57 replies to this topic

[cvo]

Hello all. Long time follower and now first time poster.

I've come across two approaches to dampers and transient load transfer which seem to contradict each other. The first is that dampers provide roll/pitch resistance when you have roll/pitch velocity, providing another layer of control over the load transfer distribution in transients. As an idealized example, going into a flat-out corner, increasing bump/rebound stiffness at the front will induce more understeer in the entry phase.

The second approach seems to be related to the damper's influence on the rate of load transfer. So using the same example as above, increasing damping would slow the rate of load transfer and possibly cure an understeer in the entry phase.

While I go about fine-tuning dampers with the roll resistance principles in mind, my physical understanding leads me back to the ideas in the 2nd paragraph. If I consider each corner as a spring-mass-damper and that the elastic load transfer is related to spring displacement, a stiffer damper would tend to slow weight transfer (ignoring settling time).
Hopefully a discussion may clear up my confusion and please correct me if I'm wrong with anything above.

[gruntguru]

I've come across two approaches to dampers and transient load transfer which seem to contradict each other. The first is that dampers provide roll/pitch resistance when you have roll/pitch velocity, providing another layer of control over the load transfer distribution in transients. As an idealized example, going into a flat-out corner, increasing bump/rebound stiffness at the front will induce more understeer in the entry phase.

The second approach seems to be related to the damper's influence on the rate of load transfer. So using the same example as above, increasing damping would slow the rate of load transfer and possibly cure an understeer in the entry phase.

While I go about fine-tuning dampers with the roll resistance principles in mind, my physical understanding leads me back to the ideas in the 2nd paragraph. If I consider each corner as a spring-mass-damper and that the elastic load transfer is related to spring displacement, a stiffer damper would tend to slow weight transfer (ignoring settling time).
Hopefully a discussion may clear up my confusion and please correct me if I'm wrong with anything above.

I agree with your first paragraph, disagree with the second. Increasing front damping will increase initial load transfer at that end and decrease it at the other. What decreases is initial roll-angle which is often confused with load transfer because load transfer through spring elements is proportional to roll angle. (as opposed to load transfer through damping elements which is proportional to roll velocity)

Edited by gruntguru, 05 August 2010 - 10:05.

[cvo]

Well that was simple... Thanks for clearing it up!

[meb58]

gruntguru,

I think some of us actual use or use to use adjustable dampers to control balance and that's not really the correct way around chassis tuning is it? I guess we can cheat a little where fine tuning is concerned - why not adjust the swaybar? I am learning, however, that dampers can have a huge influence upon how a car feels through the phases of corner despite lacking any really affect upon load transfer...they can make load transfer feel twitchy or or very composed.

Is this just a mater of properly mating a damper and spring or is it much more?

[gruntguru]

I think some of us actual use or use to use adjustable dampers to control balance and that's not really the correct way around chassis tuning is it? I guess we can cheat a little where fine tuning is concerned - why not adjust the swaybar? I am learning, however, that dampers can have a huge influence upon how a car feels through the phases of corner despite lacking any really affect upon load transfer...they can make load transfer feel twitchy or or very composed.

Is this just a mater of properly mating a damper and spring or is it much more?

I am not an expert in this area but I am beginning to work a few things out. As you suggest, there is a damper setting that will maximise grip for the given wheel spring rates (including ARB's, thirds etc) and masses present. Combining these allows tuning of the steady-state cornering characteristics. However optimising steady-state cornering does not guarantee optimal transient behavior. For example if the car runs a spool, it will tend to understeer on initial turn-in. Adjusting dampers can help here without overly affecting the steady-state cornering. Setting the rear dampers a little stiffer and softening the fronts a little will increase the transient roll stiffness at the rear an reduce it at the front - unloading the inside rear on rapid turn-in - allowing it to lose grip and unload the spool. In addition, because the relative load transfer rear-front is greater than the steady-state setup, the transient tendency has been shifted towards oversteer without significantly affecting the steady-state handling balance.

Edited by gruntguru, 09 August 2010 - 23:10.

[mariner]

If one wanted to be very clever , and/or , totally confuse oneself, you could decouple the spring and damper linkages and set up the dampers for falling rate even if the springs were linear or rising rate.

then you can get a "sharp" initial damper influence on the transient whilst still having a " softer" damper influence on the taking up of final wheel load.

[GrpB]

you could decouple the spring and damper linkages and set up the dampers for falling rate even if the springs were linear or rising rate.

Suzuki TL1000

If one wanted to be very clever , and/or , totally confuse oneself

Absolutely the latter.

[gordmac]

Bear in mind the effective spring rate in roll is likely different to pitch (anti roll bar) meaning "optimum" damping will be different for each type of motion.

[meb58]

I am still playing with the calculator you sent me I thought i understood your reply until I got to (anti roll bar)...maybe it's where you placed it in the sentence. Anti roll bars aren't influenced by pitch, correct? I thought I should ask just to confirm what I thought i knew...

Bear in mind the effective spring rate in roll is likely different to pitch (anti roll bar) meaning "optimum" damping will be different for each type of motion.

gruntguru,

I guess I've always been able to adjust my cars to achieve comfortably good steady state; reasonable grip and balance. But more often than not, because I really don't know $%@, corner entry is very spooky and I attribute this to damping. Now, I've used adjustable and non-adjustable dampers and this may be just how life is, me over driving the car or perhaps a better marriage between spring and damper exists.

Mariner,

Can you explain in more detail?

I've always thought that if someone asked if I could spend a day with a suspension engineer discussing only one subject it might have to be ride and roll.

Edited by meb58, 10 August 2010 - 17:07.

[mariner]

In more detail.

