What the 'Fric' is it?

[DaveW]

My guess is that we're looking at 5bar/75psi or less, just to keep it simple and light.

As it happens, I used 10bar (dampers are typically charged to pressures of between 10 &, perhaps, 25 bar), but recognised that P.V is what matters (higher pressure implies that a smaller volume can be used, but at the increasing risk of a leak).

[ringo]

Ringo: I sympathise with your point of view, and acknowledge that the stiffness of a fluid (& supporting structure) may be high compared with springs of a suspension. However it is a fact that the compliance of the damper fluid can have a significant effect on the performance of a suspension.

In an attempt to explain my point, here is an extract from a rig test of a vehicle that happens to be in my rig. The vehicle is a fairly conventional "tin top" race vehicle, the plot describes the suspension characteristics and the legend presents (linear) estimates of the suspension parameters (in "racing" metric units). For reference, Mu are the unsprung masses, Ki are the "installation" stiffnesses, Kw are the normal spring rates, Cw are the damping coefficients and Mi are the inertance values (zero in this case). The solid lines are the measurements, and the superimposed crosses are "theorectical" values computed from the estimates. You might agree that the estimated values are fairly representative of the measurements (apart from low frequencies, where the accelerometer based measurements fall over).

Contrast that with this plot. Here the parameter estimates are identical except for the Ki values, which have been (artificially) set to zero. The phase angles (representing the effective damping of the suspension) start to diverge observably at 4 hz, and are only half the theoretical values at 20 Hz., correspondinng to the natural frequency of the hub modes. It must be concluded that the "installation" stiffness caused a major reduction in effective damping of the hub modes. The installation stiffness of the rear axle was estimated to be 1,940 N/mm. A damper typically measures at 3,500 N/mm so the structural element was a respectable 4,350 N/mm (actually estimated directly at 4,600 N/mm). So it has to be concluded that the damper element is the principle source of compliance. You will have to trust me when I say that this was mostly caused by fluid compliance (i.e. I have yet to prove otherwise).

Does it matter? Probably not in this case, but with the damper consuming a significant part of the "stiffness budget", it would not require are huge change in structural compliance to actually lose control of the hub mode (inevitably, this has happened in the past).

Edit: An old & much missed friend has suggested that I point you to this reference for a more direct contribution to the effect of compliance. I did reference it earlier, but maybe it was overlooked.

Ok, I'll have a look at the data. As i say i'm no expert, so it's gonna take a while to digest it.
However, i must ask, how much of the installation stiffness of the damper is the gas separator section with the nitrogen gas? This section is an air spring made within the damper. It's probably making it's contribution to the installation stiffness more so than the fluid.

As for the fluid itself, it does behave like a spring when it is moving, a dynamic rate spring. The only issue is energy storage, it wont be able to do this like metal spring. So rebound is not possible after steady state. The nitrogen gas would account for the rebound we get in our dampers. Correct me if i'm lacking detail.

[ringo]

Dave the relationship seems linear in theory.
Why is it that the measured plot doesn't change between both graphs?
Is it that it would be impossible for you to have zero instalation stiffness in reality and only did the theoretical run?
edit: what in on the vertical axis on the top graph?

[DaveW]

However, i must ask, how much of the installation stiffness of the damper is the gas separator section with the nitrogen gas? This section is an air spring made within the damper. It's probably making it's contribution to the installation stiffness more so than the fluid.

No contribution, hopefully, but it may add some series stiffness (aiding the steel spring) and some "preload" equal to the charge pressure multipled by the shaft area, if the damper isn't a "through rod" design.

As for the fluid itself, it does behave like a spring when it is moving, a dynamic rate spring. The only issue is energy storage, it wont be able to do this like metal spring. So rebound is not possible after steady state.

A definition of a spring is that it gives up energy it stores without loss. Will it work in rebound? Actually yes, better than an unconstrained coil spring, actually - until the fluid cavitates.

Why is it that the measured plot doesn't change between both graphs?
Is it that it would be impossible for you to have zero instalation stiffness in reality and only did the theoretical run?
edit: what in on the vertical axis on the top graph?

The change was only made to the model (the crosses). The units are N/mm...

[DaveW]

A couple of other thoughts regarding "FRIC":

I believe that teams are allowed to replenish fluids & set tyre pressures between qualifying and race. Perhaps they are also allowed to reset "FRIC" charge pressures.

