Suspension - droop question

[AG_]

Hi, i'm new to the forum, wanted to ask a simple suspension question, which I haven't seen addressed on any of the other big suspension threads here (or my teammates).

I have read TTW and part of RVD, so I do know the basics, this is more of a FSAE specific thing that I was hoping to hear some experienced opinions. Might be totally amateur, but gotta start somewhere right?
By rules, we are required to have 1in of up, and 1in of downward travel. From what I understand you don't need much more than 2in (my team's current car has about that much) travel front and rear - rear a bit more.

So, that would mean that to meet rules the car has to sit halfway into its travel - from my limited experience tuning the 15' Ryerson car (of which I did not design the suspension, and the designer is pretty much unreachable for help), that means
a pretty soft setup (from looking at the driving, and a bit of driving myself - also installed linear pots, haven't got data yet)

So heres my question - is a half droop setup inherently too soft for our kind of tracks? (with sufficient antiroll) or is that a reasonable setup on a racecar like this?
Also, seeing teams have pretty stiff roll resistance, is there even any value in having that much spare droop, performance wise?

Thanks in advance for the responses.

Figured I should have an intro -
2nd year mech at Ryerson, suspension Co-lead along with another teammate for this year.
Car took 2 years - 1 for design, 1 for build (painfully inefficient time management). I was there for the build, and so did not even meet the guy who designed suspension -
he graduated, and didn't keep in contact. This year i'm doing a partial redesign, of which one thing is adding adjustability to get some good real world testing of different characteristics.

[Charles Kaneb]

AG_,

The driver matters quite a bit, and what she thinks (there's no minimum weight) matters to both laptimes, setup, and confidence.

Is last year's car driveable? Can you take it somewhere to test? If you can, you can find out whether it's too stiff or too soft. The driver might just flat-out tell you - if it's far too stiff, you'll get complaints about the pounding, if it's too soft, you'll get complaints about having to drive the smallest '99 Buick Century ever (sorry, BillCobb - with all of General Motors' abilities and experience in vehicle dynamics, they still made marshmallowmobiles).

If you'd rather not trust your driver, pull a zip-tie tight around the damper rod, and go for a few laps of a smooth skidpad. If it is pounded into the bumpstop, then you are soft enough in roll to be bottoming out. If you're running a soft-spring/stiff-ARB setup, this may not be the case, and you will have to set up a little oval to find that your springs are too soft to prevent you from pitching the nose into the pavement under braking.

Negative driver feedback usually does not come with good laptimes, so never let the car leave the pits without a stopwatch on it just in case you find an exception.

As for travel, you're required to HAVE 2" of kinematic travel. How much you use is up to you. If you're seeing advantages from running a soft setup right up until you bottom out or top out, then you may want to design your car to have and use more. You do not have to use all 2", either. Go to the local kart track and ask to try a TaG or shifter out, or have one of the TaG or shifter drivers drive your car - this may be especially useful if you can compare laptimes. A kart has substantially less kinematic travel than a midpack FSAE car.

Now that you've qualified what you're looking for, go quantify it. Find your center-of-gravity location, then calculate what the rolling moment about some point on the car will be at the cornering acceleration you've found in skidpad testing, and the pitching moment from the braking acceleration. Then, get a big spring scale and a ratchet strap and apply this moment, while measuring travel at both ends. If you take a few points working up and down to these peak loads, you might confirm what you'd calculated from RCVD, or learn something completely different. Either would be useful for the design of a new car.

Just remember, everything's a spring, and a 1/2" diameter "oh-two-thin" wall A-arm tube or pushrod is a soft enough one that you should probably calculate its effects!

Your job is to take this rambling and turn it into a real test plan, run it, and use its results to make your design decisions.

[AG_]

That was a fantastic answer, thanks. I will take your advice for sure and go through that systematic test procedure.
The car is drivable, just that the team was focused on getting reliability issues figured out so I couldn't get proper suspension testing- plus, I didn't have a solid enough procedure. I know how to get CG hight, am planning on doing that. But here's another question: if I wanted to get mass centroid axis, would the only accurate method be taking the car apart, weighing all components, and plugging it back to the SW assembly?