Dampers ( if I have got it right) react to velocity and not load.So as velocity is movement/time you put the damper on a linkage which is falling rate. That is the initial hub movement creates a large damper movement but each later equal hub movement gives a diminishing damper movement.

Thus the damper sees high initial velocity but lower final velocity. This will increase damper control of small bumps and initial turn in even if the dampers own characterstics are linear.

The springs could have constant or rising rate seperately from the damping rate. It is not complex mechanically but I suspect the permutations would get very complex espeically with a third spring and damper.

Anyway just an idea.

[gordmac]

The anti roll bar adds effective spring rate in roll not pitch, the different spring rates may want different damping values.
Pure damping gives a force proportional to velocity, this will create forces in transient motion in addition to the spring forces (the relevant inertia is important here). The transient motion can be due to roll, pitch etc (handling) or bumps (ride), I suspect a car spends very little time in steady state. You also get friction damping, generally thought of as a bad thing. If you can imagine a soft car, push down on it then lift it, you will find the ride heights will be quite different. You get the damper force vs velocity from a damper dyno, depending on how your damper works it takes a finite time for the damper to reach the steady state values you get from the dyno. This can be an issue if your damper motion is small.
Putting the damper on a falling rate mechanism is interesting, would the relatively lower damping with travel cause a problem? Would the idea be to have a higher than "optimum" value initially?
How would this interact with the "knee point"?

[meb58]

Okay, glad to read that the SB still works the way I thought it did

In addition to learning much more about dampers and spring relationships, what about those multicellular jounce dampers - Bump stops? I am not referring to the old hard rubber type...I am referring to the type that appear to be in constant contact with the dampers adding some progressive(?) rising (?) rate to the suspension system.

In one of my posts above I described some very unsettled corner entry feelings and no matter what I tried I could not dial this out...I often wondered if these components had something to do with the uneasy feeling. And, if a car is lowered by 10mm - mine was - this might exaserbate the effect.

[meb58]

This all makes great sense to me at the conceptual level. The testing will produce another phase of learning.

The car in question is a 2005 JCW mini - now with JCW suspension - a stock upgrade. The 'bumpstop' is in contact with the damper at ride height - sitting in the pit - and has plenty of room to expand; it is covered by a rather generous plastic dust shield. The bumpstop is conically shaped - fatter at top - and does a two grooves cut into it...the lower one being a smaller circumference since it's in a smaller part of the cone.

I have often thought about buying a bunch of these to play with...just to feel the difference...I also wondered what it might be like to drive without them, but the mini does not have lots of front end travel. In addition, I thought about flipping them upside down, but oh goofy me, the result would be the same. As you wrote, shape, among other things matter.

As you pointed out, they did appear to work a bit more abruptly in cold; while at Lime Rock in november a couple of seasons ago track temps were below freezing. I swapped the R- Compounds for my normal street fare making an actual comparison difficult but I do remember the car feeling progressive up to a point followed by what I would describe as a vary large step in rate...followed by little slides. I guess the cold made the bumpstops harder...my intuition tells me the street rubber was better suited to those track temps.

Sorry to steal the OP's thread.

Edited by meb58, 12 August 2010 - 16:39.

[meb58]

Wow! Thank you again.

Interesting out non-intuitive logic can be; I thought about making the bumpstop softer, but in fact, as you suggest making it harder may prove out to be better. In any event, I have some home work to do and some tinkering to look forward to!!!

And again, as I noted way above, I am not sure that the car will be any faster, but if my confidence improves upon corner entry then I can improve my lap times a bit.

A curious historical question comes to mind...what precipitated this type of bumpstop? Essentially a progressive rate bumpstop will make even a linear spring work in a progressive way...so why not just install a progressive rate spring? Do we have more control - greater fidelity - with a bumpstop of this type?

[gruntguru]

Hi Meb. If your damper is in contact with the bump-stop sitting in the pits, chances are you have insufficient bump travel ie the unsettled feeling is the result of extreme compression of the bump-stop and spring rate approaching infinity. If this is the situation, removing the bump-stops will result in harsh bottoming (an even steeper step in spring rate). The solution is shorter dampers (if you can get them) or moving the top mount up.

[meb58]

gruntguru,

I am told that the new mini is a bumpstop active car - meaing the bumpstops are supposed to be in contact with the top of the dampers at all times. The current set up is an upgrade, but still stock components produce for BMW/Mini by John Cooper Werks. In any event, these bumpstops have lingered in my mind for quite a while...and being distracted by many other parts of daily life, I chose not to jump into a pot I knew nothing about. But you and TC3000 have given me ample reason to concentrate on those.

TC3000,

Since this car is also a daily driver - about 50K - 60K a year - I spend a fair amount of time and money replacing things like tie rods, bushings and bumpstops. I have noticed the color change in the bumpstops - front especially - and wondered what caused this...sort of a "hmmm, wonder what causes that?" moment and nothing more.

Regarding your note about the bumpstop not 'rebounding' back to original shape...I get that. At LRP - Lime Rock Park - I am able to brake at brake point 2 but the middle of the track is somewhat bumpy. I imagine that the above condition is helping to unsettle the car...so a question comes to mind...the rate at whcih these 'rebound' is affected by shape and material? Sorry if rebound is the wrong term.

I'm actually a littel giddy about working this problem out now! My other challenge is finding a place to dyno the stock JCW springs and dampers...this stuff is a secret and each spring set is tethered to a car's VIN #. I'm told that there are a few different rates and each is matched to the car's weight...maybe i also have the wrong springs. Perhaps I will find a damper - Koni? - that is better matched to the bumpstop frequencies available.

Edited by meb58, 13 August 2010 - 12:35.

