Linked Suspensions Bible

[gt1guy]

Running 100% AS ( or even really close to it) will cause wheel hope during braking too. The more traction you have the more it will rear it's ugly head.

We also can't leave out the effects of running really low tire pressures when talking about wheel hop.

 

[Treefrog]

Been meaning to run this data for a while. So 3 link travel and flex.

The setup and travel range was the same for all 4 tests. The links are all parallel (except the panhard), except as noted. As this is a common front axle setup, it was modeled with the upper link on the passenger side, and the panhard running from driver frame to passenger axle. The panhard was flat.

The test conditions were:

All 3 links parallel and flat.
The upper link inboarded at the axle but still flat.
The upper link lowered at the frame, but still parallel to the centerline when viewed from above.
The lower links angled in at the frame.​

The first three test conditions have a flat roll axis, while the one with the angled lowers has an understeering bias roll axis.

Unsurprisingly, all the modified setups behave close to the control setup near ride (the center). The two tests with the upper link moved both showed very little difference from the control at the more extreme conditions (the corners). The test with the lowers angled inward at the frame had up to10x the flex steering as the other modified conditions. That said, 1 degree is not a lot of flex steer.

Symmetry is X" up on the driver side and Y" down on the passenger side produces the same flex steer amount as X" up on the passenger side and Y" down on the drivers side. All four 3-link tests did not have symmetry. In order from least symmetric to most: angled lowers, all parallel, upper inboarded, and upper frame lowered.

Looking at the difference between the three changed conditions and the control conditions, only the angled lowers showed enough difference to be of note.

Regarding further testing, I don't see a need to test the difference caused by the upper switching vehicle sides. given that changes to it resulted in very little change to the results. I also don't see a need to change the panhard's static orientation as that will only shift the 0" deflection lines of the plots.

 

[Treefrog]

Portals. Great for clearance. But how do they affect all of our numbers? Thanks to No way for bringing it up in the How's my numbers thread. Didn't have an answer when he asked, but that gave me a reason to do some reading.

Portals do not affect the roll axes. They are not affected by the torque from the boxes. The same applies to bump steer.

But that still leaves the anti's. These are not geometric values, they are descriptions of how the suspension reacts to acceleration. Interestingly enough, portals do NOT affect the calculations anti's.

From Milliken and Milliken's Race Car Vehicle Dynamics, section 17.3:

If the control arms react torque, either from the brakes or from drive torque, the anti's are calculated by the instant center location relative to the tire/ground contact patch. If the suspension does not react drive/brake torque, but only the forward or rearward force (e.g., inboard brakes), then the anti's are figured from the instant center location relative to the wheel center.

This means that the method of calculating the anti's for portal axles is as shown earlier in this thread. This is also the way it is done in all versions of the 4-link and 3-link calculators.


However, because portals increase the lever arm from anything pushing against the tire, the links and mounting brackets need to be made stronger.

There was also mention of caster in regards to portal boxes and determining the center of the tire relative to the tube centerline. Long story short neglect caster; to get a 5% change in height you would need to run 18* of caster. With typical portal heights being around 4.5 inches, that's more than acceptable.

 

[Treefrog]

Onto the last of the small topics that I can think of. How the anti values affect the ride going down a dirt road at a steady speed. Everything else held constant, the lower the anti value, the smoother the ride.

This has to do with lower anti values having a less upward force on the body. This upward force is transferred through the links instead of the springs. The links do not store the force, they transmit it. So when the wheels see an upward force (hitting a bump) the links transmit some of that to the chassis instead of the springs and shocks. Force means acceleration, so the occupants feel the bump. The higher the anti values, the higher the portion of the force the links see.

