Some of us have been discussing the 914 rear suspension offline and I measured my toe curve yesterday. I figured I'd share it with the club for discussion.

The axle height scale is in inches and the zero is with the arm level.

thanks for sharing!

Any idea on the front? I'd be curious to know how bad the bump steer really is with stock geometry.

I presume on that graph positive numbers are for bump and negative numbers for droop?

Sorry, I meant to explain more, but I had to run out for a sec.

Yes, along the y-axis positive is jounce and negative is rebound.... but it all depends on your starting condition. If you start nose low (the trailing arm attachments to the chassis are lower than the centerline of the wheel) then your zero might be 1 or 1.5.... if you've moved your chassis mounts higher in vehicle your starting point may be -2.

And the toe is relative, too. X-axis now... assume you start with zero toe, then look at the change in toe only.

For example: if you set your car to zero toe and the axle center line is 1 inch lower than the chassis mounts... your zero is at -1 on the y-axis. As you jounce, the rear suspension toes out first and then goes back through zero toe, then on to toe in.


I plan on measuring my front suspension, too, but it would have no relavance to a stock setup

Based on the graph it appears that your car has the axle 3" higher than the pivot when you have 0 toe and that it goes to a maximum of 1/16" of toe out when the axle drops 3" to level. I would infer that a car (such as mine) with the axle level with the pivot and 0 toe static will toe in both in jounce and rebound.

No inference required... it's right there on the chart.

Yes, level is a good compromise, but axle slightly low is better (like -0.5) so you minimize the toe change in jounce, even with a slight toe out condition. But, it still is not what you want.... zero toe change and more camber gain is more appropriate.

Chris-How did you decide level was appropriate? Do you have some measurements to share, too?

BTW... the curve changes depending on the relative heights of the inboard and outboard chassis mounting.

QUOTE (groot @ Nov 15 2005, 09:40 AM)

Chris-How did you decide level was appropriate? Do you have some measurements to share, too? BTW... the curve changes depending on the relative heights of the inboard and outboard chassis mounting.

I did raise the pickups as close to 3" as I could without being anal about it. I'm not actually sure the arms are level now - I was just saying that.
I think the axles might be up to 1/2" lower but I would have to measure when I have the car on a good surface. At that point I can take some other measurements as well.

One of my next projects may be to incline the pivots to lower the roll center, and move the bearing carrier to restore the static camber. I would be interested in seeing the toe and camber curves for such a change.

You will increase the scrub by changing the pivot point angles. Have you measured the amount scrub in the stock setup? Have you measured it in your current setup?

Do you know where the roll centers are now and where they go in bump and roll? What is your cg height? Do you know what happens when you increase the distance between the two?

You still have to fix the toe change issues with the rear suspension. Kevin is looking to solve the problems.


Heard the other day Finch is thinking about coming back to play in FP. You can bet he has analyzed all of these factors.

wow...
great thread Kevin!
I'm going to have to do some poking around on the race car now.

brant

QUOTE (Brett W @ Nov 15 2005, 08:50 PM)

Heard the other day Finch is thinking about coming back to play in FP. You can bet he has analyzed all of these factors.

All he has to do is snap his fingers to produce the data he needs, lol.

I met with Finch the other day for a few hours and had a really good discussion.

He believes, as I do, that a 4cyl can win in EP and FP (duh!!). So, if he runs the 914, I wouldn't be surprised to see him in E or FP.

We talked a lot about the 914 suspension... and he thinks the suspension is good enough to win without serious redesign required. He has tweaked the 914 suspension on that car, but nothing too drastic.

I think I need all the help I can get because I know I'm not the best driver, so I'm going crazy on the suspension.

QUOTE (groot @ Nov 16 2005, 09:26 AM)

I met with Finch the other day for a few hours and had a really good discussion. He believes, as I do, that a 4cyl can win in EP and FP (duh!!). So, if he runs the 914, I wouldn't be surprised to see him in E or FP. We talked a lot about the 914 suspension... and he thinks the suspension is good enough to win without serious redesign required. He has tweaked the 914 suspension on that car, but nothing too drastic. I think I need all the help I can get because I know I'm not the best driver, so I'm going crazy on the suspension.

