wheel hop durring Acceleration

[T.B]

Looking for some help to get rid of wheel hop during acceleration. I think its caused from the rear moving forward, normally fixed by swapping to a stiffer bushing or heim joints (witch i think this is my problem that the heim joints are bad). It's a 1971 Camaro with a g-bar that we made are own upper and lower links with heim joints. would i be better to try a jonny joint are one of the spohn joints? Are can there be another problem im not getting?

[exwestracer]

Did you just build the links, or design the whole system? I'm asking because I'm wondering about geometry. Has it always had wheel hop since you put it in, or is this something that just "happened"?

Unless you can reach under there and move the links around by hand, I wouldn't condemn the heims JUST yet...they usually don't allow wheel hop until they are worn out.

[T.B]

YES i just built the links it started out as a tri- 4link from air ride or chassis works but i think air ride designed it, and when chassis works took it over they just added a set of vari-shocks instead of air bags. No the wheel hop just started. The car only has about 800 miles and 2 track days. i thought the heims would last longer than that. o and i did just change spring rates i went up to a 200lb spring from 175lb but i don't think that would cause the problem.

[exwestracer]

...duplicate

[exwestracer]

Tim, did you reset the ride height when you changed springs, or is the rear sitting higher now? Did the wheel hop start at the same time?

I never assume ANYTHING will or won't cause a problem...

Can you "rattle" the suspension links by hand?

[Bryce]

If you raised the rear of the car with the new stiffer springs you would have changed your IC and A/S.

What kind of rod ends did you buy. I would recommend the chromoly with teflon kevlar race. I am running those on the front of my falcon and I have no wear issues after 2.5 years. I am using the same on my 3-link in the back I only have about 100 miles but no wear issues either. Also I have no wheel hop and 225 pound springs.

[T.B]

Yes I did raise the car by 1/2" or so with the new spring's it was lower in the back than the front, now it's level.

No I can't rattle the rods. Yes the first time I noticed it was after the springs were changed, but it was a few 100 miles after we changed them.

[exwestracer]

Yes I did raise the car by 1/2" or so with the new spring's it was lower in the back than the front, now it's level.

No I can't rattle the rods. Yes the first time I noticed it was after the springs were changed, but it was a few 100 miles after we changed them.

This is the setup we are talking about, right?
52954

With that short of an effective upper link, the IC will be very sensitive to ride height changes, as Bryce mentioned. I would put the rear axle on stands (weight on axle and car level to floor), and measure the distance from the upper mounts on the axle and frame down to the floor. The height number itself doesn't matter, but if the frame mounts are equal or higher than the axle mounts, the car will be very prone to wheel hop. <This is a gross over-simplification of a complex calculation, but I would start there. If the car has adjustable shocks, also make sure one of the adjusters didn't get backed off all the way somehow. About the only other thing I can think of that would cause this problem is a dead shock; and even so, correct geometry should eliminate it.

[T.B]

Yes very simular.Thanks Ray i will give that a shot.

[T.B]

Thanks for the info guy's the frame mounts are equal on one side and a slight higher on the other. Is there a magic number i need to be at with the upper links between the axle and frame mount? the lower links are equal.

[exwestracer]

Thanks for the info guy's the frame mounts are equal on one side and a slight higher on the other. Is there a magic number i need to be at with the upper links between the axle and frame mount? the lower links are equal.

Ok, here's where it starts to get complicated... With the lower links level, you are technically at the "correct" ride height. With the top links level (I'd try to figure out why they are different, too...), you have NO instant center, and therefore no anti-squat on launch.

52959

The more down angle you get in the upper links, the harder it will plant the tires under acceleration. Now, how you get them there is another story...and it IS possible to go too far on a street car.

[Bryce]

This is a general statement which may apply to your set as well.

With a stiffer spring you will need a stiffer shock, otherwise this becomes an underdamped system, which can lead to oscillation or wheel hop in your case.

A tire with too little air pressure can also lead to a oscillation and eventually wheel hop.

In your case: can you please measure all the link pivot points left and right and post up, with your weight in the drivers seat. Ray or myself can run them through a suspension calculator and tell you the A/S.

1) lower links should be horizontal
2) the upper should point slightly down towards the front of the car.
3) the rear end should be centered and the links should be centered in the car aswell
4) the lower links should be parallel to the car's longitudal centerline
5) the upper links should have an equal angle to the axle centerline.
6) both sides should be the same with your weight in the drivers seat.

[exwestracer]

I do think the problem is ride height related (if only because it didn't appear until you raised the car with the stiffer springs); however, be sure to check all the items Bryce mentioned, as well as the shock adjusters for correct settings.

