The other active suspension cars - Leyton House

[Tim.Wright]

Please explain how the "low frequency" active hydraulic actuator would be able to keep pace with movements of the "high frequency" passive suspension it is attached to. The digital controller can send command signals to the valve actuator at whatever frequency it desires. But that does not mean the hydraulic actuator will be able to keep up with the movement demanded by the controller.

The actuator has no requirement to "keep up" with the passive damper.

Think of what will happen when it gives a 0hz (stopped) input. It will act as a solid link and and passive system will respond to the road inputs.

The load on the actuator will be chaning at high frequency but its displacement will remain zero. This is not difficult to achieve.

[DaveW]

The digital controller can send command signals to the valve actuator at whatever frequency it desires. But that does not mean the hydraulic actuator will be able to keep up with the movement demanded by the controller.

You have stated the problem quite well, I think, although I would not have used those words. Tim has (patiently) described one solution, which was adopted by WGP and (apparently) by Leyton House.

The frequency and flow rate required by a full active F1 suspension system can be handled by digital servo-valves that have been available since the late 60's from companies like Sturman.

Statements like that can be (are) very misleading. I am not sure how "digital" valves help in what is essentially an "analogue" problem (perhaps you could explain that). It turns out that Sturman Industries was founded in 1989. Much happened in the preceding 20 years, so I think you might have been a victim of salesman's hype....

[riff_raff]

The actuator has no requirement to "keep up" with the passive damper. Think of what will happen when it gives a 0hz (stopped) input. It will act as a solid link and and passive system will respond to the road inputs. The load on the actuator will be chaning at high frequency but its displacement will remain zero. This is not difficult to achieve.

I think you would agree that the main objective of an active suspension system is to provide an optimum chassis condition in terms of pitch/roll/yaw and how it affects F/R and L/R tire traction. There is no way a DA hydraulic actuator working in series with a passive compression spring and dampener would be of much use, except in straightaways or similar conditions, if it does not have the frequency response capability to keep pace with the passive system. Of what benefit is the DA hydraulic actuator if it is fixed? The position of the DA hydraulic actuator also affects the force existing at the passive compression spring for a given ride height at each corner.

I can see how such a system would be helpful in reducing drag on long straights or high speed corners. But it would not seem to help in other situations.

[Tim.Wright]

I'll try one more explanation. Bear in mind that it's not just myself and Dave but also Adrian Newey who is on record as saying the series active system (with the double actin actuator in series with the spring/damper) is the better concept.

Firstly - on the subject of the locked actuator. This was ONLY to illustrate a point - that the suspension will still work with the actuator locked in an "extreme" low frequency. In reality it will be constantly moving but there is absolutely NO requirement for the active system to "keep pace" with all passive system's movements.

Here is the reason why.

The vertical role of ANY suspension is to manage the following 3 inputs
A. Inertial load transfers e.g. due to braking/accelerating/cornering (typically 0-10Hz)
B. Aerodynamimc downforce (0 - 0.2Hz)
C. Road profile inputs e.g. bumps, curbs, track camber and warp (typically 15+Hz)

...with the overall goal(s) being:
1. Minimise ride height changes to improve aerodynamic performance
2. Minimise vertical load variations to improve tyre grip

So if we look at the 3 inputs (A&B&C) - we can see how an "ideal" suspension system would deal with it to keep the 2 goals (1&2) satisfied:

A. Inertial load transfers (e.g from braking/cornering) cause a pitch/roll moment to which the suspension must react.
To maintain goal 1 you need a very HARD suspension to minimise the ride height changes in response to these load transfers.
Maintaining goal 2 (minimising the load transfer) is not possible because load transfer is a function of CG height and track/wheelbase width only. It's independent of the suspension.

B. Downforce ranges from 0 (at a standing start) to tens of kilonewtons of force and it changes relatively slowly.
To maintain goal 1 with a constant ride height from such a large change in force you need a very HARD suspension.
Maintaining goal 2 is again not possible because the vertical downforce LOADS will occur regardless of the suspension.

C. Road profile inputs typically come in as high speed single wheel bumps which are local high spots with respect to the general road surface. Given that the aerodynamic elements need to remain constant with respect to the general road surface (not the bump):
To maintain goal 1 you need to let the wheel lift up over the bump without pushing the body up (and changing the ride height). To acheive this each wheel needs a very SOFT suspension
To maintain goal 2 you need the wheel force to remain constant as it passses over the bump. Again, to acheive this you need a very SOFT suspension.

