chances are that I shall indeed end up using an Alfasud rack and pinion.

Best regards

Dawie


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: April 20, 2009 02:09AM

That link: [www.britishv8.org]


Bill Young
Bill Young
Kansas City, MO
(1337 posts)

Registered:
10/23/2007 09:23AM

Main British Car:
'73 MG Midget V6 , '59 MGA I6 2.8 GM, 4.0 Jeep

authors avatar
Re: Hypothetical geometry question
Posted by: Bill Young
Date: April 20, 2009 09:05AM

That type of rack is quite common on many cars using McPherson struts and FWD here in the US, although Alfasuds are somewhat rare! LOL I think it would do quite well in your application from the description you've given. Bump steer won't be perfect, but you should be able to design a fitment that will give minimum deflection throughout the suspension travel. I'm not a great math type guy and don't have the computer skills to use a CAD system so I'd probably just have to fab something up and then adjust while making measurements until I got the best results possible. A trial and error approach I admit, but about all that would be within my abilities. I did that type design once years ago on a street rod project using some home made A arms and a Fiat rack and pinion. Got lucky and wound up only having to move the outer tie rods from the top to the bottom of the steering arms to get the best bump steer and ackerman. Hope you are as lucky with your design.


Moderator
Curtis Jacobson
Portland Oregon
(4575 posts)

Registered:
10/12/2007 02:16AM

Main British Car:
71 MGBGT, Buick 215

authors avatar
Re: Hypothetical geometry question
Posted by: Moderator
Date: April 20, 2009 12:20PM

Dawie, I'm out of my depth here but I'm enjoying thinking about the points you're making.

I had to Google the term "Alfasud" to know which Alfa you guys were talking about...

At first look, the design of the Rover P6 front suspension seemed (intuitively) "wrong" to me too. My very first impression, when Bill Wardlow pulled a tire off his P6 so I could take photos, was that torque reaction when braking might be a problem. I was used to contemplating hypothetical racecar suspensions with upper and lower wishbones, except where the forward and rearward legs of the wishbones are replaced by straight individual links (with Heim joints at each end for tuning and adjustment.) In that scenario, whilst braking, the upper/forward link and the lower/rearward link are under compression. They should ideally be short and stocky to resist bending loads. The upper/rearward link and lower/forward link are in tension during braking, so they can be relatively thin and long. At first look, it seemed to me that the P6 suspension was missing one of those short, stocky links... it is, but they can omit the upper/forward link because the upper-rearward link is coming straight forward, parallel to the car's centerline. I'm rambling on... I'll just close with the point that although camber change and bump steer are important, so is stability whilst braking. (Make everything stiff, plan for bushing wear, and watch out for toe change whilst braking too.) You're probably on top of that already!

By the way... WELCOME to the BritishV8 message board! I hope you'll consider starting a "project journal" or at least putting an introductory note in the sticky thread at the top of the pub section. Direct link: [forum.britishv8.org]


BlownMGB-V8
Jim Blackwood
9406 Gunpowder Rd., Florence, KY 41042
(6468 posts)

Registered:
10/23/2007 12:59PM

Main British Car:
1971 MGB Blown,Injected,Intercooled Buick 340/AA80E/JagIRS

authors avatar
Re: Hypothetical geometry question
Posted by: BlownMGB-V8
Date: April 20, 2009 06:30PM

Very nice to get a fresh perspective. To me that entire Rover suspension is rather bizarre but apparently effective. Proof that at some point the designer refused to limit himself to conventional wisdom. It's great to see the creativity that went into it, and also very good to see someone willing to take things a step further.

So Dawie, am I to understand that what you want to do is move the steering arm downwards from it's location near the top of the vertical member and come off that member roughly 90 degrees from the existing location? How far down?

Interesting comment about lower leading links. For some odd reason a Model A Ford springs to mind, no doubt irrelevant. What would you consider the best means of providing lateral stability?