Since Mark H. set a lap record and beat Sargis handily I wouldn't be surprised if the CRB gives the FP car a lead trophy or other competition adjustment.
That will increase the value of modifying the suspension.

Good point!!

Let's hope there are no lead trophies handed out.

QUOTE (Racer Chris @ Nov 15 2005, 03:04 PM)

One of my next projects may be to incline the pivots to lower the roll center, and move the bearing carrier to restore the static camber. I would be interested in seeing the toe and camber curves for such a change.

I think this is valuable if you're interested in making the production semi-trailing arm work. I was going to make the curves for differences in the height of the outboard attachment point, but I ran out of time.... travel for work (and the new child) makes it hard to work on the car.

I probably will not measure it because I'm abondoning the production arm. When someone does measure it, please post it. I'm interested in discussing it.

BTW... you don't need a good surface since you don't really care what the zero is... you're just looking for trends and comparative numbers.

QUOTE (groot @ Nov 16 2005, 04:21 PM)

I'm abondoning the production arm

what are you gonna do/build?

QUOTE

Since Mark H. set a lap record and beat Sargis handily I wouldn't be surprised if the CRB gives the FP car a lead trophy or other competition adjustment

Hence my main complaint with the SCCA. kinda like the US tax system, they penalize you for hardwork and acheivement. The will up the weight or something or shrink the carb size. Just what the FP guys need. Lets hope they don't change the EP rules also.

QUOTE (Jeroen @ Nov 16 2005, 07:28 AM)

QUOTE (groot @ Nov 16 2005, 04:21 PM) I'm abondoning the production arm what are you gonna do/build?

By rule, I have to stay with a semi-trailing arm suspension. So, I'm going to build a new one. Brett was on this path a while ago and I resisted, but looking at the geometry and what I need (more rear grip), there's no way around it..... for me.

It sucks, too, because I built and rebuilt this rear a few times now and I'm going to trash it again. But, I like a challenge and this rear suspension is one.

I prevail. Yes the stock trailing arm works but it is far from optimal in a racing situation. A custom built arm is just part of it.

QUOTE (groot @ Nov 15 2005, 09:40 AM)

Do you have some measurements to share, too? BTW... the curve changes depending on the relative heights of the inboard and outboard chassis mounting.

I took a trailing arm to the shop today. I'm going to fixture it and make camber/toe measurements. Then I can generate 3D plots like the one above, with pivot inclination and axle height as the independent variables.

Now this is what I am talking about. Finches car (Mr Kirby car) is very stiff, probably in the 400lbs per corner area for spring rate. I believe that to be stiffer than any of us are running.

Erik

If you are building a competetive EP car you should be shooting for 10-15K per degree for chassis stiffness. Most production chassises will have a tough time reaching these numbers with out a properly engineered chassis. Just adding a cage won't do it. From teh cages I have seen many are severyly lacking.

QUOTE (914_Guy @ Nov 18 2005, 06:16 AM)

In theory, should the toe and camber change at all for the rear? Maybe you were measuring the change in rear track?

In theory, the toe can change a maximum of 12.5 degrees, (if you could rotate the trailing arm through a full 90 degrees).

this pic claims that the angle of the pivot is 25deg
maybe that info is usefull making calculations (not me... this is still way over my head )

QUOTE (Jeroen @ Nov 18 2005, 07:09 AM)

this pic claims that the angle of the pivot is 25deg

I think it's only half that.

hmmm... looks like your right
just did a crude measurement off the pic in graphix software and it looks like 14deg
nevermind... I'll just shut up now and listen

The factory angle is 12 degrees. In my research it looks like the best compromise of all the available angles. BMW offered two different options for the M3 on was 12 deg and the other was 15 deg. They kept there race stuff at 15 degrees, but the toe curve is not as nice for the 15 deg angle. It does gain more camber.

