Mystery steering pull, could it be the tyres??

[JohnLear]

First, apologies for the long post…
Second, I hope this is the right forum, couldn't decide whether it would fit better in the tyre / wheel forum...

At the risk of being one of those people who ask less than fascinating ‘I have a problem with my car’ type questions, I have a problem with my car, that is hopefully intriguing enough to elicit some response...

The car (RHD Honda Accord, with SLA suspension) has a strong tendency to pull left (more like a mission in life to pull left…), that I can only alleviate by running very different caster angles left vs right. I’m currently running 6.5° caster on the left (way more than stock) and about 3° on the right (within stock range), and while the left pull is still slightly there, it’s much less severe than with equal caster on both sides.

I am intending to pull even more caster out of the right side to see if I can coax the car into tracking perfectly straight, though it irks me greatly even as it is to be so lopsided (I want my steered camber changes to be equal, and they aren’t!).

The mystery arises because the suspension / steering geometries are otherwise all symmetrical to quite close tolerances, i.e.:

• Rear toe zero on both sides (i.e. thrust angle zero)
• Rear camber equal to within 0.1° (-0.8).
• Front camber equal to within 0.1° (-1.5°).
• Included angle equal to within 0.1°.
• Scrub radius equal to within 1.5mm (+8mm and +9.5mm).
• Each spring pair is equal free length (i.e. no sagged springs).
• Chassis is untwisted (tower heights all within the same plain + - 1mm).
• Front anti roll bar has been removed, and rear ARB doesn’t appear pre-loaded in any way.

All sub-frames are tight and I can’t find any suspect suspension components, all articulated joints / bushes seem to be fine. I’ve carefully measured the front uprights and both have almost identical relationships between upper / lower ball joints and hub axis (interestingly discovered that the hub is forward of the steering axis), and trail is thus equal to within 0.5mm at equal caster (though of course not with unequal caster). KPI is equal to within 0.05°.

I’ve tried numerous alignment settings, including asymmetric settings, and only substantially asymmetric caster seems to improve the problem to any appreciable degree (when making any asymmetric change I have equalised the camber and reset toe of course). An exacerbating factor (but surely not causative), might be that I like to use quite high tyre pressures (around 45psi).

The pull isn’t caused by road cambering to the left (though this does of course make it worse), the road needs to be visibly cambered right in order for the car to track straight. Steering is also heavier right vs left, more so the more equal the side to side caster is. I have accidentally managed to briefly achieve straight tracking with zero pull, but only with lesser right caster AND both front tyres having significant camber wear off to the right (countering the left pull due to this conicity).

This has been an ongoing issue since I’ve owned the car, through several sets of tyres, through swapping rims / tyres side to side, and ‘flipping’ a tyre or tyres on the rim. With the current tyres swapping and flipping merely results in changing the severity of the problem, i.e. either a minor or major left pull, so the current front tyre pair at least seem to have some pull inherent in them (conicity, ‘bias steer’..?). But, if the tyres were the root cause then I’d expect that swapping them around etc would result in a pull to the right (surely?), which I’ve never achieved no matter what I’ve tried.

I was suspecting unequal power assist left vs right (i.e. leaking seals in the rack hydraulics etc), but have recently removed the PS belt and only achieved heavy steering (which I much prefer to the overly light assist). The fact that it’s been a problem with different sets of tyres has had me thinking that it must be some mystery geometry problem, but I’m starting to wonder if it isn’t tyre related after all (considering that very careful alignment to symmetrical settings hasn’t fixed it, and I can’t think of any other geometry that could possibly be the culprit).

Now, to my suspicions about tyres; I’ve noticed that very many non directional and notionally ‘symmetrically’ patterned tyres don’t have a truly symmetric tread pattern, but have grooves that are more or less diagonally oriented across the width of the tread face. This pattern ‘diagonality’ doesn’t change if the tyres are swapped side to side or flipped on the rim, i.e. the diagonal orientation is always the same no matter how the tyres are fitted, so, all tyres on the car exhibit this tread pattern asymmetry in the same direction...

