peacekpr

So I am thinking about embarking on trying to simulate/model the vehicle dynamics of the 5th Gen Honda Prelude in my quest to find the perfect suspension setup in terms of spring rates and damping. After a quick Google search, I see there are various tools available to help me attempt this. (Since I have used Matlab in the past for school and work, I may be partial to using it.)

Has anyone out there in the Honda Tech community attempted this? I realize this may have been done and that various companies out there sell suspension components based on certain track/road conditions, but the skeptic and engineer in me wants to find out the answers for myself rather than believing aftermarket parts companies.

Some information I would need are exact static forces at each wheel, the factory spring rates and damping coefficients for front and rear, moments and products of inertia in the various directions, center of gravity, and probably other parameters. My ultimate goal would be to get a working model that I can summarize in a white paper that tabulates the main "tweakable" suspension parameters based on input conditions that the driver demands.

Any help or ideas would be much appreciated. Any preferred software packages? Is the OEM technical information listed above readily available?

Thanks,
Jason

RR98ITR

Oh Good Luck in your worthy endeavor. You know...it sounds like you're more looking to solve a problem than to wind up with a good working setup to drive. Why not take up instead something like Fermat's last theorem or the Poincare conjecture? "Yes, I know that."

Scott, who is offering better advice than at first appears...

peacekpr

I won't deny I'm looking for a problem to solve, and this is as good as any (since a Russian beat me to a problem you mentioned). With a good solution to this problem, I (and maybe even you) can be better informed on what components I need (and what settings to dial in) based on the driving I'll be doing.

Worthy information that could help me would be nice if you have any to offer.

RR98ITR

Ah, but you see I specified those projects precisely because they were known to have solutions, whereas the one you're intent on doesn't. The black box calculator that you can realistically construct can be loaded only with some basic physics of which only some is intuitive and operation of which requires iterations of faith and experimentation. And if you think you can build a model and from it produce the "perfect" solution, you won't. You won't. This honestly isn't intended to be discouraging, merely to suggest your first step be in a different direction and be propelled by a different philosophy.

I know the DC2 chassis. I don't know who knows the 5th Gen Prelude. But you can gather most of the data you want for yourself. And you probably should so that you can trust it. And you sound like you want to do your own homework, so just do it and apply and assimilate from what similar has been done by others. But it takes quite a driver to do any serious validation. And quite a budget. Have you got those covered?

Scott, who has considerable experience along the path from ambitious beginnings to ambiguous indetermination...my love of driving sustains the efforts that impede it...you have been warned...

beanbag

Yes, there is Matlab simulink for all your simulation needs.
At some point, you need to decide which things you want to really model, and which things you would rather measure. For example in the case of shocks alone, forces can be measured with load cells, velocities measured with LVT's, etc. This is easier than trying to model the tire as some kind of non-linear spring, which manages to transmit road forces thru the shock and then to the chassis, etc.

IOW, you will need to build measuring equipment.

More importantly, you need to have a more precise and well defined question or problem to solve.

If you want to sit around and theorize and simulate, then read some vehicle dynamics books and look up various people's masters theses on this. They have all that natural frequency, quarter car model, masses connected by springs, transmissibility plots (which are not valid for non-linear damping anyway), etc etc etc. They are a dime a dozen. DescartesFool has posted some links here on some of them.

Doesn't some university in Florida have some FSAE or whatever team that builds race cars? I think those are the people you want to hang out with, not the losers here.

slick1851

June racing prelude gets taken by there DC2 around the track and they have more RD into that chasie than most tuners. Mfactory sponsered a red prelude it's on there site ans they have a lower final drive for the dreaded Hseries trans.

peacekpr

All of this is a moot point, of course, if I can't find the physical characteristics... that may be discouraging enough. We shall see. This whole endeavor may require more time and effort than it's worth to me. But if I can pull it off, maybe it's useful. Maybe not. And I for sure don't have the budget to "validate" - except for the first and only attempt if I purchase what I perceive to be the "perfect" setup. But it may be interesting to see how close the answers get (if I ever get them) to the established numbers (whatever they are) one might find with a Skunk2 setup or Eibach or any other name brand setup recommended by NOPI.com.

