TA22 + rack and pinion

[LeeRoy]

heya Rob,
i had thought a little toe in on bump steer is better?

ie, if you are mid corner and hit a bump, adn get toe out, then the front of car turns toward outside of corner...
if it is toe in, then the tyre temporarily increases understeer, but keeps going same direction?

or am i missing somehting vital?

Spot on.

Under braking which makes the front end of the car dip it also provides toe in and increases braking stability.

Dimensions wise are the power steer and non power steer racks the same?

[stidnam]

So I sliced up a tie rod end tonight and got some dimensions. I have a question on what I use as a reference point for the measurements though. Look at pic 1 the top blue mark is where the steering knuckle bottom sits on the tie rod end once tightened. Do I take the centre of the main tube and add to the the distance to the mark ie 12mm + 10mm to give me the centre point or do I do as per picture 2 where I am using the centre point of the ball in the join so that the slight bend in the housing is accounted for?

This only makes 3 mm of difference I know but I am interested to know which one to use.

[LeeRoy]

What are you using to draw your pics? They shit on mine

[stidnam]

Paint.Net

Its a free mod of MS paint with some extra stuff that is in Photoshop and the like.

[oldcorollas]

or do I do as per picture 2 where I am using the centre point of the ball in the join so that the slight bend in the housing is accounted for?

jointed systems don't care what shape they are... only where they are connected...

in this case, you would take the centre of the pivot, since that is what controls the movement.

that means that the bolt part of the tie rod should be considered as part of the steering arm... and the socket part as the actual tie rod...

maybe?

[stidnam]

I was thinking that #2 was the one to use.

Unfortunately the housing has already been shortened to the old incorrect dimensions. Lucky I can fix this by chopping more off it.

[o_man_ra23]

Yep, second is the one to use.

Will remember to err on the toe in side when I do mine.

[stidnam]

I have drawn all this out on paper and measured the distance from the middle pivot point of the tie rod end and the middle steering arm hole for the ball joint. I drew this out on an enlarged scale which was probably a bad idea as it will exaggerate any slight drawing errors such as fractions of an angle etc.

There is a surprising effect that this angle makes. Instead of the offset just being 16.2mm the distance is now 26.5mm. So looking back at my calcs instead of it being

Crossmember pivot point distance - steering arm offset distance x 2 = overall length that shortened rack needs to be between pivot points.
or 548 - (16.2 x 2) = 515.6mm

Now as the crush washer and the inner rack ends wont change the rack is the only thing that can be shortened so 598mm - 515.6mm = 82.4mm that needs to be removed from the rack. I have rounded this to 82.5mm as I can measure acurately down to 0.1mm with calipers but can't mark it that finely as to where to cut.

it will be 548 - (26.5 x 2) = 495mm which then gives 598 - 495 = 103mm to be removed from the rack. As I have already shortened mine by 82.5mm I will have to remove an additional 20.5mm

Pic of my lovely drawing because a story with a pic is always better.

[Rob KE25]

Spot on.

Under braking which makes the front end of the car dip it also provides toe in and increases braking stability.

Dimensions wise are the power steer and non power steer racks the same?

Yes, Toe in will help with stability. But minute (very tiny) amounts of toe out in bump on the first inch or two of bump will help with transient response, and help the loaded wheel (out front wheel) grip better in initial turn in.
Too much toe in will be difficult to drive hard, making one feel always anxious in bumpy corners, expecting possible oversteer.


Now to throw a spanner in the works: I nearly drilled the holes in my crossmember, and shortened the rack, but went and so a (very) experienced race car builder/engineer.
Hehas made from scratch extremely quick (current quickest lap times) race cars from scratch. I told him what I'm up to.

Two things he pointed out:

1. Try fix roll centres first. That is what has caused my car to understeer on the track day.( This is irrespective of bumpsteer, rather a large roll moment resulting in too much weight on the front outside tire). My static roll centre hieght is about ground level, should be around 15-17cm about the ground.

2.Then Bump steer. Here we have forgotten something. I will still have bumpsteer, as I have not accounted for the fact that the top of the strut is a fixed position, and unless the steering knuckle tierod end is not only on the same horizontal plare as the LCA ball joint, BUT ALSO on a paralell line with this LCA ball joint (no ackerman built in) the fixed strut top will cause bump steer. I showed him the bump steer drawings I had done, and fair and true, I hadn't accounted for this. When he showed me how to draw it, it all clicked. I'm hopeless at doing drawings on the computer, hopefully someone else understands how to do this, and puts this on the thread.