Could this tread asymmetry of itself cause a pull? Could the tyre manufacturers be deliberately making tyres with an inherent pull in an attempt to compensate for pull caused by road camber??. Could any such tyres be designed to counteract road camber pull where the car is assumed to be LHD and driven on the right hand side of the road? Could such tyres (if they exist) end up being sold into markets where we drive on the ‘correct’ side of the road, and thus cause the car to hunt for the gutter????

From memory all the tyres I’ve used have had a somewhat similar ‘diagonal’ tread pattern…

This problem is irritating purely on the driving level (the pull kills steering sensitivity, and I suspect contributes to tyre wear), but it’s also driving me nuts that I’ve just failed to diagnose the problem, so it’s intellectually frustrating as well.

Might I be on the right track thinking it could be tyres??? Are there any other tyre related issues (other than diagonal tread pattern) that might cause a pull and be immune to how the tyre is mounted? Can anybody think of anything else that could be causing this? Any considered or wild theories will be entertained…

 

[patprimmer]

Have you tried measuring the weight on each wheel.

Could a brake be dragging.

Are the tie rods symmetrical on the knuckles.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[JohnLear]

Pat,
Thanks for the reply.
The brakes are fine, no dragging.

I don't currently have access to appropriate scales for measuring the corner weights (the joys of living in a rural location), but it's my understanding that FWD Hondas are typically somewhat heavy on the LF and light on the RF (hearsay).

It did occur to me that unequal front corner weights acting through the caster angle and the 'lever arms' of trail and (particularly) +SR might cause a tendency for the wheels to turn around the steering axes (this certainly occurs with karts, but the SR is massive, while my Honda's SR is quite small). However, with +SR, and assuming the chassis to be heavy on the LF, I'd expect this affect to cause a pull to the right not the left...

Theoretical misgivings aside, with this in mind I tried placing 1cm thick spacers under the spring seats on the RF and LR (thus increasing weight to some unkown degree on the RF / decreasing on LF), with no perceptible change to the pull. Since I was shooting in the dark, I also tried the opposite weight jacking and placed the spacers under the LF and RR, to equally zero affect on the pull.

The springs are not sagged, no suspension member is bent, and the chassis measures 'flat' at the tower tops. The rubber bushes (where the springs / dampers attach to the LCAs) are vertically compressed to some degree, but very close to an equal degree on both sides (and at the rear).

I've driven other Hondas that don't exhibit this steering pull (with suspension / steering / alignment in worse apparent condition than mine is), and considering that all components that might affect the 'stock' corner weights seem pretty good, and jacking corner weights on both diaginals has no affect, I just can't see that corner weights are likely to account for it...

Note that changing the weight distribution by adding passengers doesn't perceptably affect the steering pull.

"Are the tie rods symmetrical on the knuckles."
I've checked this. With equally set side to side caster the tie rod symmetry was not the same, the left side of the rack being behind the right side, and thus the left tie rod was visually angled more forward.

I took measurements and did drawings that indicated this would have some affect on steering 'leverage'left vs right (an effective unequal Ackermann like affect due to differences in the arcs of tie rod end motion left vs right), but also that this would be very slight.

Despite my drawings suggesting any effect of this on steering weight and / or pull likely to be very small, I moved the left side of the rack forward until the tie rods were at least very close to being equally angled (forward from rack ends to rod ends). Result, no change to the pull (not surprised).

Presently, with unequal caster angles (which alone has substantially reduced the pull) the tie rods are no longer equally angled. This is due to the manner in which caster has been adjusted (altering radius rod length), and also affected left vs right side wheelbase, with the right side now being about 1cm shorter. I don't think this is causing any significant problem.

I've even tried inserting (on one side) a 1.5mm spacer between the wheel and hub face to equalise the slightly unequal SR, to no change in pull (not surprised).

If joints and bushes are all good, there is no significant pre-load from ARBs, and caster, camber, KPI, SR, trail, thrust angle, tie rod geometry, corner weights AND tyres are all equal (or very close to), then I can't for the life of me see what could be causing such an obvious steering pull. If the tyres do have an inherant tendency to self steer, then I can see a reason...

