To learn and share about dampers / shock absorbers

[WilO]

I checked back on the topic, and sent a message to a couple of the guys involved with it (the topic ended without any clear resolution whatsoever), hopefully I'll hear something soon.

I do know that the method suggested by Mssr. Seger uses vertical acceleration of the hub, tire spring rate (this method assumes the tire behaves as a perfect spring), wheel mass, 1/4 body mass.
Ultimately, I believe the contact patch acceleration, hub and body accelerations are plotted in the frequency domain.
Sorry for the quick and garbled reply, just a little busy here.....more soon.

Wil

[DaveW]

A "generic" suspension set-up can be arrived at without track data or a rig test. All that is required is a representative model of the vehicle, routine to carry out a virtual rig test on the model, and a "cost function" - a measure that allows set-ups to be assessed numerically, one that will (on average) indicate "best" performance. A "generic" set-up can usually be tuned for a particular track by responding to driver comments during testing. I am sure there are ways of processing track recordings that can help, damper histograms for example.

As it happens, I do compile a simple model of a vehicle from rig test measurements, and use this to predict optimal settings. I always check the optimum, however, using a kind of "steepest descent" strategy to converge on the true optimum, because a simple model is only an approximation to a real vehicle.

I also have the ability to put together a more complex model of a vehicle. I use this as a matter of course to test analysis routines and as a learning tool to explore different possibilities. I could also use it to set-up a vehicle suspension without a rig test. I have pushed to do that on occasions, but I don't like it as an option because whenever I follow it up with a real rig test my solution was always deficient, because of errors in the data I was given (e.g. motion ratios), and unknowns in the vehicle (friction, secondary springs, "installation" stiffness, etc.).

It is a fact that a major reason for carrying out a rig test is to discover what is unknown about a vehicle. That step might be tedious, but I don't believe it can be omitted. It is, by the way, not a particularly costly exercise.

[WilO]

Dave, thank you for the insight into your processes, always fascinates me.
I had a look at the relevent section of Segers book, and it is slightly confusing (it's not just me, a good friend of mine who is a particle physicist and a pretty bright guy echoed my assesment). It does describe using data from an instrumented vehicle to determine CPL variation, but appears to do so pretty roughly. If memory serves, it uses the average of wheel position and load as the suspension input. Force at the contact patch is found by Fcp = Fsuspension + Mwheel (Gb + Asusp), where Gb is the body vertical acceleration, and Asusp is the suspension acceleration.
Perhaps others with familiarity with the text or the method could chime in here to help clarify...But one can certainly see the difference between an approach such as this, with its multitude of simplifying assumptions (tire is a perfect spring, inputs to the suspension are not well-quantified or controlled), and the method used by DaveW on his rig.
Dave's point about finding things that are previously unknown is a very good one, as a simplified model won't tell you much you didn't know.
I have told many local club racers to investigate a rig test, as there is a facility in Toronto which is only a few hours away. A day spent there would pay big dividends, I suspect.

[DaveW]

Force at the contact patch is found by Fcp = Fsuspension + Mwheel (Gb + Asusp), where Gb is the body vertical acceleration, and Asusp is the suspension acceleration.

I can't help thinking there is something awry with that relationship. If Newton was correct, then the vertical component of force acting on the hub is equal to the mass of the upright * the vertical component of acceleration of the mass. Hence, in the above notation, Fcp = Fsuspension + Mwheel*Asusp (with the appropriate sign convention).

By the way, the alignment of the sensitive axis of the accelerometer is critical for track data....

[DaveW]

It is probably worth recording that the equation presented by WilO is a quotation from Jorge Segers book 'Analysis Techniques for Racecar Data Acquisition'. It turns out, I think, that:

Fsuspension is the total vertical load of the suspension acting on the upright (must be calibrated as such).

Gb is the vertical component of acceleration of the body at the (possibly virtual) point of the sprung mass above the upright

Asusp is the second rate of change in vertical displacement with respect to time of the upright referred to the body (damper displacement calibrated in "wheel" units). It is probably misleading the refer to it as acceleration, given the context of the book.

Elsewhere, he uses a term Ktyre. This should be the taken as the vertical stiffness of the tyre & wheel....(I think). Calculations using his equation will be affected by the ability of tyres to average out the road structure. This is not necessarily a bad thing (probably good actually), just different from the way a tyre responds to a flat surface input.

[WilO]

Thanks for the clarification Dave, much clearer to me now. You noted that Fsuspension is the total vertical load of the suspension acting on the upright, and must be calibrated to give that value. I'm wondering what you meant by that, specifically does this involve using accelerometers to correct the load value, or is it a matter of careful placement and calibration of the load cell.
I suppose the accelerometer signal could be used to correct for the inertial forces caused by accelerating the unsprung mass, though I haven't convinced myself that this is necessary, or even correct.

Wil

[Jersey Tom]

I have told many local club racers to investigate a rig test, as there is a facility in Toronto which is only a few hours away. A day spent there would pay big dividends, I suspect.

Or a day spent considering the fundamentals and building on things one step at a time may pay bigger dividends. Need to have some constraint knowledge, or if you just throw a car at a rig I'm sure you'll see that just dropping spring rate gives you mechanical grip (no brainer!)... or in general the setup choices will make for an under-constrained study without clear direction. Just put up a post on my (new) blog about constraint-based design packages, actually, because of this thread.