That opens up the possibility for organising a controlled leakage, to maintain static ride height as fuel is burned off...

[Dragonfly]

A couple of other thoughts regarding "FRIC":

I believe that teams are allowed to replenish fluids & set tyre pressures between qualifying and race. Perhaps they are also allowed to reset "FRIC" charge pressures.

That opens up the possibility for organising a controlled leakage, to maintain static ride height as fuel is burned off...

I've written about such a possibility far back in this thread, but you guys seem to go in so much theoretic detail while I, with my amateur view, think that the system is much simpler than some of you are suggesting.

[marcush.]

I thought I had suggested a similar device a few times before...

[marcush.]

currently my thinking is going along the lines of a hydraulic pump actuated by the bump of the pushrod pressurising a ram which is the end eye of the damper /spring support (in case of sauber this would be the rod connecting the two torsion bars ) .The ride height itself is under governance of a bleed which is opening /exposed when the set rideheight is achieved.
This is a quite simple setup totally mechanical and only operating under wheelloads .presto .fully legal .

[ringo]

As for the fluid itself, it does behave like a spring when it is moving, a dynamic rate spring. The only issue is energy storage, it wont be able to do this like metal spring. So rebound is not possible after steady state.

A definition of a spring is that it gives up energy it stores without loss. Will it work in rebound? Actually yes, better than an unconstrained coil spring, actually - until the fluid cavitates.

Keep in mind i'm referring to a cylinder with just oil and orifices, that was claimed to be working in the mercedes. What you say is true with visco-elastic dampers of course. There is a spring component.

However for a pure damper, a viscous damper as claimed by the journos, (i feel this is only possible with a straight through damper, if it wasn't one then the piston would decrease volume in the chamber as it enters hence compressing the oil, which is a no no). I don't think it will rebound if you push down the piston. The kinetic energy of the piston will dissipate to heat as the viscosity slows it down. This was the core of my argument. There is no equation for energy storage in that case.
I can see where an automotive damper rebounds, but i feel this is attributed to the gas charge taking some of that kintectic energy from the piston and storing it, then releasing the stored energy after steady state and change of the external force. More than the fluid itself doing any springing.

Mx``+Cx`+Kx=Force.
We can see here that once the piston stops moving (x`=0)we're left with Kx=Force,this is the stored component of the energy, the spring. The damping properties go to zero.

If you have a damper that is open to atmosphere, or a straight through piston damper, it wont rebound when you push the piston, as there wont be any energy to store as there is no gas to compress. It will dissipate as heat.

[ringo]

In fact come to think of it, it is quite impossible for mercedes not to use air springs if they don't have straight through piston dampers. ...
I would say that mercedes do have springs on their car. either air or metal. If they are depending on hydraulic oil compression, then it wouldn't be much of a suspension.
Spring rate is simply too high.

Shifting gears a bit, it was mentioned that the system is probably bleeding. Any thoughts on how this is done?

[mep]

I think you got something wrong ringo. I never heard anywhere that the system uses the hydraulic compression as spring. In fact that would be way too stiff and difficult to adjust. The point of our previous discussion was that the compression of the fluid can’t be neglected.

Even with the interlinked suspension it is IMHO best to use conventional steel springs. What the links do is only to distribute the load to several corners of the car. It might offer new options of pitch and roll damping as well.
Basically it goes like this: A load applied to one end of the car is supported by the springs of the other end. A compression of the front axle for example causes the rear to compresses as well. The compliance of the link and the different spring stiffness’s determine how much.

This is over reaching. There is no piston area to pipe diameter motion ratio. You have the fluid pressure acting on the pipe wall....

This patent is a very good read and gives some insight into the effect of piston to pipe cross section area.
http://www.freepatentsonline.com/WO2011089373.pdf

[ringo]

I think you got something wrong ringo. I never heard anywhere that the system uses the hydraulic compression as spring. In fact that would be way too stiff and difficult to adjust. The point of our previous discussion was that the compression of the fluid can’t be neglected.

People have been saying that if you follow some of the discussions!
The thread has evolved to a point none of us here believe the system has no springs at all, but i was just looking into the damper itself. I can assure you though that there's still a small contingent that would believe anything once it's in a blog. So i was just trying to make the clarifications in this thread.