[gruntguru]

gruntguru,
I am told that the new mini is a bumpstop active car - meaing the bumpstops are supposed to be in contact with the top of the dampers at all times. The current set up is an upgrade, but still stock components produce for BMW/Mini by John Cooper Werks. In any event, these bumpstops have lingered in my mind for quite a while...and being distracted by many other parts of daily life, I chose not to jump into a pot I knew nothing about. But you and TC3000 have given me ample reason to concentrate on those.

Sorry - just checked your earlier post and noticed the car is only lowered 10mm which shouldn't be causing bottoming.

The colour change could be temp related. Shocks, bump rubbers and nearby brakes all contribute heat.

[DaveW]

A few comments, if I may.

Mariner: Apologies, but I think you will find that a force applied across a damper creates a pressure differential across the piston, causing metered flow of fluid from one (high pressure) side of the piston to the other (low pressure), resulting in a velocity of the damper shaft relative to the damper body. Hence the independent (input) variable is force, and the dependent (output) variable is velocity.

I am not a fan of using "active" bump rubbers in "tin top" race vehicles, for a couple of reasons. First, they are not consistent either in the short or long term. TC3000 has discussed short term issues, & I have direct evidence that they also take a permanent "set" with use. Second, they "couple" set-up changes. So, for example, increasing rebound damping will tend to "jack" the vehicle dynamically into its bump rubbers & so increase the average spring stiffness. For what it's worth, I think that long bump rubbers are used in production road vehicles to allow the static ride height to be lowered (largely a stylist's requirement) & as an alternative to fitting effective anti-roll (sway) bars.

I use a multi-post rig to characterize vehicle suspensions (that is, BTW, the best way to understand & to optimize suspension set-ups), & I help a fair few touring & GT teams each year. I guess it has been 10 years since I have seen "active" compound springs or bump rubbers used in racing tin tops, with one notable exception (& that exception provides easy income for me).

My advice for a serious "track day" conversion would be to cut down or replace "active" bump rubbers & to compensate by increasing spring stiffness. To make the revised set-up work, you may have to install a working & adjustable arb & you may have to replace existing top mounts to avoid losing control of hub modes as damper settings are increased. Tin tops (having a large roll inertia) do sometimes develop a "stick slip" roll/pitch instability, which can be solved by moving stiffness from bars to springs & re-adjusting damper settings to suit the increased spring rates. Road car dampers are often rebound-biased for "ride" reasons. Making the damping more evenly split can (depending on damper architecture & vehicle geometry) allow the static ride height to be lowered without increasing the risk of running out of damper travel. That will often help to improve lap times significantly. Before launching into any of these changes, however, my advice would be to spend a day on a multi-post rig to gain a better understanding of your base vehicle.

Edited by DaveW, 14 August 2010 - 06:14.

[johnny yuma]

Various tru-isms at times seem to contradict each other.A spring/damper/bump stop on full compression --bottomed--will lose road grip on that corner and slide to some extent.However downforce applied by aero effect will increase grip on cornering.Why is it so ?

[DaveW]

Put crudely, available "grip" is proportional to, amongst other things, the minimum dynamic contact patch load. This can be thought of (still crudely) as (static load) + (aero load) - (load variation). Hence grip will be increased if aero load is increased, and will be decreased if load variation is increased.

Load variation in response to a fixed road input will be decreased by a "good" mechanical set-up, whilst aero load will be increased by a good "aero" set-up. The two are incompatible, at least in principle, because a good mechanical set-up "consumes" more ride height variation, whilst a good aero set-up requires (usually) less ride height variation.

It follows, from the argument at least, that chasing a good aero set-up generally implies accepting a worse mechanical set-up. The chase will be successful whilst the gain in aero load is greater than the consequent increase in load variation. There are, however, other reasons why the chase after a good aero set-up to the exclusion of all else might appear to be the best option.

[gruntguru]

To take that a little further, suspension bottoming alone does not reduce grip, in fact while a particular corner is at maximum bump (and assuming zero damper force) that corner will have increased grip levels. It is the resulting load transients/fluctuations that reduce overall grip as Dave says.

[meb58]

gruntguru,

Meaning that when a suspension has bottomed any additional suspension input - bump - quickly or abruptly saturates the tire and a slide begins?

DaveW,

I am not sure I understand hub modes...do you literally mean the hub? If so, doesn't the hub move in sync - velocity - with the damping system - spring and dampers? Or is tire deflection a part of this characterisitic too? I am astonished that I actually understand some of the information in your reply. Lots of help from folks like grunturu, cheapracer, Greg L. Gordmac and others...

Edited by meb58, 17 August 2010 - 12:39.

[mariner]

Dave W

"Load variation in response to a fixed road input will be decreased by a "good" mechanical set-up, whilst aero load will be increased by a good "aero" set-up. The two are incompatible, at least in principle, because a good mechanical set-up "consumes" more ride height variation, whilst a good aero set-up requires (usually) less ride height variation."

So Lotus were definitely onto something withe the Lotus ** concept then!!

[Tony Matthews]

the Lotus ** concept

I like your use of code there, mariner! Smacks of espionage...

[DaveW]

I am not sure I understand hub modes...do you literally mean the hub? If so, doesn't the hub move in sync - velocity - with the damping system - spring and dampers?

Yes & yes.

However, a "top mount" (compliance) placed in series with the damper will allow the hub to move relative to the sprung mass even when the damper is locked. Top mounts are almost invariably used in road vehicles to isolate the passenger cell from higher frequency road inputs & suspension (mainly damper) "noise". Usually, they are selected to minimize the effect on control over the rigid body modes. However, converting a road vehicle for "track day" use generally means increasing spring stiffness & increasing damper strengths. It may also involve using higher stiffness tyres. Both increased damper strengths & higher stiffness tyres will reduce hub mode damping when top mounts are present, sometimes disastrously so...