QUOTE (914_Guy @ Nov 18 2005, 09:43 AM)

Now the x-axis will correspond to the trailing arm’s pivot axis and as the arm rotates it moves only in a plane corresponding to the y-z plane and therefore its camber angle never changes.

I dont agree with that statement. I dont think the axis of the trailing arm is parallel to anything (either the x, y, or z axis).

Remember, camber change is a fuction of instant center. The instant center is fixed on a semi-trailing arm, but that just means the camber change is linear and not exponential.

It has nothing to do with orientation and everything to do with Instant Center. Its a geometric condition. Where you orient your origin is arbitrary.

This is a good site to understand what is happening, although not exactly the numbers we are dealing with, very close.

http://e30m3performance.com/tech_articles/...nter/page-2.htm

What we are considering is obtaining a little bit more camber gain, and lessen or eliminating the toe change. In some of our cars, and everyone is running something a little bit different, Roll center MAY be TOO high. In others there is a discussion on how we can change the trailing arm to deal with some of this. There are options, but they need to be researched, not just guessed. This is what Kevin (primariliy) is doing. others are as well.

Brett,

I was talking about Finches spring stiffness in the actual springs. He is running the car alot stifffer than many of us with regard to spring rate. I would say your numbers for platform stiffness are very close. I don't think I am quite there yet. Hows the Ultimate 914 coming??

Erik

Can you try that link again? It doesn't work...

Thanks!

-Josh2

Its kinda on hold until I find some more shop space.

QUOTE (jhadler @ Nov 18 2005, 06:25 PM)

Can you try that link again? It doesn't work...

Here

QUOTE (Thorshammer @ Nov 18 2005, 06:19 PM)

He is running the car alot stifffer than many of us with regard to spring rate.

He also has several more bars in the roll cage than we do.

The math is given at the website link I provided.

QUOTE

Does anyone have a book on suspensions that states trailing arms like the 914’s do change camber and toe as the arm pivots?

You don't need math you just need a tape measure and some string, and maybe a camber gauge. Supoort the body on jack stands and pull the shock loose. Set the trailing arm at a level position. Then take a really big square and stand it up next to the wheel. Measure the distance from the front of the wheel to the square and measure the top of the wheel to the square. Now cycle the trailing arm as far as it will go into bump and re-measure. You will see the toe and camber change. You don't really need math to see that.

But if you must see the math.

If you plot toe or camber as a function of rotation you get a sine wave, assuming you could rotate about the pivot axis through 360 degrees.

The pivot axis isn't parallel to the direction of travel. Your graph does not take this into account.

Here's an interesting article... that contradicts Milliken's book.

Basically, it says that toe is a complex curve that can toe out (proven by my chart)... Milliken states that a semi-trailing arm never toes out. Camber is linear over our range.

BTW... BMW uses 12 degrees on their early race M3s and used 15 degrees on their street cars.

Article has been removed from this post... its text is in a later post.

Eric.... yes, I'm finally on my home computer to share this file. Sorry it took so long.

Table to go with text.

Pretty charts

Bigger than I expected... sorry.

Last purdy chart

kevin, if you e-mail me the .doc file you tried to post, I'll upload it to the site
just send it to jeroen@berloth.nl

Or I could just post the text.... duh....

Car Suspension and Handling, Second Edition
Written by Donald Bastow
Penntech Press


6.9 TRAILING AND SEMI-TRAILING ARMS

The fundamental difference between trailing and semi-trailing arms is that the axis of the former is at right angles to the car centre line; this implies that there is no change, in end view, of the angle of the wheels with suspension movements. The semi-trailing arm, so far confined to examples where in plan view the axis of oscillation of the arm meets the vertical plane of the wheel axes towards or beyond the other rear wheel, necessarily introduces some swing axle effect. Although it is generally considered that the instantaneous centres are where the arm axis intersects the vertical planes of the rear wheel axes across the car and fore and aft, this is not strictly true. The actual path of any point on the wheel, in side or end elevation, is an ellipse and the instantaneous centre for that part of the ellipse can be found by known methods.