Even with the interlinked suspension it is IMHO best to use conventional steel springs. What the links do is only to distribute the load to several corners of the car. It might offer new options of pitch and roll damping as well.
Basically it goes like this: A load applied to one end of the car is supported by the springs of the other end. A compression of the front axle for example causes the rear to compresses as well. The compliance of the link and the different spring stiffness’s determine how much.

I am in agreement of this, but read some of the blog articles out there, even ones sited wherever fric is discussed, and there are suggestions that Mercedes aren't using any springs whatsoever and Lotus are to follow suit in upcoming races.

Right now, i'd like to see a nice FRIC diagram, with the ride height adjustment. I have a semi mechanical method for the ride height that doesn't involve bleeding or the fuel tank weight, but it depends on "track memory". It jacks down under big braking events.
Especially since these are mostly at the end of long straights on a track and happen once or twice per lap, it acts as a good jack down clock.
I'll sketch the mechanism.

[henry]

I have been musing on what the aeroynamicists would want out of ride height control. My suggestion would be:

Requirement 1: maintain front and rear nominal ride heights at a fixed value compensating for the compression of suspension and tyres. obvious but here for completeness.

Requirement 2: in addition to 1 adjust front ride height so that the splitter runs closer to the ground at low speeds than it does at high speeds. This should increase air velocity under the floor at low speed without choking at high speed. This should increase low speed downforce without incurring unnaceptable plank wear.

Requirement 3: at some threshold speed, above which extra downforce is superfluous, pitch the car nose up/ tail down to reduce drag/downforce. If top speed is set at the norm for the track ( there always seems to be one) this would allow higher a higher downforce at lower speeds when cornering. It would also reduce the energy going into the tyres at high speeds due to camber distortion and other effects of suspension geometry. It might be useful to pitch the front up by only the amount it will pitch down under the initial braking forces. This would keep the plank off the ground and not require special measures to control extra suspension travel.

Requirement 4: during braking reset 3 and return to ride heights set by 1&2 whilst accomodating suspension and tyre deflections due to pitch.

Requirement 5: during cornering restrict roll to keep the floor parallel to the ground.


In terms of implementation I would propose that 1 to 3 would require a mechanism that raises ride height at the front in response to increasing average downforce. 4 requires the inverse and would need to react more quickly.

I think both mep and davew have proposed using the rear suspension travel to "pump up" the front ride height. I also recall davew has argued cogently that using an inertial switch to select between acceleration and deceleration behaviour would be legal. This would allow selection of the two relatively static modes.

I think controlling roll attitude is more difficult because the response needs to be much faster.

As well as pleasing the aerodynamicist an approach such as this would allow the suspension engineer to reduce spring rates, particularly at the front, so smiles all round.

I suggest that from observation Renault are probably doing something like this. And Mercedes are not.

[mep]

Ok, I have to admit that I did not read all of the comments here so there might be stated somewhere that it uses hydraulic springs. Also I don’t really give much on those (journo) articles which are around. They don’t really give any technical insight and too often are comments misstated or interpreted. Also is Mercedes trying to sell the system in a way to make them look smarter than they are. Almost the complete field is using something like this and it was originally designed by Lotus/Renault.

[marcush.]

I think the first and major concern for the teams is the varying fuelloads coming into the equation.

As you need to run the leading edge of the splitter as low as possible to maximise downforce ,the low fuel setup is a compromise as you would drag the leading edge too much under full fuelload -which would of course change the flow of the undertray dramatically as the middle section would be robbed of its throat and only air from the side would feed the middle part of the undertray robbing the outer thirds of flow -you get crossflows -that cannot be good for downforce and drag?


the second thing is to avoid the car sinking too much at the rear under aero loads -at least while cornering .

the third is to optimise downforce under braking -which is pro dive - but not so much as to drag the splitter -you need to raise the rear enough to compensate the lowering of the fronts (tyre squash) but yet avoid separation ..


So before trying to adapt all those variables it might be wiser to try and transform variables into constants taking them out of the equation....which is what we see at those teams which are in trouble.
This approach will give a chance to optimise your suspension settings for the aero you got and borrow time for the aero guys to try and find aerosoltions which enhance preferred characteristics at the car attitude you will encounter when cornering ,breaking and accelerating or going straight at vmax.