SPRING FREQUENCY 5

[willeng]

We have come across some suspension guys who bounce cars up & down for 1 min,shocks,sway-bars etc disconnected to get a spring frequencey figure. They say they are getting cycles per min of different rate springs. They are using this as a basis for setting up there suspensions.
Is there someone who can shed some light on this subject as to me the more i think about it the more it seems like BS.

 

[Metalguy]

Sounds like BS to me too. Spring harmonics matter for things like high speed valve springs, not susp.

 

[JakubMech]

Guys,

what theyre testing is chassis natural frequency, its an important part in the whole vehicle dynamics area. Check in your vibrations courses w=(k/m)^(1/2), its a measure of the stiffness of the suspension.

The higher the number of cycles per minute or Hz the stiffer the system. The softer the system the more grip you can get, the higher the system the less grip (generally). As always theres comprimises, less grip the higher frequency the more body roll control, higher anti roll moment at that end of the car - front / rear.

Its just a measure of the stiffness of the system per unit mass of chassis.

Jakub

 

[Warpspeed]

I believe relative front and rear rates of bounce can also have quite an effect on vehicle pitching motion.

The front wheels hit a bump/hole before the rears, the time delay obviously depending on wheelbase and road speed.

If the spring frequency at the rear is made suitably higher, the rear will catch up with the front and the vehicle tends to rise and fall on its springs, rather than pitch violently back and forth.

In a softly sprung luxury limo barge that is underdamped (for comfort), the overall rise and fall is much less disturbing to the occupants than a pitching motion.

That is what I have read about this, for what it is worth.

 

[JakubMech]

I also forgot to mention that spring frequency is what the ride is measured in. Passanger cars are usually within 1 to 2 Hz for comfort, below this I beleive there is the possibility of sea sickness, and above its to harsh.

Warpspeed correctly points out that usually the rear is of a higher frequency to avoid pitching issues. I beleive the figure is roughly 1 : 1.3 for an optimum setting.

Jakub

 

[exige]

Hi guys,

I'm from holland, and i've done a project on ambulances. We wanted to improve the roadholding capacity of the car without too much loss of comfort.

We also did the spring frequency test, get the car in bounce ( we call it "cadans" i don't know if its the same in english). For the best comfort the Hz frequency is about 1,1 hz. A normal van is about 1,6 hz

So it isn't all bs...

 

[Warpspeed]

Exige, that sounds about right. I believe expensive luxury cars aim for about 1Hz and and production factory sporty cars about 2Hz, as JakubMech has already said.

Anything lower feels floaty and unstable, anything higher is rather too firm for comfort. Bouncing the vehicle without the shocks, is a good way to find out what you have, because it takes everything into account and gives you a definite figure to work from at each end of the vehicle.

 

[NormPeterson]

I seem to recall reading somewhere that the low 1-Hz range is the most comfortable because it closely mimics the average normal walking pace; hence the human body does not translate disturbances in that rather narrow frequency band as being annoyances.

Actually measuring the frequencies certainly has merit, particularly for vehicles whose springs are not too heavily damped by the shocks. Although the basic math for matching front and rear ride frequencies for a given vehicle to a target speed isn't all that difficult, it is altogether too easy to assume that all of the basic inputs are constants.

Norm

 

[GregLocock]

Well, I'd always prefer to see an experiment AND a calculation, rather than the calculation alone.

If you really want to know what is going on even in a simple case like this then the experiment is vital, because you have springs in series (eg tire wall stiffness), and you have springs in parallel (parasitic springing from, for example, rubber bushes) to the road springs.

If you actually want to achieve the flat ride criterion, you need to select a cruising speed, and know the wheelbase of the car. This gives you the time delay front to back. Then build a pitch and bounce model of the car, and fiddle with the rear spring rate until you minimise the pitch after a step input to the two wheels with the right time delay. Obviously you can't do much with the initial ramp up, but the trick is to get the front and back synchronised for the next couple of cycles, which is why the back has to be stiffer than the front, so it can catch up.





Cheers

Greg Locock

 

[NormPeterson]

Greg - is there any generally accepted +/- range either side of a theoretically determined "flat ride" speed that is considered to provide a good (if not precisely optimum) ride? I have committed this problem to a spreadsheet and would like to be able to add a so-called "speed range for a good ride" as an output. Obviously, this is a judgement call, so I'm looking for the voice of experience here.

Thanks in advance.

Norm

 

[GregLocock]

No, I don't remember a figure. I think it was Olney who first proposed it, it may be worth checking Milliken's discussion and re-issue of Olney's work - I don't have it.

The tolerance is very broad, once you have the basics in place. The main reason for this is that the motion is heavily damped so you only have three cycles at most to worry about, and the first quarter of the first one is a bust, so that leaves you with 2.75 cycles, of rapidly decreasing amplitude.

1.3 seems to be a very common answer, by the way. Our solar cars have always been very softly sprung, and the ride has been excellent, just by using this approach.

If you want to check your results I have a Mathcad worksheet somewhere for this. I don't do the ride tuning of the car, that is done using the obvious instrument, so I don't have any advanced tools for examining this.


Cheers

Greg Locock

 

[JakubMech]

Im investigating that very same problem. Just running a few models in a dynamic solver.

I've found 1:1.3 a few times in literature, although I know of people who use 1:1.2 and others which are close but not 1:1.3. The k^2/ab term would be effecting what ratio it would be yes?

A quick question if I may, what kind of damping range would standard road cars have, Ive been guessing between 0.25 to 0.5 (damper damping only) for a simple quater model?

Jakub

 

[GregLocock]

I'm sure if you've looked at a damper characteristic you'll apppreciate that that is a hard number to give, but the rule of thumb in the UK is to settle within 1.5 cycles. I think (very hazy) that is about 40% critical damping

yes k^2/ab is important, as is the choice of cruising speed and wheelbase.

Cheers

Greg Locock

 

[NormPeterson]

In "Race Car Vehicle Dynamics" the Millikens offer 0.15 as the damping value for best ride in a "compact passenger car". About 0.45 is given for best roadholding for the same vehicle.

Norm

 

[Tmoose]

Don't many street shocks have pretty minimal compression damping? Is it clear if the .15 the rebound damping or maybe an average of the 2?

 

[JakubMech]

I know this is slightly off topic a little, would the absolute optimum setting be critically damped (zeta = 1.0), is this at all even possible to have this in the ride mode?? Or is it a comprimise with other modes, ie pitch and roll that less than critical is used?

Jakub

 

[xs650]

Jakub
Critical damping isn't hard to achieve, but it results in harsh ride and limits the ability of the wheels to follow the road, particularly when the road drops away from the car. That results in lousy traction on anything less than a very smooth flat road.

Dick

 

[NormPeterson]

Tmoose - The RCVD discussion appears to assume that the 0.15 is directly applied to both bump and rebound damping, as there is no indication of their being different.

Norm

 

[willeng]

Thanks:
It was good to hear your thoughts on this.
I can understand it for road cars.
Maybe i'm to thick but i just can't see why there so concerned about this method for race cars with 10,000,000 variables & a stop watch to worry about. Especially when there cars show a need for better suspension.
Thanks again.

 

[NormPeterson]

Might I suggest driver performance over the duration of an event as a function of his (dis?)comfort as being variable # 10,000,001?

Norm