[T.B]

Thanks Bryce and ray I will check all the dim tomorrow and post them. I'm going to re fab the upper frame mounts to fix the passenger side being a little off. And It would be nice to see were I'm at before having it all welded up.

[T.B]

Here's A quick sketch on were I was able to get at with a few mod's. I can get more angle down in the uppers if needed.

53053

[Norm Peterson]

Do you have the elevations of those pivot points and a rear tire size? I'm guessing that the uppers are at about 3" above the axle centerline and that the lowers are at 4", maybe 5" below, with about a 13" tire loaded radius.

What I am seeing is that starting from a relatively low amount of anti-squat with the previous springs, you lost about 1/3 of that just from the 1/2" ride height change. Your IC is probably way out around 10 feet ahead of the front axle line.

A rough plot of your anti-squat (rough because I am working from at least 3 guesses here) is suggesting that the antisquat increases rather rapidly as ride height is lowered (it gains over 20% inside the first inch of squat). I'm not sure what that's trying to tell me, but it's a lot higher rate of A/S change around the static ride height than I'm used to seeing with either OE triangulated 4-link or 3 link arrangements. Definitely a function of those short uppers and their length relative to the length of the lowers. Roll steer looks minimal, though.


FWIW, at ±20° plan view skew (I think that's what I'm seeing), if the rod ends in the uppers were fried badly enough to cause the hop all by themselves, the axle would probably be moving around laterally enough to make the car feel a bit woozy.


Norm

[T.B]

Hi Norm, the uppers are 4 3/4" up from the center line of axle and 4 1/2" back of the center line
the lowers are 5" down from center line and 1/4" back from the center line.rear tire size is a 335-30-18 witch i think would be around 26" dia.

thanks

[exwestracer]

According to your drawing, your IC is out at about the front axle centerline and about 8" above ground...WAY below the neutral line.

On a typical 110" car, the neutral line would fall somewhere about 50" ahead of the rear axle C/L at that height.

[Norm Peterson]

. . . 4 1/2" back of the center line
??? (doesn't appear to agree with the earlier sketch).

If the '4' isn't there, A/S is under 15%. Might be a little higher if the '4' is and the length is still 9-7/8", but not by much. No matter where the UCAs start, 2° down angle from there isn't much at all, and it takes quite a distance to converge with the LCA line in side view.


Norm

[ochohill]

This is the setup we are talking about, right?
52954

With that short of an effective upper link, the IC will be very sensitive to ride height changes, as Bryce mentioned. I would put the rear axle on stands (weight on axle and car level to floor), and measure the distance from the upper mounts on the axle and frame down to the floor. The height number itself doesn't matter, but if the frame mounts are equal or higher than the axle mounts, the car will be very prone to wheel hop. <This is a gross over-simplification of a complex calculation, but I would start there. If the car has adjustable shocks, also make sure one of the adjusters didn't get backed off all the way somehow. About the only other thing I can think of that would cause this problem is a dead shock; and even so, correct geometry should eliminate it.

Is there anything wrong with the 2 upper arms being all heim joints in this design?

[Norm Peterson]

Ideally, all 8 pivot points in ANY kind of triangulated 4-link suspension would be spherical pivots of some sort.

The reasons for using cylindrical rubber bushings has nothing to do with theoretical geometric requirements and everything to do with minimizing things like production cost, NVH, and in-use inspection & maintenance. You might even say they work as well as they do in spite of themselves rather than because of.


Norm

[T.B]

Yea Norm the sketch didn't agree with the dim. i posted. thats what happens when you trying to rush to get your notes on paper.My problem is the IC was way out of wack. Anyway i am fabing some new upper and lower brakets on the axle that will give me the ability to go from 5 to 10 deg's down on the upper's and from level to 5 deg's on the lowers. The car works pretty good for how far of i was off just need a little more foward bite.

[Norm Peterson]

With lowers that are parallel in plan view, I don't think you want them running uphill from axle to chassis for most driving conditions (the axle steer goes "loose" sooner and may even start out that way).

Moving the UCA axle side up about 1" from where it is now might be a good place to start. I think you'd be around 50% A/S with just that change. Don't forget that the longer bracket length increases the bending load in the brackets and in their welds to the axle tubes.


Norm

[ochohill]

Ideally, all 8 pivot points in ANY kind of triangulated 4-link suspension would be spherical pivots of some sort.

The reasons for using cylindrical rubber bushings has nothing to do with theoretical geometric requirements and everything to do with minimizing things like production cost, NVH, and in-use inspection & maintenance. You might even say they work as well as they do in spite of themselves rather than because of.


Norm

I thought there was bind when all 4 joints of the UCA's were heims. Is this not the case? I recall reading an in depth study on this at CC.com.