[NickFavazzo]

You don't have to use the entire 2 inches, nothing is stopping people running super stiff springs and driving it like a kart. We ran something like 100mm of bump travel on our aerobeam cars just to prove we had sufficient travel, so whatever you do make sure it is obvious that you have properly attempted to have the minimum amount of travel.

[Z]

AG_,

As others have said, the FSAE Rules dictate that you MUST have 2" of suspension travel AVAILABLE.

But it is the track itself that dictates how much suspension travel you NEED.

As I have noted at length before, many cars have won FS/FSAE with NO VISIBLE SUSPENSION TRAVEL. (These cars would have been even faster with a small amount of travel, but they still won... :)).

Most of the FS/FSAE tracks I have seen benefit from about +/- 1 cm of travel, preferably fairly soft, then with progressively stiffer bump/droop stops for another ~1 cm. Only minimal damping is needed.
~o0o~


... if I wanted to get mass centroid axis,...

Why would you want this?

(And what, exactly, is a "mass centroid axis"!!!???)

I suspect my response to the answer you give will be, "NO, no, no, no, nooooo....., that is NOT how it works!". :)

Z

[BillCobb]

if I wanted to get mass centroid axis, would the only accurate method be taking the car apart, weighing all components, and plugging it back to the SW assembly?

That's a NO from me, too. Simply because your car is not made up of point masses. Some of the non-point masses have their own inertia values. That makes your proposal 'pointless' so to speak. If you point it out, you'll likely only get 60% of the real matter. Sorta like Dark Matter as a matter of fact...

Instead, you swing the car + the dummy-driver (whom ever volunteers, actually, and with optional hyphen) as a suspended entity in a number of positions and compute the inertias (tri-filer pendulum rig). Then you fit these values to an inertia ellipsoid and presto-chango the ellipsoid's principle axes are the axes you axed about. You might be intrigued by a notion to align the roll axis of the car with the IXX axis you axed about or to reconfigure the car to make the inertia watermelon a design feature. If you are not into big mellons, a football analogy will do. You can make your driver cross by crossing up the cross products, though. Powertrain stuff is a key ingredient to this puzzle. (Transmission, accessory drive(s), manifolds), battery and fuel, wheels and tires are there, too, whatever junk is not at the CG and heavy. If your motor and transmission are soft mounted, then the debate that follows is one of inertia based mounting system vs. torque axis mounting. Do you want good ride and handling or smooth acceleration ?

[AG_]

As I have noted at length before, many cars have won FS/FSAE with NO VISIBLE SUSPENSION TRAVEL.
Z, Believe it or not, I have read quite a bit of your comments on certain suspension threads, back when I joined my team last year and also a bit a while ago. Pretty much on any suspension thread, you seem to pop up....and always with the same opinion :)
As a new member, I totally agreed - I was all for going bare bones simple. After all, a flat track car can use its tires for most of the suspension (assuming the driver can take it...). Then the new member attitude faded...and I tried to learn the "proper"
way of doing it. After driving it for the first time just recently, I am back to my original belief - we can get away with something hilariously simple, and from what I talked to the judges at michigan, they may just like it - if it is justified. But, this year it is
more complex. The team has alot to prove, university needs results or they may shut us down. Chassis is changed minimally, everything to having a running, well finished car at comp. I would want to do something, but its best to actually go through the proper process first, understand it, so I can prove to my team that it is unnecessary. Right now it is a theory, by a basically new member. So at the start of that "proper path" (also backtracking through a design that I have virtually no access to knowledge of it, so I know what to change and why), I want Centroid axis. So that I know where the roll axis (at ridehight) are in relation to the mass,
which would explain what behavior it is supposed to generate. I assume the judges would want that too.

[AG_]

BillCobb - I would have to check if the university has this kind of equipment, and read up on the math of it, but sounds like that is doable, thanks for the explanation. The engine is hard mounted, didnt think soft mount is even a consideration for fsae

[Z]

AG_,


So at the start of that "proper path" ... I want Centroid axis. So that I know where the roll axis (at ridehight) are in relation to the mass,...