I can't really help you in your quest for a mathematic model I don't think, I tend more towards the concrete when I can. Some of the best design work gets done that way, even in this day and age. Mock it up, move it around, try it, see what works and what doesn't. It's amazing what can get overlooked either way.

But this is something I find very beneficial. Hard to get your head around, but it spawns new ideas, a very good thing.

Jim



Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: April 21, 2009 01:40AM

Thanks guys!

The intent with the "inverted P6" model is a conversion on my Morris Minor, whose steering arms are about two-thirds of the way up the steering knuckle. I'm hoping to avoid having to modify the knuckles.

The leading lower links would be wide-based wishbones, like the suspension arms on a Citroën 2CV, or the trailing arms on an early VW Beetle. In fact they might well end up being Beetle parts, but I'd have to figure out the bearing arrangement in lieu of torsion bars; and I'd have to fabricate lugs so the arms can work as bell cranks.

I also prefer working with the concrete, but for lack of having the physical parts at hand I often end up working graphically.

Dawie


BlownMGB-V8
Jim Blackwood
9406 Gunpowder Rd., Florence, KY 41042
(6468 posts)

Registered:
10/23/2007 12:59PM

Main British Car:
1971 MGB Blown,Injected,Intercooled Buick 340/AA80E/JagIRS

authors avatar
Re: Hypothetical geometry question
Posted by: BlownMGB-V8
Date: April 21, 2009 10:14AM

The one thing I can see being an issue is tire clearance at full lock, and this is really where the unequal length "A" arm setup shines because it gives maximum rigidity in a minimalist package. It's also undoubtedly the reason for the P6's leading arm being on the upper link. The Vdub rear trailing arm really doesn't give adequate clearance for a steering application, but if you have room under the engine or transmission to mount it, a transverse cross tube design might work nicely and give a great location for a sway bar also, using something like a large diameter tube running from a long and therefore rigid pivot at the frame rail forwards with a couple of angles in it allowing room for the tire and still ending up at the spindle or (steering knuckle) attachment. With this type of a lower link providing lateral and longitudinal stability the upper link could be about anything you want, A arm, unequal length links, even McPherson struts. I wouldn't get overly excited about the steering until the basic suspension is nailed down. Wherever you put it there is going to be a geometry that works well enough.

Jim


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: April 23, 2009 01:43AM

I was thinking of a VW front arm, but I don't think it'll work, as it runs inside its bearings. Because I'll need something that works as a bell crank in different planes I'll need it to be a sleeve that runs over its bearings on a stub-axle. I'll probably have to fabricate. -D


Bill Young
Bill Young
Kansas City, MO
(1337 posts)

Registered:
10/23/2007 09:23AM

Main British Car:
'73 MG Midget V6 , '59 MGA I6 2.8 GM, 4.0 Jeep

authors avatar
Re: Hypothetical geometry question
Posted by: Bill Young
Date: April 23, 2009 08:06AM

Dawie, now that I understand your project your problem becomes a lot clearer. Since you want to use the Subaru drive train why don't you just use their suspension as well? If you take careful measurements off the donor car you should be able to recreate the suspension pick up points on your frame. Then all you need to adjust would be the shock and spring rates to match the Morris. By the way, I've owned to Subarus and I love them, great cars. A good choice if you want a front mounted boxer.
If you still want to fabricate your own suspension have you just considered a forward reaching "A" arm top and bottom. If you keep the rotational axis parallel and in line with the theoretical axis you can make an A frame that reaches forwards or backwards quite a distance and still describes the same arc as one much shorter and directly inline with the spindle. The same would apply in the most part to the arc of the steering arms and allow you to position the rack further back if necessary.
A frames.JPG



Edited 1 time(s). Last edit at 04/23/2009 08:20AM by Bill Young.


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: April 23, 2009 09:36AM

Thanks Bill, but my concept for this project is somewhat different! I'm looking at using one of the Toyota W5-series gearboxes, and rear-drive only. Those 'boxes are plentiful here, and more than strong enough for anything that will physically fit in the Morris engine bay. Furthermore, the engine will be considerably further back in relation to the front wheels than in the Subaru, which would preclude awd even if I'd considered it.