At this rate, I can see the car ending up with strong + caster on the left and some - caster on the right...

 

[patprimmer]

If tyres typically pulled that much, it would be a common problem.

If the diagonal tread pattern created it, flipping the tyre on one side would rectify it, if by flipping you mean mounting the original inboard face on the outside and thus having the diagonal in the other direction. You say you tried that, so you have eliminated that to the same extent that you eliminated suspension settings. It will be something you have not changed, not something you have changed several times.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[JohnLear]

Thanks again Pat.

"If tyres typically pulled that much, it would be a common problem."

Which is why it's taken me so long to start considering tyres as a possible cause, along with the problem surviving a number of tyre changes.

Could it be that relatively high caster angles (with = side to side caster I had it at a bit over 6°, which isn't much for a RWD, but a lot more than seems typical for FWD) and high tyre pressures conspire to make the tyre much more sensitive to tread groove angle in a front drive application??

This might make the problem (if it does exist as a tyre problem) much less common since relatively few cars will have the combination of caster and pressure that I prefer?

"If the diagonal tread pattern created it, flipping the tyre on one side would rectify it, if by flipping you mean mounting the original inboard face on the outside and thus having the diagonal in the other direction."

Yes, that's what I mean.

I also assumed that doing this would 'obviously' change the pattern orientation, until I actually tried to make it happen. Turns out that it's just not possible to change the diagonal orientation of the tread grooves by either flipping the tyre on it's rim or by swapping the whole wheel to the other side of the car (which amounts to the same thing as far as the angles of the tread grooves are concerned).

It's easy to test this while sitting at your desk by marking the edge of a suitable disc (jar lid is perfect) with a diagonal mark representing the tread grooves, then trying to make the diagonal mark change orientation by flipping the disc in any way. I'll bet you can't make it happen either...

"You say you tried that, so you have eliminated that to the same extent that you eliminated suspension settings. It will be something you have not changed, not something you have changed several times."

There lies my problem. I can't think of anything that might conceivably affect steering pull and relative side to side steering resistance that I haven't with reasonable confidence eliminated, save for the tyres (however tenuous this possibility might be). I was hoping someone might be able to confirm my tyre suspicions, or else point out a gorilla sitting in the corner...

 

[patprimmer]

Are the struts a pair or was the car ever repaired and a "same model" replacement strut used. There may be an in model changes to steering knuckle or what I understand as mechanical trail. I might have the mechanical trail terminology wrong as I am not a suspension guy.

Are you sure the stub is exactly the same distance behind the king pin or strut axis. Is it the same distance, but forward on the other side.


Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers for professional engineers

 

[JohnLear]

Pat,
The 'knuckles' measure out almost identically. The only significant measurable difference between the two is the location of the upper ball joint relative to the hub face, with the left side upper ball joint being 2mm further inboard of the hub face than is the right side UBJ.

The uprights themselves are of the type with a tall upper section placing the upper BJ higher than the top of the tyre, which is why this 2mm difference in lateral UBJ location only equates to a 1.5mm difference in side to side SR and only a 0.3° difference in KPI.

I've proved to my satisfaction that the 1.5mm SR difference isn't the problem, and I'm having a hard time convincing myself that a difference of 0.3° KPI could be to blame. In any case, greater left side KPI would cause the left wheel to 'lower' more for X left steer than the right wheel will lower for X right steer, thus requiring more effort to 'lift' the body mass in left corners, which would create the opposite problem to what I have, i.e. a pull to the right rather than left...?

The hub axis is equally forward of the steering axis on both uprights, at least it's so close it doesn't matter. According to my measurements / drawings the right side hub axis is 7.8mm forward of the steering axis vs 7.7mm on the left knuckle (with the steering axis at 90° to the ground). This is well within the accuracy of my ability to measure this, so the 'cross' could be zero, or, slightly greater, but not by much.

This (as measured) 0.1mm difference between the hub axis / steering axis displacement equates to a 0.5mm difference in trail distance (once I tilt the caster back to 6°), right trail being 25.5mm vs left at 25mm which couldn't possibly be the problem, especially as right trail will now be considerably less at the lesser right side caster angle.