So basically in the end there is a pattern:
Low frequency inputs (interial load transfer and downforce) - HARD suspension
High frequency inputs (road profiles) - SOFT suspension

In a series active system, both the passive and the active part are subjected to all of the loads - both high and low frequency. The difference is that the passive spring/damper responds to everything while the active system only reponds to what the controller tells it to do.

So the concept of the series active system is to use a SOFT passive suspension to satisfy goals 1 and 2 under input C (high frequency). And then control the active system to respond to only the low frequency inputs (A and B) to cancel out the low frequency component of the spring movements.

To make the suspension effectively "HARD" for the low frequency stuff - the actuator needs to "cancel out" the low frequency spring movements by doing the opposite. I.e. if the spring compresses by 10mm under braking, you need to extend the actuator by 10mm. If the car rolls by 5mm a side in cornering then the actuator needs to "roll" in the opposite way (extending the outer wheel and compressing the inner wheel) to keep the ride height constant. In reality, to also compensate for tyre deflections - you need to add an extra bit compensation from the actuator.

So, how do you control the actuator to move only at low frequencies? Well for a given set of passive springs - you will know, up-front, how they compress as a reaction to the 2 low frequency inputs (load transfer and aerodynamics).
Compression due to braking/accel load transfer is basically a function of longitudinal acceleration
Compression due to cornering load transfer is basically a function of lateral acceleration
Compression due to aerodynamics is basically a function of air speed (from pitot tube)

If you put these three sensor measurements through a low pass filter - then use them to calculate the predicted passive spring responses - you can then command the actuators to do the opposite movement and cancel these low frequency COMPONENTS of the passive spring system.

If, during a particular corner you hit a curb, the soft passive system will let the wheel raise up over the bump in order to maintain goals 1 and 2 - but this movement will not be compensated by the active system because a. it doesn't change the lateral acceleration and b. it's a high frequency input so it's effects will be filtered out by the low pass filter in the controller.

The advantage of this system is precisely that your sensors and controls are able to work at low frequencies/high force conditions which means you can keep your actuator power requirements low and therefore lightweight. A series active system like this shifts the management of the high frequency stuff to the passive systems which does not have bandwidth problems.

[autogyro]

Great explanation Tim and almost exactly how Tony Rudd explained it to me when we were considering the turret stabilising systems on tank gunnery as a comparison.

[DaveW]

Mmm... Remarkable memory, ag. Did you ever meet Bruce Maclaurin?

p.s. How's that man powered autogyro coming along...?

[stresseddave]

Did you ever meet Bruce Maclaurin?

I was Bruce's underling for a couple of years - learnt a hell of a lot in a short time.

As for ag - by the time we'd finished fettling Dave's system on a tank, it was approaching the platform stability of some of the gun stab systems.

[riff_raff]

Statements like that can be (are) very misleading. I am not sure how "digital" valves help in what is essentially an "analogue" problem (perhaps you could explain that). It turns out that Sturman Industries was founded in 1989. Much happened in the preceding 20 years, so I think you might have been a victim of salesman's hype....

It is correct that Sturman Industries was founded in 1989, but the development work Eddie Sturman did on the digital servo valve was done for the Apollo program back in the late 60's. From his bio on Sturman's website: " Digital actuation, our core technology, was developed by Eddie Sturman, for the Apollo program in the 1960s. We continue to advance the technology today."

Digital control of hydraulic valves allows a higher frequency response and more precise control than analog. This should be obvious based on the fact that analog control of high frequency hydraulic servo valves is not used.

[DaveW]

I ... learnt a hell of a lot in a short time.

So did I....

It is correct that Sturman Industries was founded in 1989, but the development work Eddie Sturman did on the digital servo valve was done for the Apollo program back in the late 60's. From his bio on Sturman's website: " Digital actuation, our core technology, was developed by Eddie Sturman, for the Apollo program in the 1960s. We continue to advance the technology today."

So...We agree that Sturman Industries was a late-comer to the party. I wonder if Eddie Sturman once worked for Moog? That would make some sense.

Digital control of hydraulic valves allows a higher frequency response and more precise control than analog. This should be obvious based on the fact that analog control of high frequency hydraulic servo valves is not used.

That is not obvious to me, I'm afraid. The plots at the bottom of this page from the Sturman website appear to compare analogue (blue) with digital (red) actuator performance The left hand plot shows velocity (in some units) and the right hand plot shows position (presumably the integral of velocity). On the face of it, a velocity demand of 1 unit would cause the analogue version to move, but not the digital version. Is that an important difference? Certainly. I would not choose the digital version as an active suspension element because it could not control ride height accurately (for example), and the "chuckle" when the actuators did "let go" would probably be annoying.