 

[AgitatedPancake]

I've been lurking around this thread since it began, but haven't had the patience to sit down and collect a somewhat coherend thought process hah. It's been a long time since I've had time to properly digest a tech heavy forum post like this vs the average social media crap. These are some of my thoughts on AS, and I'm absolutely down for discussion/correction of what i currently think. I just posted this in the new IFS 101 thread, but figured it's very applicable here as well

One of the best downright tech threads just about grasping some of the dynamics of anti squat/anti dive was by ThinAirDesigns back at the old place years ago, I'll be excited to rehash those topics here. One of the biggest ah-hah lightbulb moments I had back then was the fact that your antis (Anti dive and Anti squat) are actually comprised of two significant geometries/forces that work together to create the effect you feel. Torque based anti squat, and thrust based anti squat. Thrust based anti squat is your wheel recession/kick/whatever term you want to describe the wheels front to rear motion through travel. An example of this would be in a rear axle, where the wheel moves rearward in compression, and forward in droop. As you accelerate, the fact that the tire is trying to drive forwards against the chassis causes it to force the suspension into droop, cancelling the squat from weight transfer to whatever degree (based on AS %). When it comes to torque based anti squat, it's the actual rotational force of the axle housing that causes the suspension to unwind under torque, even if the tire is moving perfectly vertically at that point in travel. The reality is 99.9% of the time you're working with a combination of the two which create your final AS number, but the proportion of which is which is rarely touched upon. But it directly impacts what I would call "compliance" of the suspension, how resistant it is to compression when you hit an object at speed, as that directly changes. The more your AS is based on thrust, the more compliant your suspension is to impacts at speed because the wheel regresses through compression. While AS that's primarily based on torque doesn't have that inherent compliance because the wheel is moving vertical.

The easiest way to see it on paper and your standard calculators, thrust based AS is when you have an extremely long IC. For example, if you have 100% AS but your IC is set near infinity, essentially your links are exactly parallel to your AS line so your wheel has a silly amount of front to rear movement through travel. Compared to torque based AS, where to acheive the same 100% you have an extremely short IC that rests on the AS line at the same height as the centerline of your axle. So you still have 100% AS, but your axle is moving directly up and down with no progression/regression (at that static point). Those are the two extremes of the example, but help me visualize the concept.

 

[AgitatedPancake]

Sorry I seem to have derailed this thread a bit, but I wanted to expand on my above thoughts with at least one more note, and a big reason why I like to analyze the two independent actions that combine to create these AS forces. My understanding is fluid and I'm always open to expanding/correcting my thought process, so let me know if I'm off anywhere!

The scenario: Two gear portal boxes.

With a two gear portal solid axle, you flip the pumpkin and run the axle shafts backwards so they can be reversed by the two portal gears (converted into forwards motion). The torque applied to the axle housing is reversed. Now under load, instead of your pinion wanting to walk upwards toward the sky, it's trying to twist and lever it's way down towards the ground.

In my eyes - this means the Torque portion of your anti forces are now having a negative effect on your suspension. That driveshaft torque will now pull the axle towards the chassis, instead of shove the axle away form the chassis as in traditional calculations. So calculating true anti squat with 2 gear portals gets messy because you have the thrust aspect of your suspension doing it's job like normal and creating anti squat, but the torque based component of your geometry is now pro-squat, and cancelling AS forces created by thrust/compliance to a calculated degree.

At least that's my current mindset on the subject!

 

[No way]

Interesting theory G.
But, with the portal gears working in opposite directions and the lower stub having the traction force of the wheel - the rotational forces will fight each other and also vary depending on the amount of traction you have.

Or not.. I may be having a minor aneurism

 

[Treefrog]

Sorry I seem to have derailed this thread a bit, but I wanted to expand on my above thoughts with at least one more note, and a big reason why I like to analyze the two independent actions that combine to create these AS forces. My understanding is fluid and I'm always open to expanding/correcting my thought process, so let me know if I'm off anywhere!

The scenario: Two gear portal boxes.

With a two gear portal solid axle, you flip the pumpkin and run the axle shafts backwards so they can be reversed by the two portal gears (converted into forwards motion). The torque applied to the axle housing is reversed. Now under load, instead of your pinion wanting to walk upwards toward the sky, it's trying to twist and lever it's way down towards the ground.

In my eyes - this means the Torque portion of your anti forces are now having a negative effect on your suspension. That driveshaft torque will now pull the axle towards the chassis, instead of shove the axle away form the chassis as in traditional calculations. So calculating true anti squat with 2 gear portals gets messy because you have the thrust aspect of your suspension doing it's job like normal and creating anti squat, but the torque based component of your geometry is now pro-squat, and cancelling AS forces created by thrust/compliance to a calculated degree.