[johnny yuma]

Thanks DaveW for your patience and expertise.
If can ask further to top mount/compliance referred to,does this take the form of A-Arms or struts mounts deflecting a little when road irregularities
are hit,or something more complex? Can't form a mental image of the hub moving without the damper moving.What damper or damper mount
is that strong ?

[DaveW]

Suspension compliance (installation compliance) exists in all vehicles. It is caused mainly by structural deformation. A "top mount" is another compliance introduced deliberately in road vehicles to improve subjective "ride". It is an elastomer bush or bearing and is normally installed between the damper (specifically) and the sprung mass (chassis). Hence "top mount".

Both compliances reduce damper "efficiency". The effect is to reduce damping ratios of the vehicle rigid body modes, particularly those of the hub modes because their natural frequencies are the highest. The effect can be large. In one actual example, hub mode damping ratios without top mounts were 60 % of critical, reducing to just 10 % of critical with top mounts installed. The effect will increase when/if damper settings are increased. Hence the concern when an existing vehicle is to be converted to a "track day" role.

[DaveW]

So Lotus were definitely onto something withe the Lotus ** concept then!!

Not sure what you mean, mariner. If you are referring to the Lotus active suspension, then that was introduced specifically to minimize the interaction between aero & mechanical set-up requirements. It was designed to react to road inputs, but not to driver inputs or to changes in aerodynamic forces. That is not something that can be replicated in a purely passive suspension, not when the suspension is the only connection between the sprung & unsprung masses, anyway. Sadly, active suspension has been banned from all major race series, apart from one.

[mariner]

Dave/Tony , my ** was a tribute to Quintin Tarantino and "Kill Bill" , the crazy 88's! No actually of course I meant to type the Lotus 88 - which seperated the mechanical grip from the aero grip by having a set of springs for each load ( and chassis).

I know it was banned but it seemed such an elegant solution to the aero versus mechanical grip problem plus it could have meant no sliding skirts with all their hassles.

[Tony Matthews]

It wasn't a mystery as to how ** appeared! I always thought of the Lotus 88 as being like a cat in a cardboard box with no base, the box skating around the room while the cat had full movement. Amazing what you see when you haven't got your gun...

[gruntguru]

gruntguru,
Meaning that when a suspension has bottomed any additional suspension input - bump - quickly or abruptly saturates the tire and a slide begins?

No, in fact while the suspension is bottomed that wheel has more vertical load and therefore will have more grip. The primary downsides are:
- the increase in vertical load of one tyre usually means at least one other tyre suffers a reduction in vertical load
- the large increase in roll stiffness results in a reduction in lateral compliance. Lateral transients are more likely to overload the contact patch.

[meb58]

DaveW,

Okay, got it...and amazing it is to me, a novice, what factors skip right by the back yard mechanic's gaze!

The car in question had a different setup a while back that inlcuded a very nice camber kit with a hiem joint or pillow ball (not sure where that term came from) and was attached under the strut tower - a solid connection - strut bearing gone. Indeed the top mount or strut tower suffered. Using back yard mechanic's license, I used longer studs and attached a billet piece of aluminum on top of the strut tower. This piece was formed to fit only the shape of the top of the strut tower and the strut tower became sandwiched between this piece and the camber plate below. Until now I never knew about hub frequencies...I am not sure that my modification hurt or helped...but the towers certainly did not deform after.

As you might have guessed, my track setup was not suitable for the roads I drive. The stock setup was slower on the track but it made me a better driver at some level - more patient.

Yes & yes.

However, a "top mount" (compliance) placed in series with the damper will allow the hub to move relative to the sprung mass even when the damper is locked. Top mounts are almost invariably used in road vehicles to isolate the passenger cell from higher frequency road inputs & suspension (mainly damper) "noise". Usually, they are selected to minimize the effect on control over the rigid body modes. However, converting a road vehicle for "track day" use generally means increasing spring stiffness & increasing damper strengths. It may also involve using higher stiffness tyres. Both increased damper strengths & higher stiffness tyres will reduce hub mode damping when top mounts are present, sometimes disastrously so...

gruntguru,

But if a suspension corner has bottomed - assuming there is no more travel - then any additional input in bump essentially makes the whole thing a solid stick or nearly so if we inlcude tire deflection. If the suspension has bottomed but the tire has not saturated then we have nearly maxed the that corner...? Am I making sense? If so, the characteristic of a tire at precisely this moment is crucial to what happens next, not really the fact that the suspension has bottomed...?

Edited by meb58, 17 August 2010 - 12:56.

[DaveW]

Dave/Tony , my ** was a tribute to Quintin Tarantino and "Kill Bill" , the crazy 88's! No actually of course I meant to type the Lotus 88 - which seperated the mechanical grip from the aero grip by having a set of springs for each load ( and chassis).

I know it was banned but it seemed such an elegant solution to the aero versus mechanical grip problem plus it could have meant no sliding skirts with all their hassles.

I should have looked at my keyboard! You are correct about the 88 solution - except that skirts had been banned by the time the definitive version was in build, & that is probably a large part of the reason other teams objected. Would it have worked? I like to think so, although unresolved teething problems remained, thanks to it being black flagged every time it appeared in public....

[gruntguru]

gruntguru,
But if a suspension corner has bottomed - assuming there is no more travel - then any additional input in bump essentially makes the whole thing a solid stick or nearly so if we inlcude tire deflection. If the suspension has bottomed but the tire has not saturated then we have nearly maxed the that corner...? Am I making sense? If so, the characteristic of a tire at precisely this moment is crucial to what happens next, not really the fact that the suspension has bottomed...?

Depends what you mean by "saturated". Adding more vertical load (which is the only direction affected by bottoming the suspension) will not cause loss of traction.