At any rate, I just can't see minute differences being enough to cause this pulling problem in the severity that it has (at equal caster angles), there is something significantly wrong somewhere, and I'm just failing to find it.

I suspect I'll just have to continue pulling caster out of the right side until I get zero pull (if I actually do...), but I won't be happy knowing that I've had to substantially bodge the geometry to achieve it, and that the caster induced steered camber changes won't be the same left vs right.

Regards and thanks,
John.

 

[evelrod]

Just a BTDT...I had a 124 Fiat that exhibited a similar problem. It took quite a while to figure it out. In my case it was a slightly bent lower control arm on the left front. Not sure how it happened as it was this way when I bought the car. Anyway, I did all the stuff you have done until one day I accidentally 'bumped' a curb a bit and really bent the thing.
when I replace the lower control arm...voila...no more pulling. I could not tell exactly what the problem had been aside that it was most definitely the lower control arm, but, as I said, I believe it had been slightly bent all along.

Rod

 

[cibachrome]

If your car is RHD and not being driven in an island nation, then the tires may have the wrong plyrat settings (residual aligning torque). This is a tire construction detail intentionally produced to correct for road crown effects. What's needed is a tire with the belts reversed within the carcass so that they pull the car to the right instead of the left. This tire design feature is related to lead/pull and headlight aiming issues, so don't think that it is a "defect". It should be somewhat pressure sensitive, so put a lot more air in the tires to lessen the effect. (40 - 50 psi for a test on a smooth road, no they won't explode, don't worry get permission from your spouse, etc.). If this is the cause, get tires from the car manufacturer that are intended for it. Your dealer may not be aware of the simple fix (or may not want to fix it in a low cast manner since he will bill Honda for it). He may choose to alter the car's cross camber alignment to fix it (which you could do, too)

You can check the left and right wheelbase dimension to make sure the rear axle isn't the cause. And, check to see if one or more of your wheels is a different rim width.

 

[JohnLear]

Thanks cibachrome,
I think you're more or less confirming my suspicions about the tyres being at least a possible cause.

The car has all stock rims, is RHD and IS being driven in an island nation (I'm assuming the reference is to driving on the left side of the road in Australia, Britain and Japan, and on the right just about everywhere else?).

Would I be correct that you're suggesting this designed in 'pull' is unresponsive to the manner in which the tyre is fitted to the car?

My suspicion relates to possibly surplus tyre stock originally designed / intended for LHD / right hand side of road markets perhaps being dumped here in Australia (or some tyre manufacturers simply not bothering to make tyres with constructions tailored for RHD markets). Perhaps I'll need to make sure I buy Australian manufactured rubber next time?

I already use fairly high pressures (usually around 45psi, depending on what I find works to my liking with individual tyres). I've already tried cross camber to no avail, i.e. I can alter the pull characteristic to some degree, but not eliminate it while keeping camber cross within reason, and sometimes I've found the camber change to have an opposite effect to that I was expecting.

It occurs to me that while the accepted wisdom is that a car will pull in the direction of the side with the least neg camber or most pos camber (rolling 'cone' effects), that there are other factors involved that may cause a pull in the opposite manner.

My best guess as to why this may be:

If we assume slight positive or zero SR, then neg cambering the wheel significantly will place more load on the inner part of the contact patch and less on the outer part, and as such may create an effective SR (as opposed to the nominal SR) that is negative, causing an additional but opposite pull around the steering axis at that wheel that to some degree counteracts the rolling 'cone' pull effect?

Such a pull might possibly be caused by chassis weight interacting with the caster angle (wheels 'wanting' to 'rise' and 'fall' and thus steer around the rearward leaning caster angle), and perhaps by the car being FWD (tractive effort 'pulling' the effective contact patch centre forward around the steering axis, thus causing the wheel to 'steer' outward).

If this is a reasonable hypothesis, it would be likly that the affect might be greater the wider the contact patch, the higher the psi and / or the greater the sidewall stiffness? And I am using fairly high psi, and do have significant neg camber...