At least that's my current mindset on the subject!

Interesting theory G.
But, with the portal gears working in opposite directions and the lower stub having the traction force of the wheel - the rotational forces will fight each other and also vary depending on the amount of traction you have.

Or not.. I may be having a minor aneurism

I think you are both on to something by looking at the forces and torques. The ratio between the diff and hub gearing probably changes how much pro-squat it adds.

I think this leads to finding the forces on the links, then taking the sum of the forces in the longitudinal and vertical directions. Finding the slope of a line from those forces and using the IC to define the line's position. The height of where this line crosses the CG*drive bias is then used for the anti-squat and anti-lift calculations.

This also means that differential gearing affects the values as well, even without portals. Or am I headed down the wrong path?

 

[No way]

This also means that differential gearing affects the values as well, even without portals. Or am I headed down the wrong path?

I think you'd have to get into extreme diff ratio variations for that to be a significant factor. I'd say tire size is a much more significant torque factor in this perspective.
Most of this is best guess, experience and put shit where it fits anyways so I don't see it as an exact science.

You'd have to calculate gear ratios and portal reduction ratios and even look into what I'm doing with setting a certain amount of rake to the portals.
I don't math goodly enuff fo dat sheets.

 

[Wendle]

I got shut down by some pretty knowledgeable people on the old board for bringing up what is mentioned in the last few posts about portals, but from the drivers seat that is exactly what it feels like - when it gets real steep the pinion is trying to suck the car back into the face of the obstacle, rather than trying to climb the ring gear and throw the car over.

I've run portals for 15 years now and I reckon I can now feel the difference not having backward-spinning shafts makes if I'm in someone else's rig...

 

[Treefrog]

I got shut down by some pretty knowledgeable people on the old board for bringing up what is mentioned in the last few posts about portals, but from the drivers seat that is exactly what it feels like - when it gets real steep the pinion is trying to suck the car back into the face of the obstacle, rather than trying to climb the ring gear and throw the car over.

I've run portals for 15 years now and I reckon I can now feel the difference not having backward-spinning shafts makes if I'm in someone else's rig...

The more I think about it the more I think that the pinion wants to do what it does without portals. The wheels and tires require torque to go forward, and being solid axle the housing takes the torque. Equal and opposite reactions says that that torque can only go opposite the wheel and that it has to equal the torque to the wheel no matter what. Which would imply that nothing changes in terms of determining antis on portal axles. The only thing portals change is the leverage on the links and add torque to the tubes.

 

[AgitatedPancake]

I've been slow to respond as I crunch on that same dilemma Tree. Equal and opposite seems to dictate it would be such. I think I just need to physically tinker with a portal housing to make things click in my mind. The pinion is rotating the opposite direction which means it would want to walk around the ring gear opposite of a non portal housing. But then you have to factor in the fact that the portal boxes themselves are creating a torque upon the housing which a traditional wheel bearing does not. And if I'm interpreting it correctly, the torque of the portal boxes themselves would try to make the pinion rise, so those two examples combat each other.

I'm really not sold on either end of the spectrum yet, but I want to confidently understand.

 

[Wendle]

The more I think about it the more I think that the pinion wants to do what it does without portals. The wheels and tires require torque to go forward, and being solid axle the housing takes the torque. Equal and opposite reactions says that that torque can only go opposite the wheel and that it has to equal the torque to the wheel no matter what. Which would imply that nothing changes in terms of determining antis on portal axles. The only thing portals change is the leverage on the links and add torque to the tubes.

Does my head in. If you think of a rear tyre jammed into an undercut it can't possibly climb, the car is now essentially a stationary machine. Now, no matter what happens upstream the rotational force at the wheel hub is the end product of that machine and should be trying to roll the car over backwards.

But sitting in a car with two-gear portal hubs in that exact situation, you can physically feel the the car fighting itself to stay exactly where it is.

I gave up trying to rationalise it years ago, it's just how it is.

 

[Treefrog]

Does my head in. If you think of a rear tyre jammed into an undercut it can't possibly climb, the car is now essentially a stationary machine. Now, no matter what happens upstream the rotational force at the wheel hub is the end product of that machine and should be trying to roll the car over backwards.