More importantly, I'm hoping to engage more with Issigonis' thinking. He left a zone in the design for torsion bars: I'm saying, let's go with that but suppose they're not torsional elements but tensile elements; which opens up a whole new avenue of thought around interlinkage between front and rear, which has led me to an arrangement that has all the springing, front, rear, and roll-control, under the trunk floor. That's what got me wanting to change the lower arm pivot axis from longitudinal to lateral. The roll-centre migration and negative-camber-gain advantages are merely by the way - though that was another exploration, and sometimes things converge.

And Subarus are way too wide!

-Dawie



Edited 1 time(s). Last edit at 04/23/2009 09:38AM by Dawie.


t.lay
Tom Lay
Grayslake, IL
(93 posts)

Registered:
05/13/2008 09:53PM

Main British Car:
72 mg b gt

Re: Hypothetical geometry question
Posted by: t.lay
Date: May 13, 2009 11:30PM

You might try the McDermott Wishbone software. You can plug in various pick-up points and rack placement and see what dynamics you are introducing. You can get a free copy on the locost usa site. It's designed for unequal length a-arms, but it should work for other configurations.


Dawie
Dawie Coetzee
Cape Town, South Africa
(25 posts)

Registered:
04/17/2009 07:53AM

Main British Car:
1958 Morris Minor Subaru EJ25 (planned)

authors avatar
Re: Hypothetical geometry question
Posted by: Dawie
Date: May 14, 2009 03:33AM

Thanks Tom. I took a cursory look, and it promises to be able to handle my situation. I'll give it a try at home tonight. -D


roverman
Art Gertz
Winchester, CA.
(3188 posts)

Registered:
04/24/2009 11:02AM

Main British Car:
74' Jensen Healy, 79 Huff. GT 1, 74 MGB Lotus 907,2L

Re: Hypothetical geometry question/anti dive?
Posted by: roverman
Date: March 20, 2010 06:52PM

From a "side veiw", has anyone touched on this yet ? I can't seem to find a formula for calculating anti dive for ifs. Simplistically, it's, the rear of upper control arm will be "x" amount, lower than front,(side veiw).Since I haven't went around measuring various cars, My WAG. would be 4-6 deg. ? As for the % this would equate ? Too much should act like excessive spring rate ? It provides,"roll caster gain", in a turn,(good thing). I'm using Vette C5, front suspension clip, of which upper control arms are mounted separately. This item is not listed with regular alignment specs. Anybody ? Thanks, roverman.



Moderator
Curtis Jacobson
Portland Oregon
(4575 posts)

Registered:
10/12/2007 02:16AM

Main British Car:
71 MGBGT, Buick 215

authors avatar
Re: Hypothetical geometry question
Posted by: Moderator
Date: March 20, 2010 09:08PM

Well, the usual method of approximating anti-dive percentage geometrically is explained in one of the photo captions of this article: [www.britishracecar.com]
As the article shows, the approximation is dependent on how accurately you can estimate the center-of-gravity. You won't need much anti-dive.



Edited 1 time(s). Last edit at 03/20/2010 09:21PM by Moderator.


roverman
Art Gertz
Winchester, CA.
(3188 posts)

Registered:
04/24/2009 11:02AM

Main British Car:
74' Jensen Healy, 79 Huff. GT 1, 74 MGB Lotus 907,2L

Re: Hypothetical geometry question
Posted by: roverman
Date: March 21, 2010 03:41PM

Ok, so too much creates, "darty" and "bump nervous". C5 calls for 7.4 deg. + caster=straight line stability,(anti-darty) ? Sounds like easiest , is measure a Vette . I like concept of, roll caster gain, but at 7+ deg's at static and fat/short sidewall tires, I must be carefull ,not to un-plant the footprint. Thanks, roverman.


Sorry, only registered users may post in this forum.