And now starteth the properly steep part of the educational path you are on. But don't worry, because when you get to the top of the hill the view is just wonderful! :)

As I said before, with regards to centroid and roll axes, "NO, no, no, no, nooooo....., that is NOT how it works!"

And as Bill pointed out, you can do various tests to determine your (assumed rigid) car+driver's "...inertia ellipsoid and ... the ellipsoid's princip[al] axes are the axes you axed about...". But Bill did add "...You can make your driver cross by crossing up the cross products...".

Putting Bill's comments a little less elliptically :), your (assumed rigid) car+driver has three "principal axes" about which it can be spun in perfect balance. These three axes ALWAYS pass through the CG (obvious, really), and are ALWAYS mutually perpendicular (astonishing, really!). Of course, the inertia ellipsoid also has an infinite number of other axes passing through its centre (ie. the CG), but spinning the body about any of these other axes makes it wobble. Oh yes..., those cross-products!
~o0o~

So, what has the above got to do with Vehicle-Dynamics? Well, everything ... and nothing.

Everything, because the above principal-axes stuff is straightforward "Eulerian Rigid Body Dynamics", which is a sub-field of "Classical Mechanics", itself a sub-field of "Applied Mathematics", etc. As such it is currently the best way of figuring out how stuff, like a racecar on a racetrack, works. Note that you can use the above to make successively more and more accurate models of "reality", until eventually the correspondence between your model and the real reality is less than the tolerence of your best measuring instruments (well, providing you don't go down to the quantum level). You can start this more accurate process by modelling the car+driver as multiple idealised rigid bodies connected with idealised springs... (Edit: but at this point in your path, you should just stick to a single rigid car+driver.)

But, sadly, the above "Eulerian RBD" has nothing to do with the many popular flavours of VD that are heavily promoted nowadays. This is because these "cottage industry" VDs choose to ignore Classical Mechanics and all the good results that have come from it, and instead they dream up their own magical "Voodoo Dynamics" (TM applied for).

So they invent impressive sounding things like "mass centroid axes", which presumably can point in any direction you want? If not, then how do you measure their "proper" direction?

And then there are those magical "roll" and "pitch" axes, which again come completely UNDEFINED. If not, then what is an unambiguous way of finding them?

And what, if any, is the connection between the mass-centroid-axis(es?), and the roll/pitch axes?

And will Claude ever give a clear, rational, justification for his insistence on the use of the "parallel-axes" theorem when calculating damping-ratios and the like?
~o0o~

Ooops, I seem to be ranting again...

Claude, can you please put AG_ back on the "proper path"?

Z

(PS. This "steep" part of your educational climb involves you finding out that your "teachers" have many different opinions on how stuff works, and it is up to you to figure out how to choose between said many contradictory bits of advice!)

[Tim.Wright]

The (incorrect) use of the "mass centroid axis" in vehicle dynamics was popularised in Carroll Smith's "Tune To Win" which was subsequently used as a bible for many vehicle engineers so the idea spread like wildfire. I've heard that Mr Smith corrected himself on this subsequently but the book was not updated so the myth propogated.

In short - forget the mass centroid axis and anything written about it.


PS. This "steep" part of your educational climb involves you finding out that your "teachers" have many different opinions on how stuff works, and it is up to you to figure out how to choose between said many contradictory bits of advice!
This process took me about 4 years - but then again I'm stupid. Hopefully it will be quicker for you.

[BillCobb]

I would like an explanation of what the comment about incorrect use of the "mass centroid axis" is all about, the before and the after.

Never mind, I found it.

His theory, was that TWO CG locations were in play during vehicle dynamic maneuvers, one at the front suspension reference point and one at the rear. Maybe in the days when cars were made out of jellow and connected somewhere with a soda straw. SO, cars are not all about dumbells, but the reverse is sadly true (in many cases).

[AG_]

forget the mass centroid axis and anything written about it.

Wow, thanks. Had no idea. TTW really had a few weird parts, especially in how antiroll was explained. Never thought too much about centroid axis though...


