Return of active suspension - 2017

[BanMeToo]

-if the system were truly active, could it not be programmed to behave as a passive system does?

To what end? That sounds like the worst of both worlds.

[prokopi]

Active suspension in its complete meaning is to control and govern the travel of wheels in a most efficient way.
Generally speaking i vote PRO/FOR sophisticated active suspension in F1 cars.

[DaveW]

Without knowing anything about the nature of the active system being proposed (if there is such a thing)...

I think that WilO has it right. The term "Active Suspension" simply means that power can be applied to the suspension. It might mean, perhaps, that the current cross coupled systems will actually work as the designer intended (most don't, in my view), or (less plausibly, I think) it might refer to a fully controlled irreversible suspension.

To what end? That sounds like the worst of both worlds.

Either way, it won't be worst of both worlds...

[Kiril Varbanov]

Brundle warns over active suspension, while Paddy Low thinks it could be simpler than the current complex systems - http://www.autosport.com/news/report.php/id/113629

[SiLo]

Seeing as it was such a performance booster back in the day, it would make sense for teams to try and emulate that with creative ideas for passive systems, which obviously cost money. Look at the system Renault tried to get pushed through a few years back whilst under braking, it got banned. I bet they wasted a whole host of money on that one.

I can see it easily reducing costs after the first year possibly, but then I'm pretty sure the aero guys would start designing cars that react differently in different states of suspension travel which could cause a few headaches and empty pockets.

[Chuckjr]

but it would cut down on the number of parts that have to be developed, manufactured and installed for each race

It makes sense that it would cut costs, being much simpler system than trying to achieve complex behaviour with conventional mechanics.
Also much simpler to change the setup, just load a different software setting... they can test multiple settings on one run, no need to develop, produce and bring that many mechanical parts and no need for mechanics to work like crazy to make a small change.

So active suspension will simplify the car as well as reduce parts count and cost by replacing a few metal bars with a an entire electromechanical subsystem.

Seems legit.

-if the system were truly active, could it not be programmed to behave as a passive system does?

If you're going to make your active system behave like a passive one... why don't you just stick with the passive system?

Consistency between tracks?

[DaveW]

There are various ways to implement "active" suspension.

One would be to control ride height of each "strut" actively. That was the principle developed successfully by Williams, using hydro-pneumatic actuators. Another way would be to support passive springs and dampers with a hydraulic actuator mounted in series at each corner of the vehicle (I like to call that the "Mumford" system because, historically, he was responsible for implementing the system on a Jaguar). It is not dissimilar (in principle) to the passive "FRIC" systems. All of these systems might be called "ride height" controllers (fast or otherwise) because, whilst they do not interfere directly with the springs & dampers, they do control average ride height of each "strut" actively. Power is required to control ride height, but not to drive the passive elements. Control systems are relatively straightforward, the major issue being to achieve control without coupling with the passive elements of the system.

Another way would be to provide an irreversible hydraulic actuator at each "corner", and to have complete control over the motion of the actuators in response to vehicle measurements. This was the system developed by Lotus. Perhaps it should be added that the Lotus system did run springs, but they were simply used to support the actuators - to minimise actuator CSA's, and to provide a soft failure mode. Such a system would require no power to maintain ride heights, but would consume power when responding to road inputs. The system was made to work successfully by controlling "transport delay", the time required to (start to) respond to an input. The Lotus system had a millisecond iteration rate, but a fixed transport delay of around 120 microseconds. That "nailed" most of the potential instabilities (mainly structural feedback), but it still required some filtering to achieve stable operation. I am not aware of the capabilities or the structure of the "SECU", but I think that the Lotus solution would probably best be implemented using a dedicated processor and I/O system. For that reason I thought the solution, though attractive, was the least plausible of the options.

If he will forgive me, here is an extract garnered from one of Scarbs' articles (in Autosport);

"Developing purely mechanical systems to provide the non-linear compliance required for the key suspension characteristics - heave, roll and individual wheel movement - has required massive investment and huge complexity at each end of the car. Therefore, each end may have some or all of the following: spring, damper, inerter and bump rubber for each of the three elements, such that the engineer may have as many as 32 different compliance parameters for the car! The further level of interlinking only compounds that complexity. It has long been argued that, given the advent of the FIA standard ECU (SECU) and software, active suspension's reintroduction would be a positive move for cost control without the fear of driver aids creeping back in."