The knowledge that roll-steer effects are at a minimum when the semi-trailing arm axis is parallel to the ground is useful but we need to know more. For this we must study how toe-in changes on bump and rebound vary with changes in position of the semi-trailing arm axis, in plan view and in the angle which the axis makes with the ground. A basic case has been chosen in which the semi-trailing arm axis is at road wheel centre height, meets the rear wheel axis of the car at one wheel plane, point 2, and intersects the plane of the wheel whose movements are being studied 400 mm forward of its centre, point 1. Figure 6.19 shows this. An alternative plan position shows the semi -trailing arm axis meeting the rear wheel axis at twice the car track from the wheel being studied, point 2A. Points 3, the wheel centre, and 4, 100 mm away along its axis, are for calculation purposes to obtain the toe-in changes. The heights of points 1 and 2 in Table 6.1 define the axis changes studied, in height and inclination to the ground: rising to the front, dropping to the front, high or low; and parallel to the ground but above or below the original position. The resulting roll centre heights are also shown; they indicate the amount of sideways 'scrub' of the contact patch with bump and rebound movements.

Anthony Best Dynamics Ltd. has kindly computed the results and its co-operation is gratefully acknowledged. The results are summarised in Fig. 6.20(a). Figure 6.20(B) shows the camber changes.

Fig. 6'20 (a) Here the results of the calculations in terms of change of toe-in with bump and rebound movement are shown. There are effectively no differences between cases 1, 7 and 8 where the axes are parallel to the ground but at different heights. Cases 2 and 4, axes up towards the front but at different heights, are also effectively the same as each other but now favour "bump at the expense of rebound. The converse applies to cases 3 and 5, where rebound is favoured. Case 6, axis parallel to the ground and more nearly so to the rear wheel axis, halves the toe-in changes.

Fig. 6.20 (B) The camber changes depend only on the distance from the affected wheel to the intersection of the axis and the vertical transverse plane containing the wheel axes

Within the limits studied, i.e. axis height::l: 20 mm parallel to the ground, and a height change of 20 mm in sloping axes, nose high and nose low, and bump and rebound movements each of 75 mm, axis height does not affect the results (to two significant figures). The nose high axis position favours bump movements, and vice versa. Camber changes depend only on the distance from the studied wheel to the intersection of the semi-trailing arm axis, point 2, with the transverse vertical plane containing the undefIected rear wheel axes. From the trends within the range studied, we see that a sufficiently large nose-high axis angle wilI give toe-out on bump and toe-in on rebound; and again vice versa.
If the distance between wheel arches is important it must be remembered that the introduction of any swing axle effect has to be accompanied by an increase in the track to maintain that distance between wheel arches.

That is some impressive math and I am sure the results are correct for the input you used, but the inputs are wrong. The trailing arm pivots in a different axis. You also don't account for anything other than a single plane view. The semi trailing arm axis is located in a 3D plane. It will change toe because the trailing arm moves in an arc in two planes. Does the suspension based on your numbers gain camber? If so how can it change toe?

I don't understand what the 3d surface plots represent since you haven't labeled the axes. Also, since the units on each axis change from graph to graph, any trend is distorted.
I think it would be more meaningful to plot camber and toe independently, as a function of trailing arm rotation using 2d coordinates. By plotting them simultaneously on one graph their interdependence would be apparent.

I agree seperate them so that they aer more easily read and a 2D curve will suffice for each one.

Yes you will not see large changes in either, because that just doesn't happen. The semi trailing arm does not make big changes, but the little changes will make all kinds of problems in a race situation. When your suspension travel is limited to 1-2 inches in both bump and droop, small variations will become big problems. When your chassis is setup to take advantage of 20K lb/degree of twist, and your suspension is setup so that there is no uncontrolled movement, any suspension movement that is not controlled keeps you out of the winners circle. When your chassis is so stiff that your race car setup changes each time you change tires you know you have done it right.

Whats the x axis represent?
Can't you convert your graph y axis labels to degrees?