But sitting in a car with two-gear portal hubs in that exact situation, you can physically feel the the car fighting itself to stay exactly where it is.

I gave up trying to rationalise it years ago, it's just how it is.

But if the wheel is going one way, the housing is trying to go the other. Hold the wheel stationary instead of the axle and the axle spins opposite to how the wheel was.

 

[Wendle]

But if the wheel is going one way, the housing is trying to go the other. Hold the wheel stationary instead of the axle and the axle spins opposite to how the wheel was.

Yeah I know. The science doesn't marry up with how it behaves (or feels as though it's behaving through the seat).
I'd be keen to know if others with portal rigs have noticed the same thing, I don't think there's anything too unusual about my setup...

 

[gt1guy]

I don't think what's being described with/about portals is actually a function of AS per se. AS is merely the suspension geometries ability to carry xx% of the longitudinal load transfer (weight) in the links. The suspension is still free to move the whole time. It's also a very short period of time that AS is in play.

The example of a tire jammed into an undercut takes AS out of the equation. There is no load transfer taking place. But it does sound geometry driven.

 

[AgitatedPancake]

I don't think what's being described with/about portals is actually a function of AS per se. AS is merely the suspension geometries ability to carry xx% of the longitudinal load transfer (weight) in the links. The suspension is still free to move the whole time. It's also a very short period of time that AS is in play.

The example of a tire jammed into an undercut takes AS out of the equation. There is no load transfer taking place. But it does sound geometry driven.

One of the factors that combine to create AS is the torque on the axle housing levering against the links and forcing the axle towards or away from the rig, that part I'm confident on. But trying to figure out that action with the added complication of portals still has me spinning in circles

 

[gt1guy]

One of the factors that combine to create AS is the torque on the axle housing levering against the links and forcing the axle towards or away from the rig, that part I'm confident on. But trying to figure out that action with the added complication of portals still has me spinning in circles

Agreed. I probably didn't explain my thoughts very well, because I'm only really scratching the surface of understanding it all.

There does have to be a torque applied for AS to carry a % of the lateral load transfer in the links. Proof of this comes from the Mark Ortiz newsletter where he talked about the torque split in a transfer case. In 2wd, all the torque goes to the rear tire pair. If your rear geometry is set up for 100% AS, you'll have 100% AS, but, in 4wd with a 50/50 transfer case split, you'll have 50% AS in the rear with the same link geometry. If you design for 100% AS in 4wd, you'll have 200% when in 2wd and so on.

Without a torque applied there is no AS. There also isn't any load transfer.

I think we need to look at portals the same as you would the brakes. Under braking it's no longer Anti-Squat in the rear..............it's Anti-Lift. That's due to the lateral load transfer now moving forward. Here's the big but.............but the torque being applied to the rear axle is exactly opposite as it is during drive. At least as far as the link geometry knows.

 

[Treefrog]

Agreed. I probably didn't explain my thoughts very well, because I'm only really scratching the surface of understanding it all.

There does have to be a torque applied for AS to carry a % of the lateral load transfer in the links. Proof of this comes from the Mark Ortiz newsletter where he talked about the torque split in a transfer case. In 2wd, all the torque goes to the rear tire pair. If your rear geometry is set up for 100% AS, you'll have 100% AS, but, in 4wd with a 50/50 transfer case split, you'll have 50% AS in the rear with the same link geometry. If you design for 100% AS in 4wd, you'll have 200% when in 2wd and so on.

Without a torque applied there is no AS. There also isn't any load transfer.

I think we need to look at portals the same as you would the brakes. Under braking it's no longer Anti-Squat in the rear..............it's Anti-Lift. That's due to the lateral load transfer now moving forward. Here's the big but.............but the torque being applied to the rear axle is exactly opposite as it is during drive. At least as far as the link geometry knows.

There is two types of torques on the system. Drive torque and reaction torque. Drive torque does not affect the suspension. Reaction torque does, depending on suspension type. Reaction torque is applied at the differential. This is why ifs suspensions start the line to the IC in different places for anti-squat and anti-lift.