 

[JohnLear]

Rod,
Thanks for the suggestion, I'll measure up the LCAs, but I have my doubts that this might be the cause. Surely the shape of the LCA is irrelevant so long as the geometry is correct, and the spring load isn't affected by the LCA being bent down or up?

With your Fiat I can only assume that either the geometry was affected or the corner weight was affected?

 

[cibachrome]

Tires for an island market need a reverse layup tire from the same mold. This means the belt package is inverted. Because the belts are separated by a finite distance from the rubber matrix. when you stretch them, they twist. That's your residual aligning torque (RAT). Its the steer moment at zero lateral force. You need tires made for your roads.

One other thing that comes to mind: Perhaps the steering valve is staked with an offset to compensate for the road crown. You're running without power steering, eh? Perhaps that's a consequence.

BTW: I said cross camber, not just camber. You need an asymmetric inclination angle. Oh, camber stiffness is pretty weak in most radial tires. Don't be surprised if you need a LOT of camber...

 

[JohnLear]

cibachrome,
Appreciate the responses.

I'm not sure that I fully understand all of your answer, but probably well enough to see how this might possibly be affecting my situation.

So in a given tyre the RAT is unchangable because the orientation of the belt layup is 'handed' in much the same manner as the tread groove pattern is (if the grooves have a diagonal character as is common), and so the tyre will always pull in a given direction no matter how it's mounted on the car?

This would be as opposed a pull caused by conicity, where the pull will change direction if the tyre is reversed on it's rim or the wheel moved to the other side of the car?

You said; "Because the belts are separated by a finite distance from the rubber matrix. when you stretch them, they twist", which I'm not quite getting. I can see a dynamic twisting of the belts causing a pull, but not the cause of the twisting in a particular direction / manner. Do the belts have a diagonal 'grain' woven into them, and if so is it the angle(s) of this 'grain' what is manipulated to create the intended RAT direction (and strength?)?

I've done some searching around and found at least one reference sugggesting that the tread pattern also has at least some affect on RAT.

BTW: I'm sure I said cross camber...

 

[cibachrome]

The RAT and conicity are completely separate. Conicity should cancel if your wheel weights are reasonably symmetric. I presume your tires are from the same build date. Sometimes this gets into the mix.

The tire's belt package with 2 steel belt layers (most common) are woven wires set a ply angles (15 - 30 degrees lets say. They are usually the same recipe so they are identical. However, there is a layer of rubber bewteen them to hold this assembly matrix all together. Hence, the net effect like a Chinese finger puzzle: You pull, it twists. I doubt that the tread has much to do with your issue because the effect of wear on lead and pull sensitivity would not be acceptable to the vehicle manufacturer. My experience is that asymmetrical treads are more for wet and snow traction effects than dry cornering. It never shows up on machine tests.

The tire manufacturer also wants to keep the number of molds down ($1M + per) so the reversed belt package is a low cost work around for the RHD vs LHD models.

Google "Aligning torque static phase" and see what you come up with. There is a paper on a method for setting tire and vehicle specs for alignment and tire RAT, Conicity, Plysteer and PlyRat etc. values to balance the road crown forces and moments. "Analysis of Tire Effect on the Simulation of Vehicle Straight Line Motion" JUNG-HWAN LEE, Vehicle System Dynamics, 33 (2000), pp. 373–390 0042-3114/00/3306–373 $15.00 that should enlighten you.

Most problems in the industry are due to difficulties measuring, sorting, screening and specifying these important but often ignored tire properties from a customer satisfaction and pleasability standpoint. The aftermarket replacement tire issues are even more complicated...

Even my BMW dealers are not all aware of the "STAR" tire label. Get new skins for a 5 series from a dealer and they don't have the star symbol, I'm gonna be cheesed.