If I make any progress, I'll attempt to keep this thread alive. And I welcome any help or info from anyone out there in the Honda-Tech community.

RR98ITR

Jason,

If I might then suggest...nail down the mission of this vehicle of yours, and nail down the budget...then call a reputable front wheel drive race shop with experience with Honda chassis' and ask them if they can sell you a solution at your price point. Don't waste their time - buy it from them. Then, with all the time you've saved, drive the car and continue your education in vehicle dynamics and enjoy them both concurrently. And since you said "road" - BE SAFE and don't be stupid - your time on the track should innoculate you from being a danger to others on public roads so see that it does.

Best of luck to you.

Scott, who recommends King Motorsports to people who need this kind of help...

beanbag

ok, then let me put it this way.

Read Dennis Grant's site, Far North Racing.

Choose front and rear sprung 1/2 car natural frequencies to match terrain and balance.
65-75% low speed rebound damping on the shocks.
Approx 1/2 to 1/3 compression damping, depends on terrain

That's 90% of your perfect setup right there. Any more will require significant math/modeling, actual datalogging and custom shocks.

spAdam

iTrader: (1)

UCF used to have FSAE, but they only do the SAE Baja now. I'm actually going back in the fall, I'll probably head over to the shop and see if anything exciting is happening, but last time I was there the program was kind of dead in the water.

http://chassissim.com

I talked to these guys last year at PRI, the software is really cool and appears to cover all of your bases to the point where trying to do it all in Matlab is a waste of time. Tires are going to be your biggest hurdle. The only real reprieve is that there's no need to model them at any point but threshold.

uniseriate

So much good information there. Here's where he walks through the modeling process.

RR98ITR

DG's stuff on fwd is hopelessly out of date, advocating the okayness of inside rear lifting which we now know to cause uncontrollable irremediable and inconceiveable understeer.

Scott, whose favorite chassis tuning guru is Woody Allen....

descartesfool

If you want to spend some time to teach yourself something on suspension geometry, try Wingeo3, like Scott uses, but it won't do the Honda rear geometry. Of course Scott would argue that with just one tire on the ground, who needs a suspension there anyway, let alone an analysis.

Chassissim looks interesting, but big bucks.

The best bang for your buck is Susprog3D, which is like Wingeo3 but does more types of suspensions: http://www.susprog.com/susptype.htm
At $258, it is the lowest cost software I know of that calculates virtually everything related to geometry.

I got him to add the Integra Type R geometry many years back.

Getting the data is the hard part, as is figuring out the coordinate references required to input the data into the program.

Check this out, as you may find it interesting: http://www.susprog.com/tutorial/tuto.../HerbAdams.htm

RR98ITR

OH THAT"S JUST RIDICULOUS!

Scott, who..."incessant, interminable and incontinent understeer"...

suprmods

You could do the models in simulink like suggested, but as pointed out validation will be your main problem.

You can start small with quarter-car, then go half-car, then full-car (2 DOF, 4 DOF, 7+DOF) simulink models, but the more complex you get the more input info you need. Measuring vehicle cog height and location, unsprung mass per corner, wheel base, track, spring rate, damping coefficient is easy. But you'll start to need tire stiffness, tire damping, vehicle inertia, and a way to model the non-linearity of the damper, as your models add more dof's.

You can provide a swept sine input and evaluate performance in various ways, but you'll end up finding some limitations to these models. Validation of these models is usally done on the four/seven post rig ($$$).

Chassis-sim is a decent package, but it only comes with a few pre-set vehicles so you'll have to build the prelude. If you have a local track you visit you can build the track in chassis-sim and simulate the vehicle around the track, but validation will require you to have run this track with sufficient data logging to match up all the data. Racecar Engineering had a good article in the past few months by Mr. Nowlan about tire modeling in the program.