Chances are that the rack will need to sit higher than the LCA pivot points, with different length tierods than what the LCA are in length.

Cheers, Rob

[oldcorollas]

2.Then Bump steer. Here we have forgotten something. I will still have bumpsteer, as I have not accounted for the fact that the top of the strut is a fixed position, and unless the steering knuckle tierod end is not only on the same horizontal plare as the LCA ball joint, BUT ALSO on a paralell line with this LCA ball joint (no ackerman built in) the fixed strut top will cause bump steer. I showed him the bump steer drawings I had done, and fair and true, I hadn't accounted for this. When he showed me how to draw it, it all clicked. I'm hopeless at doing drawings on the computer, hopefully someone else understands how to do this, and puts this on the thread.

Chances are that the rack will need to sit higher than the LCA pivot points, with different length tierods than what the LCA are in length.

Cheers, Rob

i don7t understand the highlighted bits ?

what does strut top have to do with bump steer?
the tie rod needs to be parallel to the LCA, and the same length, but not in the same horizontal plane.
the rack should sit the same distance above the LCA pivots, as the tie rod end ball joint pivot does above the LCA ball joint pivot...

can you draw what he showed you? maybe he is wrong just building fast cars doesn7t mean you understand steering geometry??


i don7t think i have ever seen a suspension where the steering connects on the same horizontal plane as the LCA.. not even formular cars?

[Rob KE25]

So I sliced up a tie rod end tonight and got some dimensions. I have a question on what I use as a reference point for the measurements though. Look at pic 1 the top blue mark is where the steering knuckle bottom sits on the tie rod end once tightened. Do I take the centre of the main tube and add to the the distance to the mark ie 12mm + 10mm to give me the centre point or do I do as per picture 2 where I am using the centre point of the ball in the join so that the slight bend in the housing is accounted for?

This only makes 3 mm of difference I know but I am interested to know which one to use.

Yep, go with the Number two. It is a fixed part of the steering knuckle. 2-3mm of difference on the vertical plane seems to make a lot more difference to bump steer than in the horizontal plane. 3mm will make a fair amount of difference.

[oldcorollas]

i gotta say here that 3mm vertical, at the end of the arm.. doesn7t make a heck of a lot of difference to the bump steer..// the length of the tie rod ends makes a much bigger difference...

i'm making some calculations on suspension atm.. if someone can feed me some dimensions, i cna do some quick calcs for simple bump steer (no turning of wheel tho), with different LCA and tie rod lengths and positions and offsets etc...
the maths is pretty easy. (assuming a simple radius rod setup)

what is the offset of the radius rod to the LCA (sideways and vertically?)

[Rob KE25]

i don7t understand the highlighted bits ?

what does strut top have to do with bump steer?
the tie rod needs to be parallel to the LCA, and the same length, but not in the same horizontal plane.
the rack should sit the same distance above the LCA pivots, as the tie rod end ball joint pivot does above the LCA ball joint pivot...

can you draw what he showed you? maybe he is wrong just building fast cars doesn7t mean you understand steering geometry??


i don7t think i have ever seen a suspension where the steering connects on the same horizontal plane as the LCA.. not even formular cars?

No, not suggesting that the steering and LCA should be on the same horizontal plane. Probably didn't need to add that bit in my previous bit.

No, I can't draw it for you either (not being much help!!!) Don't know how to do that on computers.

i'll try explaining it: The drawings I did were all about having the tierod end keeping the same distance (horizontal and vertial) IN RELATION TO THE LCA BALLJOINT. For this to happen, the strut top would have to move exactly with the curve of the steering knuckle (for example lets say the strut top pivot point is always 80deg inbound- nearly 90degrees square to the LCA).

But it doesn't.

The steering arm's tierod end will change it's position in relation to the LCA balljoint.

Think about this.
A car wheel moves up and down.

As a result of this, the LCA balljoint will move in or out (since the LCA is fixed at the inner pivot point).

As a result of this, the bottom of the strut will move closer or further away from the vertical plane.