He adds:

"Hydraulic fluid is delivered to the servo valves by the pump driven off the engine and the valves opened by electronic control from the FIA SECU. Each of the actuators used in coordination with these can replicate every aspect of the car's current suspension through software algorithms. The current development process for suspension is; design, simulate, build, rig test and track test. With active suspension, it would be to code the software and send the results straight into the simulator for assessment."

That implies, I think, that Scarbs favours a "fully controlled irreversible" solution. He is quite right that it would expose the algorithms and parameters to inspection, although the number of parameters might be a problem (57 in the case of the Lotus system, I recall, not including measurement scaling and offsets). Overall, I'm not sure that I would relish the task of inspecting and interpreting the algorithms, however, or indeed know where to draw the line, unless they are supplied as standard building blocks. Neither am I sure that such a system can be integrated into a car without rig and track testing. However, I applaud Scarbs article.

[gixxer_drew]

There are various ways to implement "active" suspension.

One would be to control ride height of each "strut" actively. That was the principle developed successfully by Williams, using hydro-pneumatic actuators. Another way would be to support passive springs and dampers with a hydraulic actuator mounted in series at each corner of the vehicle (I like to call that the "Mumford" system because, historically, he was responsible for implementing the system on a Jaguar). It is not dissimilar (in principle) to the passive "FRIC" systems. All of these systems might be called "ride height" controllers (fast or otherwise) because, whilst they do not interfere directly with the springs & dampers, they do control average ride height of each "strut" actively. Power is required to control ride height, but not to drive the passive elements. Control systems are relatively straightforward, the major issue being to achieve control without coupling with the passive elements of the system.

Another way would be to provide an irreversible hydraulic actuator at each "corner", and to have complete control over the motion of the actuators in response to vehicle measurements. This was the system developed by Lotus. Perhaps it should be added that the Lotus system did run springs, but they were simply used to support the actuators - to minimise actuator CSA's, and to provide a soft failure mode. Such a system would require no power to maintain ride heights, but would consume power when responding to road inputs. The system was made to work successfully by controlling "transport delay", the time required to (start to) respond to an input. The Lotus system had a millisecond iteration rate, but a fixed transport delay of around 120 microseconds. That "nailed" most of the potential instabilities (mainly structural feedback), but it still required some filtering to achieve stable operation. I am not aware of the capabilities or the structure of the "SECU", but I think that the Lotus solution would probably best be implemented using a dedicated processor and I/O system. For that reason I thought the solution, though attractive, was the least plausible of the options.

If he will forgive me, here is an extract garnered from one of Scarbs' articles (in Autosport);

"Developing purely mechanical systems to provide the non-linear compliance required for the key suspension characteristics - heave, roll and individual wheel movement - has required massive investment and huge complexity at each end of the car. Therefore, each end may have some or all of the following: spring, damper, inerter and bump rubber for each of the three elements, such that the engineer may have as many as 32 different compliance parameters for the car! The further level of interlinking only compounds that complexity. It has long been argued that, given the advent of the FIA standard ECU (SECU) and software, active suspension's reintroduction would be a positive move for cost control without the fear of driver aids creeping back in."

He adds:

"Hydraulic fluid is delivered to the servo valves by the pump driven off the engine and the valves opened by electronic control from the FIA SECU. Each of the actuators used in coordination with these can replicate every aspect of the car's current suspension through software algorithms. The current development process for suspension is; design, simulate, build, rig test and track test. With active suspension, it would be to code the software and send the results straight into the simulator for assessment."

That implies, I think, that Scarbs favours a "fully controlled irreversible" solution. He is quite right that it would expose the algorithms and parameters to inspection, although the number of parameters might be a problem (57 in the case of the Lotus system, I recall, not including measurement scaling and offsets). Overall, I'm not sure that I would relish the task of inspecting and interpreting the algorithms, however, or indeed know where to draw the line, unless they are supplied as standard building blocks. Neither am I sure that such a system can be integrated into a car without rig and track testing. However, I applaud Scarbs article.