We used ADAMs/View to build a vehicle model (formula student based) that raced around a track layout and were able to make changes to improve performance.

http://www.youtube.com/watch?v=cEdO0gF9QJY

Here you'd basically build the prelude as a ball representing the chassis/cog and suspension links connected to your uprights and suspension. The program has its own tire model feature where with clever programming, you can input your Pacejeka coefficients to get the tire to match whatever tire your using. Otherwise you can use cylinders to represent tires, but must program in all the calculations required to derive slip angle, then come up with methods to measure vertical load, camber, and theoretical longitudinal slip to get a decent tire model. Here again, gotta validate the tire model and vehicle performance to ensure the model is realistic.

I've had my grand visions of using a combination of all of these for my H4 car (just for fun), but know that finding sufficient tire data is going to be near impossible. Secondly getting the data from a track run will be tough since I never ran a data logger. I suppose I could get a model in chassis sim or adams that laps "buttonwillow" in a realistic lap-time window, but would still have no way to know if the models are doing everything correctly.

Don't give up just because validation will be difficult. Its lots of fun and quite the learning exercise for an inquisitive engineer.

descartesfool

Chassissim might be nice, but I can't imagine anyone actually buying it considering the price. If you are in college and can use software for free, that is different. There have been many interesting articles in RCE by the author.

Here is a slide show presentation by the author:
http://www.vehicledynamics-expo.com/...y_3/nowlan.pdf
and another
http://www.ukintpress-conferences.co...eam_nowlan.pdf

Bosch LapSim is free for laptime simulation, and doesn't need any Pacejeka coefficients. It has a built-in basic tire model in the free version which seems to work quite well for me.

You can also run a Race Technology data logger and input your data into LapSim and compare your data to the simulation and adjust simulator parameters to match real data. And it's free.

As for 7 post rigs to validate, are you serious???? How about a wind tunnel then?

My solution is:

Race Technology DL-1: $859.00 http://www.race-technology.com/dl1_8_936.html
Bosch LapSim: $0
Susprog 3D: $258
Total: $1117

No programming required. Just learn to use 3 software programs. One does data logging and lap timing, one does suspension geometry and spring, roll bar analysis, and one does laptime simulation with all vehicle parameters available for view. DL-1 can be used for much of the data validation. Not the perfect package, but software just costs less than $300 (2 tires?)

And as for getting accurate data of the suspension pick-up points on a car with a roof, it's not easy. I'd like to hear from anyone who has actually measured all the points required front and back and accurate to about 1mm which you need for good toe calculations on a tin-top and tell me it was easy.

Center of gravity? Ever tried to measure that on a tin-top and think it is easy? Just lift the rear of a car 20" off the rear scales and see what that 2500 lbs wants to do to the front scales. You have to try it.

suprmods

If full vehicle simulation were easy everyone would be doing it, no?

Ok, maybe cog height isn't THAT easy, but there are still methods to do it. Never said finding suspension pickup points was easy, and I too would be curious to hear from anyone who managed to measure their own. The OP mentioned inertia...isn't that tougher than any of the rest to find on your own? Having to actually swing the car?

Bosch LapSim tire model - so how do you verify if the model you are using represents the tires that are on the vehicle? Certainly you can still do simulations and setup sweeps to see the effect a certain parameter change has on laptime. Just filled out an inquiry form, so I'll take a look when I get ahold of it.

Your setup sounds like a good way to go if one can afford the tools.

If I were Jason I'd take my previous knowledge of Matlab (if sufficient) and work with that for now (its free). Through just equations of motion he could build 2,4,7 dof models, and play around with them a bit. Quarter car model is quite simple, only needs sprung mass, unsprung mass, spring stiffness, tire stiffness, and damping coefficient. Simulink help has a 2-dof model included. I've got up to a 4dof so far.

Here's a paper with the equations of motion for 7dof.
Damper Tuning with the use of a Seven Post Shaker Rig

descartesfool

Some pics of a setup to measure suspension points. Start with a reference grid, take suspension apart, make some jigs, and with plumb bobs, rulers, calipers and anything else you can think of, make some measurements. Have some beers around. It's going to take a while.

suprmods

yeah thats kind of how i imagined it.

Mr Tickle

Anymore photos as I'm interested to see how you get on