[DaveW]

Good observations have been made in this thread. I would like to summarize these if I may &, perhaps, add to them on occasion.

Dampers are multi-functional. They control the dissipation of disturbance energy arising from road & inertial inputs, they affect vehicle transient response to discrete inputs, they provide driver "feel", and they affect the rate of heat input to tyres. They also affect contact patch load variations and platform disturbance whilst disturbance energy is being dissipated. The first affects "grip" directly, whilst the second affect down-force variations for an "aero" vehicle. Not all of these variables can be "optimized" at the same time. It follows, therefore, that the "best" damper settings will depend to a greater or less extent upon the vehicle, its tyres, its driver & the track.

So, setting up dampers for a race vehicle is a complex task, & is one of the responsibilities of a race engineer. Tools are available to him include "hardware in the loop" tests, mathematical modelling, track tests (on-track measurements and driver feedback) & (last but not least) his experience. I think it fair to state that none of these can achieve an optimum set-up in isolation. It is worth noting, in passing, that "single corner" calculations are unlikely to be productive, except in very specific (& rarely encountered) circumstances. Other simple "rules of thumb" can be helpful, but for minor "tuning", or for "hiding" a persistent problem when track testing. Changes of this nature rarely improve actual vehicle performance, although they may help a driver to make better use of what is available.

Mathematical modelling is the most versatile single tool because it can yield "sensitivity" information at different points during a manoeuvre (essentially, identifying the most efficient changes that will achieve a desired effect), but modelling is unlikely to be useful without hardware in the loop tests, comprising tyre tests, kinematics & compliance tests & multi-post rig tests. The last are, perhaps, the most useful, because they can also yield sensitivity information, but also reveal some important dynamic properties directly & can, at least, achieve a "generic" optimal track set-up, no matter what the starting point (a biased view, I admit, but one that is based on experience).

It can be assumed, I think, that "optimal" overall damping coefficients exist for any given vehicle, spring selection & tyre set-up. Their values will depend upon vehicle & tyre properties, and the definition of optimum. The relationship is complex, & a multi-post rig test is a good (&, perhaps, the best) way to optimize overall damping coefficients. Linear dampers will dissipate disturbance energy most efficiently (provided the vehicle itself is linear), but are unlikely to satisfy other requirements noted above. The "style" of damping is important for these.

It is, perhaps worth noting that the damping "coefficient" describes the energy dissipated per cycle, whilst the damping "style" describes the "shape" of the load-velocity trajectory - where, within a cycle, the energy is dissipated. In general terms, a "digressive" style is one where the damping coefficient reduces with increasing peak velocity, whilst a "progressive" style is one where the damping coefficient increases with increasing peak velocity. There are other properties that are, or can be, important in a damper, including internal compliance and dependencies on temperature, acceleration, position, & time (frequency), but I will ignore these for now.

It is possible to make some generalized comments about damping style, assuming that the overall damping coefficients remain unchanged:

- Moving damping from the rebound to the compression half-cycles will improve minimum contact patch loads (& hence "grip"), but will increase platform disturbance, and the vehicle will tend to increase average ride height on-track. The last should allow a reduced static ride height, which may result in a nett performance benefit.

- Excessive platform disturbance can be controlled by introducing "blow-off", limiting the maximum compressive damper loads (in effect giving up mechanical control if favour of aero control).

- Moving damping from the compression to the rebound half cycles will reduce minimum contact patch loads (& hence "grip"), will reduce platform disturbance, and will tend to reduce average ride height on-track. The last may be a benefit when a regulated minimum static ride height applies.

- Some damper architectures lend themselves to rebound-biased damping, but a modern 4-way adjustable damper is likely to be tolerant of a large range of damping styles.

- A digressive damping style for dampers attached to the steered axle will decrease steering time constant, may introduce a noticeable transient under-steer, and will increase the rate of heat input to the tyres.

- A digressive damping style for dampers attached to the driven axle may compromise traction, may introduce a noticeable transient over-steer, and will increase the rate of heat input to the tyres. The transient over-steer may reduce the magnitude of steering inputs (& hence may reduce turn in speed loss) on the entry to high speed corners.

- A progressive low speed characteristic at either axle will usually improve contact patch load variations for small inputs, but will reduce platform stability and will reduce the rate of heat input to the tyres (not necessarily beneficial for slicks).

- Digressive front damper compression & digressive rear damper rebound can help to reduce vehicle pitch velocity under braking (noting that driver inputs are essentially low velocity).

- Digressive front damper rebound & digressive rear damper compression can help to reduce vehicle pitch velocity under acceleration.

Apologies for the length of this post. I hope some of it may be useful.

[meb58]

Thank you for this follow up DaveW...most of this is above my head so I will have to reach a bit to put it all together.

What exactly is a mulit rig test? I've read about it here before but didn't want to drag down the other threads with that question...this seems to be an appropriate place and time to ask...

Regarding damping coefficients and style - and please keep in mind I am really nothing more than a back yard mechanic. Can I assume that coefficient = velocity and that style = the character of the damping curve? I am not sure what you mean by "Moving damping from the rebound to the compression half-cycles..." or the reverse?

Good observations have been made in this thread. I would like to summarize these if I may &, perhaps, add to them on occasion.

Dampers are multi-functional. They control the dissipation of disturbance energy arising from road & inertial inputs, they affect vehicle transient response to discrete inputs, they provide driver "feel", and they affect the rate of heat input to tyres. They also affect contact patch load variations and platform disturbance whilst disturbance energy is being dissipated. The first affects "grip" directly, whilst the second affect down-force variations for an "aero" vehicle. Not all of these variables can be "optimized" at the same time. It follows, therefore, that the "best" damper settings will depend to a greater or less extent upon the vehicle, its tyres, its driver & the track.