Found it. Post 3 is a copy of the original post.

http://www.corner-carvers.com/forums/showthread.php?t=390&highlight=LINK+BIND

[exwestracer]

I thought there was bind when all 4 joints of the UCA's were heims. Is this not the case? I recall reading an in depth study on this at CC.com.

Found it. Post 3 is a copy of the original post.

http://www.corner-carvers.com/forums/showthread.php?t=390&highlight=LINK+BIND

I think you were reading it wrong. Case 1 was LCA with Heims, stock UCA bushings, and findings were that ALL the bind came from UCA bushings. The case where the bind went up with all heims (case 3 LCA with torque arm) they identified the bind as coming from side load on the torque arm. All heims were assumed to be bind free...which was Norm's point.

[Norm Peterson]

Thanks, Ray. :cheers:


From a purely theoretical point of view, a stick axle sitting all by itself in space has 6 degrees of freedom ("different ways it can move"). What you need to do to make it useful under the rear of a vehicle is freely allow only 'heave" (2-wheel bump mode) and "roll", and restrain the other four motions to be very small.

For the purposes of this thread, restraining those other four requires exactly four links that are not all parallel to one another in plan view. When you add any more restraint than that - which includes the force that it takes to "bend" a cylindrical bushing about any axis other than the center of the bolt that runs through it - you get what's commonly called "bind". IOW, where something has to "give" a little in order for the suspension to move in the direction(s) you want it to move it in. The amount of this "bind" may be a little (think OE bushings) or it may be huge (ladder bar drag car).

By definition, a link has end connections that are completely free to pivot in pure rotation in any direction. A few inch-lbs of end connection rotational resistance due to friction in a Heim or other spherical joint can be ignored. A lot more coming from having to "squish" a hard rubber or poly cylinder or actually bend some steel a tiny bit matters, but can usually be designed around anyway. More or less.

Side note - although a Watts link looks like it's two (three?) links, it functions as if it was only one.



I wonder if Ehren ever anticipated how far off into the future and how widely his study would be quoted.


Norm

[ochohill]

Perhaps I misunderstood the report.

Case 4 is stock rubber bushings at all 8 points.
Case 1 (least amount of bind in this test) has heims at all 4 LCA joints. Stock rubber uppers.
Case 10 only has 2 heim joints to the otherwise stock system. There is 1 each on the uppers at the chassis side. (why would adding bind free heims induce greater bind in the sytem?)

This cuts the amount of fore aft compliance in half and induces a huge amount of bind. Elminating the other 2 sources of compliance would induce even greater bind.(the system the OP recently added)

Case 1 shows the benefit of the heims at the LCA and the need for twist compliance in the LCA.
Case 10 shows the need for fore/aft compliance required of the upper control arms.



These cases had no panhard bar in the system.

[exwestracer]

Upon further reading, a couple of things he said have me wondering how relevant that information is...

Nowhere does it say that they are testing a triangulated (as opposed to parallel) 4 link. They may be in some of the cases, but it is never specified. In his intro, Ehren says they cycled ONE wheel through 3" of bump and droop to simulate roll. If it was a tri4, and the other wheel was fixed (axle end on jackstand, for instance) that would explain the increased bind. The roll center is at the convergence point of the trinagulated links. That point should be fixed in space. With one end of the axle supported, that support point becomes an effective second roll center, which causes geometric bind. He mentions in Case 10 that the upper links needed an "effective length change". In a tri 4 link with freedom to move, this should never be the case. I believe they were actually testing a parallel 4 link, but again this is not specified one way or the other.

I think we have to make too many assumptions here to take those bind vales as gospel. Reading through Ehren's explanations, I do get the impression that they view sphericals (Heims) as having the least amount of bind. And I KNOW from my own experience that a properly designed tri 4 link with Heims has less bind in roll than one with rubber bushings. No doubt.

[Norm Peterson]

I doubt that even Ehren intended them to be "design values", only a general indication of what was going on (and perhaps as justification for MM's 3-piece poly bushings). Just changing the plan view skew of the links could be expected to result in somewhat different numbers still with similar trends.

Fixing the height of one wheel and cycling the other isn't really a pure roll condition - it's really a roll + heave situation, more than likely with some unintended lateral axle constraint due to friction at the supported axle end as well. Not that you would never have roll + heave together - you certainly do at times.

The matter of "the uppers having to change length", meaning a significant length change as that wording seems to imply, would only be true if the roll center was a true mechanical constraint located out of the plane of the uppers. But a geometric roll center is only a virtual point, and once the axle is over-constrained its motions start having to consider the relative stiffnesses of everything between it and the chassis. IOW, the neat little geometric sketch isn't good enough any more (and you've just stepped into kinematic & compliance modeling/testing). I'm pretty sure that you get a combination of the axle's actual axis of roll rotation moving away from the geometric construction (probably upward at the axle line toward the UCA "plane"), any compliant bushings "squishing" a bit, and the control arms actually changing length - but only by only a ten-thousandth of an inch or so.