SO, cars are not all about dumbells, but the reverse is sadly true (in many cases).

Ive got to admit, a bit confused about what you mean by that. Are you saying cars were less torsionally rigid back then? and that now they are rigid to a point that it is basically one solid sprung mass?
If so, I guess the 2 mass center points at each axle I guess wouldnt make that much sense, but im not quite sold on your explaination just yet

[AG_]

Z, I'm a bit confused as to what exactly is the "inertia ellipsoid". I'm only started second year so some of the concepts Bill and you mentioned are a bit new to me.
Also, what is really the difference between the "mass centroid axis" and the inertia ellipsoid/ axies of rotation that you mentioned? This may be a meaningless question, coming from my very primitive knowledge in the subject of inertia.
As for roll and pitch axies - I've noticed that, how the roll axis is not constant throughout the motion. I figured best I can do is make the front change hight in the same way as the rear, but just start from a different position depending on the location of the supposed mass centroid axis - which I am still not fully grasping the implications of what you are explaining about it being irrelevant. I should probably say, thanks for taking the time to explain it. If I am asking questions about too simple physics, feel free to say "go read book xxxx and come back with some real knowledge" - I would accept that answer, and gladly go read. Although, reading it in this form seems much more interesting

[AG_]

Also, what is the parallel axis theorem? I've actually listened to Claude in person at Michigan, very interesting, but very much by the books.

[Charles Kaneb]

The moment of inertia measured by an axis through the CG is the lowest moment of inertia that can be measured about any parallel axis. To consider the inertia of an object measured on an axis away from its center of mass, you add the inertia of a point mass that distance from its center of mass.

It's a basic concept of point-mass models.

http://hyperphysics.phy-astr.gsu.edu/hbase/parax.html

[AG_]

Z, I was speaking to an alumni about it in the past, I've come to a conclusion that this supposed distance between the roll center and the mass centroid point at each end would mainly affect corner entry behaviour. But, the more I think of it now, the less that conclusion seems valid. Weight usually transfers diagonally, and apart from the roll moment there are so many things different for front and rear in terms of actual conditions, that even if I allign the roll centers with the mass centroid, it would probably make little to no visible difference. Especially without a fully tuned chassis, and a world class driver. But, that's based on my traditional knowledge...what you are saying is that this whole thing is invalid, but for other reasons

[AG_]

Charles - That sounds pretty straightforward, I'll read up on it though!

[onemaniac]

AG_, I got two things for you.

First, drop your pen, paper or CAD station, forget about the math and work on establishing 'proper' protocol for documenting design and communication.
If your team cannot get hold of someone who designed last year's car and even worse if you cannot find the documentation to allow you understand the design, your team will repeat the same mistake yet again. (I'd understand if you can't reach the guy who designed your 2005 car, not 2015 car)
Someone WILL drill wrong size holes, someone WILL install something backwards, someone WILL neglect the rule book.

Second, doing a 'proper' design does not mean complicating things. Simple design does not mean it isn't 'proper'.
It's funny how you mentioned 'adjustability' to get the real-world testing done for different characteristics.
How many parameters are you looking to adjust? at what increments?
Let's say your suspension design is heavily limited and you are only allowed to change the track widths (front and rear), spring stiffness (front and rear), damping (front and rear), tire pressure(front and rear), and front toe angle.
That's already 9 parameters. To make it even simpler, let's say you can only choose 2 levels - high and low - for each parameter.
Now you have 2^9 = 512 setup variations. How would you keep track of these changes? and How many more do you think you can handle? More importantly, how would you process your data?
In my opinion, the 'proper' way of designing a suspension is to get out there to test your current car and focus on one or two parameters at a time, and gather consistent data. You can THEN go backwards and figure out what does what and why.
I think too many students focus too much on Research and neglect Development. They should go together.