So, setting up dampers for a race vehicle is a complex task, & is one of the responsibilities of a race engineer. Tools are available to him include "hardware in the loop" tests, mathematical modelling, track tests (on-track measurements and driver feedback) & (last but not least) his experience. I think it fair to state that none of these can achieve an optimum set-up in isolation. It is worth noting, in passing, that "single corner" calculations are unlikely to be productive, except in very specific (& rarely encountered) circumstances. Other simple "rules of thumb" can be helpful, but for minor "tuning", or for "hiding" a persistent problem when track testing. Changes of this nature rarely improve actual vehicle performance, although they may help a driver to make better use of what is available.

Mathematical modelling is the most versatile single tool because it can yield "sensitivity" information at different points during a manoeuvre (essentially, identifying the most efficient changes that will achieve a desired effect), but modelling is unlikely to be useful without hardware in the loop tests, comprising tyre tests, kinematics & compliance tests & multi-post rig tests. The last are, perhaps, the most useful, because they can also yield sensitivity information, but also reveal some important dynamic properties directly & can, at least, achieve a "generic" optimal track set-up, no matter what the starting point (a biased view, I admit, but one that is based on experience).

It can be assumed, I think, that "optimal" overall damping coefficients exist for any given vehicle, spring selection & tyre set-up. Their values will depend upon vehicle & tyre properties, and the definition of optimum. The relationship is complex, & a multi-post rig test is a good (&, perhaps, the best) way to optimize overall damping coefficients. Linear dampers will dissipate disturbance energy most efficiently (provided the vehicle itself is linear), but are unlikely to satisfy other requirements noted above. The "style" of damping is important for these.

It is, perhaps worth noting that the damping "coefficient" describes the energy dissipated per cycle, whilst the damping "style" describes the "shape" of the load-velocity trajectory - where, within a cycle, the energy is dissipated. In general terms, a "digressive" style is one where the damping coefficient reduces with increasing peak velocity, whilst a "progressive" style is one where the damping coefficient increases with increasing peak velocity. There are other properties that are, or can be, important in a damper, including internal compliance and dependencies on temperature, acceleration, position, & time (frequency), but I will ignore these for now.

It is possible to make some generalized comments about damping style, assuming that the overall damping coefficients remain unchanged:

- Moving damping from the rebound to the compression half-cycles will improve minimum contact patch loads (& hence "grip"), but will increase platform disturbance, and the vehicle will tend to increase average ride height on-track. The last should allow a reduced static ride height, which may result in a nett performance benefit.

- Excessive platform disturbance can be controlled by introducing "blow-off", limiting the maximum compressive damper loads (in effect giving up mechanical control if favour of aero control).

- Moving damping from the compression to the rebound half cycles will reduce minimum contact patch loads (& hence "grip"), will reduce platform disturbance, and will tend to reduce average ride height on-track. The last may be a benefit when a regulated minimum static ride height applies.

- Some damper architectures lend themselves to rebound-biased damping, but a modern 4-way adjustable damper is likely to be tolerant of a large range of damping styles.

- A digressive damping style for dampers attached to the steered axle will decrease steering time constant, may introduce a noticeable transient under-steer, and will increase the rate of heat input to the tyres.

- A digressive damping style for dampers attached to the driven axle may compromise traction, may introduce a noticeable transient over-steer, and will increase the rate of heat input to the tyres. The transient over-steer may reduce the magnitude of steering inputs (& hence may reduce turn in speed loss) on the entry to high speed corners.

- A progressive low speed characteristic at either axle will usually improve contact patch load variations for small inputs, but will reduce platform stability and will reduce the rate of heat input to the tyres (not necessarily beneficial for slicks).

- Digressive front damper compression & digressive rear damper rebound can help to reduce vehicle pitch velocity under braking (noting that driver inputs are essentially low velocity).

- Digressive front damper rebound & digressive rear damper compression can help to reduce vehicle pitch velocity under acceleration.

Apologies for the length of this post. I hope some of it may be useful.

Edited by meb58, 19 August 2010 - 13:54.

[DaveW]

Thank you for this follow up DaveW...most of this is above my head so I will have to reach a bit to put it all together.

What exactly is a mulit rig test? I've read about it here before but didn't want to drag down the other threads with that question...this seems to be an appropriate place and time to ask...

Regarding damping coefficients and style - and please keep in mind I am really nothing more than a back yard mechanic. Can I assume that coefficient = velocity and that style = the character of the damping curve? I am not sure what you mean by "Moving damping from the rebound to the compression half-cycles..." or the reverse?

A multi-post rig (in my case) comprises 4 hydraulic actuators acting under position control that can excite a vehicle through its tyres via wheel platforms. When required, two additional pneumatic actuators are attached to the sprung mass to simulate a constant down-force. During a recorded run the wheel platforms move sinusoidally with a logarithmically swept frequency and a constant peak velocity. Recordings are processed to derive a complete "equivalent" vehicle model, amongst other things. The model is used to predict the effect of damper. spring, tyre & ballast changes, & these are used to iterate to "optimal" damper settings for a range of spring, ballast & down-force selections.

This shows the output of a damper model. The green squares joined by a red spline shows an actual damper trajectory, The magenta "loop" shows the load-velocity trajectory that should be generated by a displacement-controlled damper dyno at 5 Hz if the damper had a pure internal stiffness of 6 KN/mm (positive values are compression in both cases). The blue line shows the trajectory of a pure linear damper having the same damping coefficient (5.277 N/mm/sec) over the 204 mm/sec velocity range. The actual model is compression-biased and is digressive in both compression & rebound. The characteristic might have been obtained from a symmetrical starting point by "moving" damping from rebound to compression.