In solid axles, the drive torque creates a force on the axle, because the wheel/tire is pushing against it and it creates a reaction torque on the housing. When in 4wd, the force is being split between both axles. The % that it is being split is a function of how much force each end of the vehicle is contributing to the acceleration. It is very possible to have 100% front and 100% rear while in 4wd.

The problem with portals that we are scratching our heads over, is that you have a combination of reaction torques. You have the torque from the diff, that is opposite in direction when compared to a normal solid axle. But you also have the reaction from the gears in the portal boxes. The part of this that is being tricky is how the combination of these torques affect the suspension. Is it trying to rotate the pinion opposite to the tires like with a normal axle, or is it trying to rotate the other direction?

 

[gt1guy]

It kind of sounds like the two opposing torques should cancel each other out until one wins.

In cave man speak that I can understand, if one torque input is say 100lb/ft (R&P) and the other in the opposite direction is 75lb/ft (portals), then wouldn't the result be 25lb/ft of torque in the R&P direction that the suspension geometry has to work with?

The axle tubes are pretty much being used as a torque tube at this point are they not?

 

[Treefrog]

It kind of sounds like the two opposing torques should cancel each other out until one wins.

In cave man speak that I can understand, if one torque input is say 100lb/ft (R&P) and the other in the opposite direction is 75lb/ft (portals), then wouldn't the result be 25lb/ft of torque in the R&P direction that the suspension geometry has to work with?

The axle tubes are pretty much being used as a torque tube at this point are they not?

Right on all cases. The next problem is that with no torque reaction the antisquat line goes to the hub, with torque reaction, it goes to the contact patch. So where does it go with %50 of the torque or torque in the other direction?

 

[No way]

As I see it, the rotational force of the wheel will still attempt to jack the pinion upwards. The wheel has considerably much more force and leverage than the portal gears so even if the counter rotating upper portal gear will "dampen" the wheel's ability to lift the pinion, the force is still there.
I see all these portal rigs with monolink/triangle setups getting mad wheel hop on inclines and the ass end rising as they throttle on. A properly linked portal axle does not behave like that.

 

[AgitatedPancake]

I just had a small "ah-hah" moment. Tree, thank you for the reminder on how the toque is isolated by the pumpkin chassis mounts in independent suspensions, and how that brings the start of the AS line to the center of the wheel hubs due to the lack of torque based AS. Something I knew and analyzed in years past, but had forgotten until you just brought up, and gives a solid numerical comparison of torque based and thrust based anti's which my memory was failing on.

That is a blessing in disguise though, because it also helps me mentally separate the pinion based and portal based torques directly. Just for this thought process, the analysis becomes - what do portals do to independent suspension anti numbers? In this example that ignores housing torque at the third member, 4 gear portals and 2 gear portals will be making a direct difference in antis with no other factors dilluting the nubers. For this discussion, we assume the same amount of reduction between 2 and 4 gear (which has it's own interesting side tangent, what happens at a 1:1 ratio with 4 gear portals?)

I still don't have confidence in my opinion, but this re-realization is helping me visualize the forces more directly

 

[Treefrog]

I just had a small "ah-hah" moment. Tree, thank you for the reminder on how the toque is isolated by the pumpkin chassis mounts in independent suspensions, and how that brings the start of the AS line to the center of the wheel hubs due to the lack of torque based AS. Something I knew and analyzed in years past, but had forgotten until you just brought up, and gives a solid numerical comparison of torque based and thrust based anti's which my memory was failing on.

That is a blessing in disguise though, because it also helps me mentally separate the pinion based and portal based torques directly. Just for this thought process, the analysis becomes - what do portals do to independent suspension anti numbers? In this example that ignores housing torque at the third member, 4 gear portals and 2 gear portals will be making a direct difference in antis with no other factors dilluting the nubers. For this discussion, we assume the same amount of reduction between 2 and 4 gear (which has it's own interesting side tangent, what happens at a 1:1 ratio with 4 gear portals?)

I still don't have confidence in my opinion, but this re-realization is helping me visualize the forces more directly

No problem, but thank ThinAir for it. Been rereading that thread trying to figure out portals.