Norm

[Bryce]

I think they were using a stock fox mustang or SN95 tri-4-link.

I have a tri4-link in a 65 mustang, with heims on all attach points, there is no bind in normal operating conditions.

I have been using some of my spare time to develope a kinematic model of a tri4-link to show bind in roll and bump. I have one that shows the total deflection of a watts link.

[ochohill]

Upon further reading, a couple of things he said have me wondering how relevant that information is...

He mentions in Case 10 that the upper links needed an "effective length change". In a tri 4 link with freedom to move, this should never be the case.

The test disputes your last statement, at least on a mustang.



I doubt that even Ehren intended them to be "design values", only a general indication of what was going on (and perhaps as justification for MM's 3-piece poly bushings). Just changing the plan view skew of the links could be expected to result in somewhat different numbers still with similar trends.

The matter of "the uppers having to change length", meaning a significant length change as that wording seems to imply, would only be true if the roll center was a true mechanical constraint located out of the plane of the uppers. But a geometric roll center is only a virtual point, and once the axle is over-constrained its motions start having to consider the relative stiffnesses of everything between it and the chassis. IOW, the neat little geometric sketch isn't good enough any more (and you've just stepped into kinematic & compliance modeling/testing). I'm pretty sure that you get a combination of the axle's actual axis of roll rotation moving away from the geometric construction (probably upward at the axle line toward the UCA "plane"), any compliant bushings "squishing" a bit, and the control arms actually changing length - but only by only a ten-thousandth of an inch or so.


Norm

I don't see this test as a justification for MM's 3 lower poly bushings. They tested worse than heims and stock components in the LCA's.

If the uppers cannot change length due to heim joints what does the 0.010" (your number not mine) length change do?

It changes the UCA bracket into a 3rd spring.

A bit of history:

Some time back I frequented a local mustang shop. The shop owner was kind enough to allow me to peruse the mechanics area. I did so to view the fine vintage racers and restomods (you guys call them PT). One 69 sportsroof had a canted 4 bar installed. I did not note the manufacturer. The car was in for the second time to replace the bracing and sheet metal at the chassis for the UCA's. I thought perhaps the install was not correct when I started doing some research. That is when I found the info. on CC.com. The UCA's can't follow the same arc in roll. The joints are at different angles. I do not have the means to build and test this, few do. This is why the MM test is often quoted.

Bryce,

How is your canted 4 link different than a fox or SN95?

[Bryce]

Mine is a modified TCP gbar. The chassis attachment of the UCA's is closer together than the axle mount. This will usually yield a slightly lower RC and other things....

[Norm Peterson]

I don't see this test as a justification for MM's 3 lower poly bushings. They tested worse than heims and stock components in the LCA's.
But better than plain one-piece cylindrical poly, which I suspect was the intent.



If the uppers cannot change length due to heim joints what does the 0.010" (your number not mine) length change do?
It is only a ballpark number for the axial compression of the steel link under the load it will frequently see in moderate driving (PL/(AE) = 500 lbs x 8" long /1 sq. in. / 29,000,000 psi). Perhaps I should have written the number, but a ten-thousandth of an inch is 0.0001". This is not enough movement to change the geometry or relieve the stress in the poly and is mostly recognition that while the lengths do change it's not significant and doesn't really count as an explanation. Even under very hard driving, you'd still be under 0.001" link compression.



It changes the UCA bracket into a 3rd spring.
All suspension brackets have finite stiffness, which makes them all "springs". FWIW, the stiffness of some brackets can be as low as only a little higher than the bushings in the ends of the attached links (this last little note comes from an OE engineer having international experience).


The details of the installation that you saw are very important. A bracket attachment will repeatedly fail somewhere nearby due to cyclic stress/fatigue cracking if the stress developed is high enough (and this stress level may be lower than you think, what with things like impact loading due to clutch-drop, wheel hop or brake hop also involving an impact multiplier). If the bracing was not a closed tube of some sort with the bracket welded directly to it, I'd absolutely expect this sort of thing.



Norm

[exwestracer]

Hmmm...maybe GM will start recalling 1964-2000 full frame cars.

The more down angle introduced into the upper arms, the more bind in roll the system will have. If the upper links are close to parallel with the lowers AND the system is free to rotate about it's natural roll center, there should be no need for the upper links to change length.

And now if you'll excuse me, I have to pop over to another thread, because I think I just figured out a problem...