I hope this helps.

[meb58]

...thought that's what it might be like...I'll have to read this a few times so I can follow...Thank you! Has anyone ever set up a multi-rig in a wind tunnel? I imagine that might prove very helpful for aero cars...

A multi-post rig (in my case) comprises 4 hydraulic actuators acting under position control that can excite a vehicle through its tyres via wheel platforms. When required, two additional pneumatic actuators are attached to the sprung mass to simulate a constant down-force. During a recorded run the wheel platforms move sinusoidally with a logarithmically swept frequency and a constant peak velocity. Recordings are processed to derive a complete "equivalent" vehicle model, amongst other things. The model is used to predict the effect of damper. spring, tyre & ballast changes, & these are used to iterate to "optimal" damper settings for a range of spring, ballast & down-force selections.

This shows the output of a damper model. The green squares joined by a red spline shows an actual damper trajectory, The magenta "loop" shows the load-velocity trajectory that should be generated by a displacement-controlled damper dyno at 5 Hz if the damper had a pure internal stiffness of 6 KN/mm (positive values are compression in both cases). The blue line shows the trajectory of a pure linear damper having the same damping coefficient (5.277 N/mm/sec) over the 204 mm/sec velocity range. The actual model is compression-biased and is digressive in both compression & rebound. The characteristic might have been obtained from a symmetrical starting point by "moving" damping from rebound to compression.

I hope this helps.

Edited by meb58, 19 August 2010 - 17:03.

[johnny yuma]

almost on topic,I have spent thousands of kilometres in the back seats of ordinary late model vehicle such as Ford Focus,Mitsubishi Magna and Outlander,Honda CRV,Toyota Corolla ,Holden Commodore, and in most the road noise right in your ear on most grades of bitumen is such that you cannot always hear what front passengers say at highway speed.The best were the Honda CRV and Holden Commodore probably because both were Station Wagons.
Question is does this noise travel up the DAMPERS,into the bodywork and into your ear,and is there a fix ? Stacking coats over the parcel shelf seems to do little.Years ago when replacing the rear dampers on a 1971 Holden sedan,a coverplate removed in the boot/trunk area to access the top nut revealed a small,apparently factory fitted,lead shield.Early NVH ?

[gruntguru]

Good observations have been made in this thread. I would like to summarize these if I may &, perhaps, add to them on occasion. . . . .

. . . . Apologies for the length of this post. I hope some of it may be useful.

A great post Dave. I am sure thet everyone who took the time to read it carefully learned a great deal - I did. A couple of questions if I may.
1. What is "platform stability"?
2. Are there any "starting point" rules for dampers for purpose-built, high-mechanical-grip racecars? For example most dampers these days are digressive and favour rebound damping, but would reducing both of these characteristics be a better starting point for such a car?

[DaveW]

A great post Dave. I am sure thet everyone who took the time to read it carefully learned a great deal - I did. A couple of questions if I may.
1. What is "platform stability"?
2. Are there any "starting point" rules for dampers for purpose-built, high-mechanical-grip racecars? For example most dampers these days are digressive and favour rebound damping, but would reducing both of these characteristics be a better starting point for such a car?

Thanks, I appreciate that, GG.

Platform stability... The best way to illustrate that, perhaps, is to consider a vehicle negotiating a single kerb. The suspension velocity will initially be positive (when the damper is being compressed) & will then become negative as the damper recovers. The sprung mass (or platform) will be accelerated upwards during the first phase, & will then recover, accelerated by gravity. If most of the damping is compressive, then the initial platform acceleration will be high, but wheel recovery during the second phase will be fast. Alternatively, if most of the damping is in rebound, initial platform acceleration will be low (good comfort, incidentally), but the wheel will retract more. Here, however, wheel recovery will be slow because the spring is pushing against a strong damper. Hence the wheel will stay off the road for longer. The platform will be disturbed more by compression biased damping, but contact patch load recovery will be slower for rebound biased damping (up to a point, anyway). For a pure mechanical vehicle, the first strategy will probably be preferred, but for an "aero" vehicle the second strategy is likely to be better. I did say "up to a point". If the kerb is big enough, then an explicit "blow off" control will help whatever the damping strategy (which, incidentally, is what Penske's latest bit of "magic" implements - rather awkwardly, in my view).

Actually, I don't really agree with your "for example" in item 2. The majority of "mechanical" vehicles I see these days (Touring cars, & GT vehicles) use mildly compression biased damping. However, they do tend to use four way, relatively symmetrical, dampers (DSSV or TTX, mainly). Both can be fitted with "blow-off" controls, however, to help with occasional large, track specific, disturbances. More "traditional" shimmed dampers (e.g. 3-way TT44) generally work best with rebound biased damping. A good compromise in that case might be to use a digressive compression & progressive rebound style. This will give compression-biased damping for low inputs, & revert to rebound-biased damping for high inputs. The disadvantages are that static ride heights must be higher to accommodate the increased compression travel, damper builds will tend to be more track-dependent, and the dampers can't easily be used to control manoeuvring response. Fat Boy (I recall) published elsewhere on this forum a simple set of "rules" for setting up that type of damper. His logic might have been suspect (I think, anyway), but it does (I recall) yield a sensible starting point. I should add, I suppose, that silly "minimum static ride height" rules can be nullified simply by adopting rebound-biased damping whatever the damper architecture. Lowering the average on-track c.g. height is often dominant.

[Tony Matthews]

Fascinating stuff, DaveW, thanks.

[munks]

... if the damper had a pure internal stiffness of 6 KN/mm ...

Trying to follow you here. Is this 'pure internal stiffness' the same as the 'hysteretic damping coefficient' or 'hysteretic damping ratio' (as described in the "Alternative models" section of this Wikipedia page: Damping)?