Just my $0.02

[AG_]

Onemaniac, I'm fully with you on the communication issue. I've mentioned the time management issue in my initial post, but the really knowledge transfer was absolutely terrible. Sure I have the SW of the design, but unless I understand the reasoning for it (which is hard by just backtracking), I would like you said make the same mistakes over again - or just different ones too. I've started a onenote design notebook to track my work, super convenient so far. Rest of the team? I'd have to constantly remind them probably haha. I will also try to reach the designer. For adjustability, of coarse there are the standard things, but I was thinking more in depth- leverage ratio (or motion ratio, if you prefer that name) behaviour - a linear setup as a basis, with different gradients of progression as adjustment holes. Is it necesary? I don't know, perhaps it is all so minor of a difference that it won't be noticed. I guess I just want to find out on track. Testing the current car - for sure, haven't done as much as I'd want to. This is why the redesign for this year is a minor one, in order to get that testing I need for a properly redesigned suspension.

[Z]

AG_,

Two very SANE points made by Onemaniac! They are so good that I will restress them here.
~o0o~

1. The very proper path for an FSAE Team is good "Knowledge Transfer". In fact, it is your School that should be setting up the right conditions for such, and then ensuring that there is lots of it going on. After all, that is their job!

The main repository of KT are the theses of all the previous Design-Leaders (assuming they did their FSAE work as a thesis/major-project). These theses should be compulsory reading for future Design-Leads. At the very least, a quick read of these theses' abstracts/intros/conclusions should give the following year's students a good idea of what the thesis writer thought most important.

Similarly, each Team(or Section)-Leader should write some brief notes (a few pages at most) stressing what they considered most important to get right, and what "fashionable must-haves" they found unnecessary. For example, a Suspension-Lead might write "MOST important is toe-stiffness. It (edit: Toe-Compliance) MUST be below (whatever...). On the other hand, motion-ratio is NOT very important, and can be anything between ...".
~o0o~

2. Don't take this the wrong way AG_, but your team is currently taking baby steps. If you run out onto the stage (ie. attend your next comp) and try to do a triple back-flip with half-twist, you are all but guaranteed to fall flat on your face. Even the experienced teams struggle to get their simple tumble-rolls right.

So a highly adjustable suspension might be something you play with away from competition, when, and if, you have lots of free time. But your comp-car should be fairly simple. And stiff, toe-wise! Otherwise you just get lost in the never-ending permutations and combinations of the adjustability...

Similarly, at this stage it pays to keep your theoretical analysis fairly simple too, least you get lost in the never-ending mumbo-jumbo of Voodoo Dynamics.

Given your apparent understanding of the subject so far, I suggest you try doing an accelerating(or braking) + cornering weight-transfer calculation of a car with four-wheels/suspensions, but only a POINT mass at the CG. Say, wheelbase = 1.6 m, F&R tracks = 1.2 m, with CG = 0.3 m high and at some given F:R%. Try to include the effects of "anti-roll" and "anti-pitch" (*) from the suspension geometries, in addition to the weight-transfer that comes from spring stiffnesses. (* These also known as "RC/PC heights", or "lateral and longitudinal force-line-slopes", or "n-line slopes", and covered briefly in this Jacking Force (http://www.fsae.com/forums/showthread.php?4063-Jacking-force) thread.)

Much, much later, you can move on to Eulerian Dynamics with its continuum-mass-distributions, principal-axes, MoIs, gyroscopic-couples, and so on...
~o0o~

Finally, as easy weekend reading, here is a 2004 thread about Vehicle Dynamics (http://www.fsae.com/forums/showthread.php?6542-Vehicle-Dynamics). This thread is easier reading than the Jacking Force thread, and it covers "front and rear CGs". You can think of it a bit like a soap-opera on the state of the art in modern VD. Yes, it was 2004, but nothing much has changed...

Z

[AG_]

Got sucked into the semester, didnt get around to looking at it till now. Thanks all you guys for the advice, really couldnt ask for any more (although I probably will, eventually ;) )
Z, I will take your advice, start simple and work my way up from there - and perhaps from the top I will see that an even simpler still solution is better. Definitely wont be making big changes this year, mostly try to learn and validate as much as I can - mess with tire data, have a yaw sensor, and see if the car is taking advantage of its tires properly.
I've learned an interesting bit about the wheel size choice from a past member, but it is fairly unrelated to this and will probably start a big discussion, so I will post it separately.