 

[JimPeop]

One dimension that you have not mentioned and could cause the steering pull issue is the diagonal center to center. If you place the car on four jack stands and set them as close as possible to the same distance from the hub face as possible drop a point to the floor directly from the center of the flange face center on all four corners and then also drop a point from the center of the center of the lower pivot point on the front lower ball joints and drop a point from the center of the tie rod centers then measure from the right rear to left front and left rear to right front. On the front you want to take the measurements to all three points that you have dropped and see if the dimensions are the same. If the measurements are not equal then this could be causing this pull affect and would be overcome by running an uneven caster on the front as this would increase the steering pull on one side therefore compensating for the unequal pull caused by the out of square wheelbase and steering pivot points. Hope this helps and may get you a point to start finding the component that is not in spec.

 

[JohnLear]

Thanks Jim,

If these diagonals measure unequally then this would have to be indicative of the front (or rear) subframe being offset to one side, or (/ and...) the wheelbase being shorter on one side due to a subframe not being longitudinally straight and / or a suspension component being damaged?

In the former case, the rear wheels would not follow directly in the paths of the front wheels, that is unless rear toe was set to the centreline as measured from points central to the hub flanges at both axle lines? (which it has been).

And, if so then one rear wheel would be farther back than the other?

And, if so then even if the wheelbases were equal side to side, both wheels on one side would be farther back in the chassis than on the other side?

 

[JohnLear]

Cibachrome,
Thanks again.

Quote; “Conicity should cancel if your wheel weights are reasonably symmetric.”

??

”The tire's belt package with 2 steel belt layers (most common) are woven wires set a ply angles (15 - 30 degrees lets say. They are usually the same recipe so they are identical. However, there is a layer of rubber bewteen them to hold this assembly matrix all together. Hence, the net effect like a Chinese finger puzzle: You pull, it twists.”

I’m not quite seeing the CFP (Chinese finger puzzle) analogy, from childhood memory they narrow when pulled not twist, but then I suppose it’s twist related, with each strand ‘untwisting’ when pulled…

But, is this close to your meaning?:

Might not the oppositely angled ply steel belts be supported differently within the tyre, with the inner belt being supported / backed up by the fabric belts / case plies, then the relatively soft rubber layer between the inner and outer belts. This might mean that the outer steel belt is less firmly supported and could ‘twist’ more as it rolls under load, and thus have a greater effect on tyre lead than the inner belt does. Thus the tyre may lead left or right depending mostly on the orientation of the outer belt’s ply angle?

If this is correct then it begs the question; do all radial tyres have a directional lead?

You said the two belts are usually identical (ply angle), but could lead be manipulated by using differing angles in the inner and outer belt plies, to produce a tyre with less lead or more lead, or maybe zero lead (perhaps with a greater angle for the inner belt plies?)?

“Most problems in the industry are due to difficulties measuring, sorting, screening and specifying these important but often ignored tire properties from a customer satisfaction and pleasability standpoint. The aftermarket replacement tire issues are even more complicated...”

I’ve heard (read somewhere, maybe on this site?) that there have been or are issues within ‘the industry’ relating to steering lead, with car companies blaming the tyre companies and vice versa?

 

[GregLocock]

Each tire is measured for more more than 10 characteristics, including conicity. Those that don't meet the tire manufacturer's specs are scrapped. Those that meet the OEM specs are fitted to new cars. The rest are sold on the aftermarket.




Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[cibachrome]

Take two conic styrofoam coffee cups and stab a pencil between them. (Tires with conicity on an axle.) As long as the rolling radii are about equal, the axle will go straight.

Yes you could make dis-similar belt angles, but the difficulties in building, managing ply stocks and fitting them up to the build machinery is more trouble than tire makers would bid the jobs for. Especially since a controlled amount is desired anyways. A 3rd belt could also be added, with large cost, mass, rolling resistance and high speed uniformity penalties. Single ply jobs have durability and wear issues.

The industry issues arise because marketting zones and replacement tire inventories confuse the vehicles. Sharing of unique tire constructions on different vehicle platforms also causes some serious headaches. And, net build vehicle alignment designs (no ability to alter cross camber or cross caster, maybe even static camber) makes the vehicle builder or the tire supplier want the other partner to change because of cost, inventory and timing constraints.

(Are overweight people too heavy or too short? Two very different surgeries are necessary).