Thanks.

[DaveW]

Trying to follow you here. Is this 'pure internal stiffness' the same as the 'hysteretic damping coefficient' or 'hysteretic damping ratio' (as described in the "Alternative models" section of this Wikipedia page: Damping)?

Good spot - but no, I'm afraid.

All real dampers have an internal stiffness, caused by fluid compliance, damper structure compliance, trapped air, etc. The stiffness is normally different for the compression & rebound half cycles, even for through rod designs, an inconvenience that is frequently glossed over by taking an average value when modelling a damper. A reasonably representative model is a spring placed in series with a "pure" damper - rather like a "top mount" in fact. A spring & damper combination has a complex stiffness in the frequency domain. Typically, when damping is linear with coefficient C & the spring has a stiffness K,

The transfer function is (K + i*w*C) for a "pure" viscous damper

The transfer function is (K + i*C) for a "pure" hysteretic damper

The transfer function is K*(1 + i*w*T1)/(1 + i*w*T2) for a viscous damper with internal compliance

Apologies about the maths, but I can't think of a better way of explaining. For case 1, the spring dominates at low frequencies (w), whilst the damper dominates at high frequencies. The spring is normally dominant at all frequencies for case 2 (no "w" term here), whilst for case 3 the parallel spring (K) is dominant at low frequencies, but the "series" spring (the internal compliance) becomes dominant at high frequencies, i.e. it starts as a spring, migrates to a spring/damper combination at intermediate frequencies, and becomes a spring again (with little or no damping) at very high frequencies. The transfer function is actually a "lead/lag" filter for case 3 with time constants T1 & T2. I leave it to the enthusiast to define T1 & T2.

Actual values of internal compliance normally lie between 2 & 6 KN/mm, depending on the piston CSA. Values less than around 2 KN/mm will have a measurable effect on the performance of the damper. One of the perennial problems faced by a damper designer is that race car designers demand dampers that weigh nothing (hence small CSA) & then arrange geometry to have a large motion ratio. In one actual example the nett effective series stiffness was reduced from a respectable value of 3.5 KN/mm for the damper by itself to just 640 N/mm, courtesy of a poor vehicle installation stiffness & an unfriendly motion ratio of 1.6. Vehicle on-track performance was seriously deficient but I suspect that the problem was understood & resolved only because the vehicle was evaluated on a multi-post rig.

Edited by DaveW, 20 August 2010 - 16:34.

[munks]

OK, thanks, DaveW. I understand the internal stiffness now, even if I don't quite get all the math(s) yet. I'm trying to get a better understanding of damper modeling so this is good stuff.

[mariner]

Dave as the damper doctor is in may I overload you you with one more question please?

Have you ever had a car fitted with the Delphi rheological dampers ( magna ride) on your four post rig and if so did it show any notable characteristics?
I ask this because I once had cause to talk to the original Delphi engineer on magna ride on a different damping subject and he tried to explain to me how magna ride could get close to "sky hook " damping.

As I understand it this is where you define the position of the car in vertical space and then use the suspension position sensors, accelerameters and a set of custom algorithms to control the rheological fliuid so the car stays as far as possible in that point in vertical space. He did not claim that the system could actually acheive this but he suggested (IIRC) that it could get close.

It sounded to me a bit like getting 90% of the benefits of full active ride without having to use any external actuator power as with AR. It all seemed very remarkable but as he had developed magna ride I have to respect his level of knowledge but it would be nice to know if the technology had been independently tested as I know more mfrs than just Chevy now use it.

[DaveW]

Have you ever had a car fitted with the Delphi rheological dampers ( magna ride) on your four post rig and if so did it show any notable characteristics?

Yes. I currently own a vehicle fitted with Magnaride. Put it on my rig, of course. It had almost no damping, apart from rather high levels of friction. I guess wheel speed signals are required for activation.

Subjectively, the system is just OK. Loose & a little inconsistent in normal operation. Much improved with sport mode selected, but then suffers from persistent shake. I did some research & discovered that it (probably) runs a stroke-limiting algorithm whereby damping is modified as a function of damper position & velocity. I could believe that is the cause of the apparent inconsistency. I implemented something very similar for the Lotus active system. That was not inconsistent (I thought at the time), although we did play with various "shaping" algorithms before the GA's were satisfied.

Would dearly like to do a job on Magnaride.

p.s. 90% of the benefits of true active suspension? Mmm... but I admit to being biased.

Edited by DaveW, 20 August 2010 - 21:01.

[meb58]

What a wealth of information here DaveW! I hope you don't mind persistant questions from left field...

Regarding progressive and digressive damping...in your experience have you found that either suits a particualr driving style? I imagine a driver prone to over-driving his rig might over work the tires if damped progressively. I understand that there are many other potential factors here but have you seen better tire wear using digressive damping?

My question is track biased, not public road.

[DaveW]

For a conventional four damper rear drive race vehicle, I would expect the front dampers to be digressive, at least in compression, & the rear dampers to be relatively linear. Digressive front dampers will make for a quick steering response & an increased rate of heat input to the tyres. It will cost some "grip" but not as much as running with low tyre temperatures. Normally, rear tyres will tend to be heated fairly quickly anyway, so a linear rear will help mechanical "grip".

Other types of vehicle have different requirements.

[meb58]

My intuition tells me that a similar goal might benefit a front driver since the rear tires are at some level 'along' for the ride. In addition, would this type of roll rate balance front to rear help tidy up mid corner and exit?

I realize that there are many other factors that may inlfuence how a car performs DaveW but I tend to graps a concept if I read it frame by frame at first.

Edited by meb58, 24 August 2010 - 13:45.