As a result of this, the steering knuckle will move in the vertical plane (rolling ever so slightly over).

As a result of this, the tierod on the end of the steering knuckle will move in relation to the LCA balljoint (as the balljoints are below(mostly) the steering knuckle's boltlike attachment.

SO
to minimise bump steer, one needs to also plot the effect of the strut's change on the steering knuckle's balljoint with respect to the curves of the LCA balljoint too.

[oldcorollas]

>i'll try explaining it: The drawings I did were all about having the tierod end keeping the same distance (horizontal and vertial) IN RELATION TO THE LCA BALLJOINT. For this to happen, the strut top would have to move exactly with the curve of the steering knuckle (for example lets say the strut top pivot point is always 80deg inbound- nearly 90degrees square to the LCA).

yup, that is true. for straightline

>But it doesn't.
>The steering arm's tierod end will change it's position in relation to the LCA balljoint.

yes it will but.... the end of the steering rack also moves out.. so the parallelogram still exists, but it is shifted..

ie, the knuckle end of tie rod moves in relation to the LCA, but at the rack end, the tie rod end also moves in relation to the LCA pivot...


>Think about this.
>A car wheel moves up and down.
>As a result of this, the LCA balljoint will move in or out (since the LCA is fixed at the inner pivot point).
>As a result of this, the bottom of the strut will move closer or further away from the vertical plane.

yes, i'm calculating camber curves atm

>As a result of this, the steering knuckle will move in the vertical plane (rolling ever so slightly over).

very very slightly.. the steering knuckle is attached to the strut.. so if the strut has a 1deg camber change, the steering knuckle also does. on the radius the knuckle is, this roll will be 0.x or 0.0xmm

>As a result of this, the tierod on the end of the steering knuckle will move in relation to the LCA balljoint (as the balljoints are below(mostly) the steering knuckle's boltlike attachment.

the steering knuckle is fixed to the balljoint pivot...

>SO
>to minimise bump steer, one needs to also plot the effect of the strut's change on the steering knuckle's balljoint with respect to the curves of the LCA balljoint too.[/QUOTE]


in pure bump, the steering knuckle and LCA ball joints will move almost exactly the same distance... the difference is very very small.
i can make a simulation to demonstrate tonight i guess..
the reason is the distance from the strut top is so similar..

there is a difference but compared to everything else.. it is tiny... then again, everything adds up.

when turning, the movement of the tie rod in and out is fine if the rack is positioned correctly, as the parallelogram is maintained..
however, due to the steering arm turning, and the requirement of ackerman, there WILL be some changes... but the trick is to make those changes help you, rather than hinder...

ie, wheel has toe out on droop, and bump toe in when compressed.. so that bumped wheels always point toward middle of corner...
or maybe that is not true...

anyway, i will try to make some diagrams tonight....

[LeeRoy]

ie, wheel has toe out on droop, and bump toe in when compressed.. so that bumped wheels always point toward middle of corner...
or maybe that is not true....

It is and it isnt.

With a rack that is positioned behind the front axle toe in on bump will occur if the tie rod pivot is inboard of the LCA pivot and the ball joint is mounted below the LCA pivot (with reference to the horizontal plane that would be made if you drew a line through the LCA pivots.

If everything is how i have described above then as the wheel moves up, the longer arc of the tie rod compared to the LCA pushes the back of the hub out (where the tie rod is mounted) and creates toe in. As the wheel travels further, maximum toe in will be reached when the LCA reaches horizontal. Past horizontal and toe in will begin to reduce again.

The biggest trick is mounting the rack and pivot points so that during suspension travel you dont have toe in and then toe out. If your running your car so low that during normal operation the LCA passes the horizontal axis then you will have some serious suspension dramas having toe in both increase and decrease on compression, rather than increasingly toe in in an exponential fashion relative to compression. You would want at the very least imo, maximum toe in to occur at a point just past where the suspension is capable of reaching.

However, bump steer (either toe in or toe out) will only occur if the LCA arc is smaller or larger than that of the tie rod. If they have the same pivot points respectively and maintain a paralellogram throughout movement then bump steer wont be present. The opposite will occur if the rack if mounted in front of the axle. Being mounted below the axle will also cause the opposite, and being mounted both in front of and below the axle would deliver the same results as above and behind.

This is how i understand it so feel free to correct me if im wrong