Let's leave independent for the 101 thread, but they move the point off of the hub center. But for sake of discussion...
For clarity, 2 gear being what you see in Mogs and Hummer portals. 4 gear really being a 3 gear, with the load split between 2 idler gears. I believe that the FBD show that 1:1 with a 4 gear, you have no change in the reaction torque.

Another though comes to mind with the torques, you have the portal gear reaction torque, the pinion gear reaction torque, but is there also torque from the portal offset?

 

[AgitatedPancake]

Those were great discussions, I should revisit those thought processes myself. I agree independent is another subject all together, but as I try to grasp what portals are doing it helps me remember that what forces they apply to a suspension can be isolated at least to help wrap my head around what it's doing.

That comment about the actual leverage offset is pretty interesting as well - it would make sense that the thrust based forces apply a certain amount of housing torque based upon the offset distance, but calculating it will be quite a bit more interesting

 

[Treefrog]

Those were great discussions, I should revisit those thought processes myself. I agree independent is another subject all together, but as I try to grasp what portals are doing it helps me remember that what forces they apply to a suspension can be isolated at least to help wrap my head around what it's doing.

That comment about the actual leverage offset is pretty interesting as well - it would make sense that the thrust based forces apply a certain amount of housing torque based upon the offset distance, but calculating it will be quite a bit more interesting

I don't see how calculating it would be all that interesting, it's just the force from the wheel times the offset distance, with a little bit thrown in to account for the angles that the portals are at.

 

[vetteboy79]

Drive torque does not affect the suspension.

Well, it does, through torque lean. It's true that a 2-gear portal rig will experience torque lean in the opposite direction of a regular-rotation rig, which goes back to what I mentioned before about asymmetric links acting on the chassis. Side-view analysis of a 3-link won't really quantify that as the upper will either counteract or exacerbate that torque lean behavior depending on what side it's on and the angle it's at.

Otherwise - I think you guys are right that the only difference is regarding the offset. The housing is a fixed member, portals or not - it's only reacting the torque to the wheels, any other reaction forces are internal. So if you have a standard axle with the lower links mounted 6" above the axle tube, and the uppers 6" above that, vs a portal axle with the links at the 'center' of the axle tube (6" above the portals)....that should be the same effect.

In a standard axle with the lowers mounted near the centerline, the lower links are primarily in compression with the uppers in tension to resist axle roll. With them above centerline, now you've got a moment acting around the lower joint, so in addition to handling the axle roll, the uppers are reacting to that moment as well. Basically in my mind - portals or raised lower links will both serve to place more tension on the upper links (compared to the lowers being close to the torque centerline), and looking at the angle those links act on the chassis, the effect will be enhanced.

 

[Austin]

I'm going to bump this up to the top and apologize for kicking these off then bailing at the end of the first one. I got so wrapped up in KoH prep you would of thought I was racing in every event lol.

Treefrog has done an amazing job on this one

I'm going to work on converting this into an article, then we can proof it, push it and as new thoughts/ideas/corrections come up we can post them here.

 

[Treefrog]

Taking a break from portals for a bit. The newest version of the 4 link calc allows me to bring up the subject of what happens when the you allow the body to move while calculating geometry values like antis and roll centers/axis. It goes without saying that exactly what happens depends on the geometry. Of note, it seems that anti values climb to more extreme values than they do with the body held steady. An example of max rear compression and max front droop. Solid lines are the pitched links, dashed are the ride level links. Interestingly enough, with 8 inches rear bump travel and 8.5 inches of droop in the front, the vehicle only pitches 9.1 degrees. However the sprung mass CG raises an inch.

Looking at what else happens when the body can move from flat and level, namely going up or down. My understanding is that the frame of reference is held parallel to the current ground plane, or in other words, rotates with the ground. Therefore the method of calculating the geometry stays the same. But what does change is the load on the springs. For the case of going up the front droops out, but depending on the conditions, the rear can travel in bump or droop.

 

[Toreadorranger]

I'm in the middle of my current build which is leafs front and rear, Ive been contemplating linking the rear (triangulated) based on ease and available space. I know a lot of the conversations are centered around rigs linked front and rear. Is there any differences in design numbers that should be accounted for when only doing the rear vice both or just the front?