Coming from admittedly a much different professional background than Dave, in my mind you have to get your steady-state or flat road handling parameters nailed down first or you will be junk... and then good mechanical grip tuning is the icing on the cake that differentiates the best performers. Not to say that I'm necessarily right or wrong on this, but the more fundamental point is that you have to make a list of what you want to work with to find speed. Are you going to chase dampers? Springs? Bars? Tires? Probably won't have time to do it all, so choose wisely!

Or.. you go to F1 where your aero department gains probably blow away anything you can do mechanically!

[marcush.]

I tend to agree with JT on this .
When working with lesser drivers you find huge chunks of time simply by providing a readable product to the driver .Confidence is all ,so when the car does not bite you it is inspiring to go closer to the limits of adhesion .Simple fact.
When the car does things you are unable to react to you will not push as you want to survive.
That´s one reason why you don´t want a pro setup the car when mere mortals have to drive it later on.

The tricky bit is really to understand what is holding you back in terms of lap time .You may think more power will ,or braking later ,but a thorrow discussion with the driver about what he really needs to be able to go deeper into a corner or floor the peadl on exit should reveal the areas where the gains can be made.

A modern damper is of course a strong tool ,as you can influence the timing of events happening -and this dramatically changes the behaviour -response of the car ...from frightimng to inspiring and the other ways round...

So if you know what to look for a session on a rig will be a valuable input into your decision process on car setup.

[hardingfv32]

or if you just throw a car at a rig I'm sure you'll see that just dropping spring rate gives you mechanical grip (no brainer!)...

Can you expand on how the post rig/tester makes the evaluation that 'dropping spring rate gives you [more] mechanical grip'?

Can you post a link to your blog?

Brian

[DaveW]

I too can understand where JT is coming from. I am certainly not happy about attempting to set up a car without some (previous) track testing, for various reasons. I would, however, question his comment:

...or if you just throw a car at a rig I'm sure you'll see that just dropping spring rate gives you mechanical grip (no brainer!)...

That dependents very much on the starting point. It is not uncommon to find that a track based set-up is too stiff for good mechanical grip. In that case his statement would be correct. On the other hand, I regularly encounter set-ups on a non-aero car that, for various reasons, are too softly sprung. For that reason I don't agree that the statement is a "no-brainer"

... in my mind you have to get your steady-state or flat road handling parameters nailed down first or you will be junk... and then good mechanical grip tuning is the icing on the cake that differentiates the best performers.

I like the logic, I would only add that you might be surprised how many discussions I have had with professional engineers (from F1 down) about the wheel rates of their springs and bars. The logic surely falls down without accurate estimates of those.

You might agree that "optimizing" damper settings for a spring and tyre selection is not completely straightforward

Or.. you go to F1 where your aero department gains probably blow away anything you can do mechanically!

Without question, aero dominates F1 performance. Arguably, however, is it not the only thing that affects performance - I would guess that Jenson Button might agree (you might be able to suggest other examples).

[hardingfv32]

Can you evaluate 'mechanical grip' on a post test rig?

Brian

[Jersey Tom]

The damper tuning bit is certainly not straightforward, I agree. That's certainly an aspect of things I don't have much working feel for. Trying to get a little more knowledge there. The series I work in has its own unique damper requirements... as with every other damn thing on the cars.

Brian - you can't measure it directly.. but you can evaluate it in a way, yes.

[DaveW]

The series I work in has its own unique damper requirements... as with every other damn thing on the cars.

You have my sympathy. If I was being unkind, I would say that the cars appear have developed around a tradition that is a maze of fix on fix, & all controlled by a magician called the "crew chief". I believe that the "COT" was an attempt to unravel some of those complexities, but I'm not sure it was entirely successful in that respect.

Can you evaluate 'mechanical grip' on a post test rig?

I seems like a good idea to minimize the time required to dissipate disturbance energy of the rigid body modes, and to achieve that whilst minimizing the transient reduction in vertical loads under the wheels.

It should also be acknowledged that slick tyres have to be "worked" to reach & maintain "chemical" grip. This is something that appears to be an issue with the current F1 tyres. The conundrum is not helped by nonsensical rules that forbid suspension setting changes between qualifying & race, & cockpit adjustable bars.

[hardingfv32]

It should also be acknowledged that slick tyres have to be "worked" to reach & maintain "chemical" grip.

Is one of the tools to 'work' the tyres the use of excessive changes of vertical loads? Does controlling the vertical loads then become a compromise?

Brian

[DaveW]

Is one of the tools to 'work' the tyres the use of excessive changes of vertical loads? Does controlling the vertical loads then become a compromise?

Not sure that I would put it quite like that. It is true that increased spring & bar rates will tend to work tyres harder, as will increased damper digression and the use of inerters. It is also true that those changes will affect the minimum loads under the tyres. The effect is not linear, however. Typically, a race car suspension will be set to optimize grip & tyre life. So for single lap qualifying damper digression and bars will both be increased to bring the tyres to "life" quickly, and then softened for a race to increase tyre life. The current F1 rules complicate set-up decisions, I'm sure, and will lead to some teams deciding not to try for grid position, preferring to reduce the number of tyre stops.