I considered the Lotus solution the most promising and I think it also had the most implications for passenger cars. I am honestly surprised that we did not see more development in this area over the years. I have often considered picking up one of these rare cars that were built with active suspension to develop it on my own. It seems like the manufacturers developed versions of them that just did simple tasks and did not contribute much to ride comfort or handling. I have been searching for some way to get a custom solution done that I could apply to racing cars, similar to the simpler systems.

[DaveW]

... that just did simple tasks and did not contribute much to ride comfort or handling.

That, if I may say so, is an opinion spread mainly by people who have never experienced the Lotus system. I admit to being biased, but if I could buy one, I most certainty would. They were far superior to the current crop of "actively" damped systems, both for ride and handling, even without preview - which they didn't need (Greg).

[gixxer_drew]

... that just did simple tasks and did not contribute much to ride comfort or handling.

That, if I may say so, is an opinion spread mainly by people who have never experienced the Lotus system. I admit to being biased, but if I could buy one, I most certainty would. They were far superior to the current crop of "actively" damped systems, both for ride and handling, even without preview - which they didn't need (Greg).

Perhaps I misunderstood but I think you just said the same thing as I was trying to say. That the lotus system could contribute well to ride comfort and handling in a production car reality.

I just felt that how the systems were actually implemented in many road cars did not. One system I was looking at a few months back did nothing more than "leveling" and I thought it was quite unfortunate to go to all the effort of a system like that to use it as a stabilizer bar with an on/off switch.

To tell you a sad truth. When I explain how the Lotus system worked many engineers don't even understand why it would be a benefit over the other systems... It seems people have this idea that aero ride height control was the only value, hence no tie to production car value. I am often disappointed by current crop of people I meet in vehicle dynamics circles. The Lotus system made the entire damper concept look mid 20th century to me.

[beelsebob]

Personally, I find it quite unfortunate to go after complex, heavy mechanical solutions, when an active system can mimic any mechanical system you care to throw at it, and do more interesting things, while still being lighter and smaller.

[Tommy Cookers]

with respect, as they say ........

for 20 years we have had F1 tyres doing the suspension's job
though the servo-hydraulic Lotus F1 approach with suitably lavish software could work well even with solid eg metal 'tyres'
it was overkill, demonstrated by the need for action to prevent the car's structure being 'destruction tested' by the suspension

this approach should be 'designed down' to a rational frequency response ie only as required for suspension with pneumatic tyres
(no doubt others thought about this)
the job could then be done with linear electric actuators or electromechanical actuators (ie no hydraulics)
these could even generate electricity
more so if such actuation was the basis of a Williams F1 style approach
and such a system would better match the direction of road and F1 car technologies
(at the time of F1 active ride I designed electromechanical substitutes to match the performance of in-house servo-hydraulic systems)

but of course F1 is already stuffed with hydraulics
and one wonders what actuation MB has in mind for their their road car F1 style active ride application 5 years from now

[DaveW]

Also with respect....

for 20 years we have had F1 tyres doing the suspension's job

True for (some) front suspensions, but not for rear suspensions.

... though the servo-hydraulic Lotus F1 approach with suitably lavish software could work well even with solid eg metal 'tyres' it was overkill, demonstrated by the need for action to prevent the car's structure being 'destruction tested' by the suspension

Wow. That would be impressive. Out of interest, where did you hear about that.... Incidentally, we concluded that an active suspension roughly halved the vertical suspension loads generated by road inputs compared with an equivalent passive set-up.

... the job could then be done with linear electric actuators or electromechanical actuators (ie no hydraulics)

I don't think so, not without gearing down (voice coils are not good at carrying a DC load, for example aero loads).

[Tommy Cookers]

IMO an electromechanical active strut system could have a frequency response up to 5-7 Hz (ie less than the servohydraulic)

what frequency response is necessary ? ....... and how/why ?

[siskue2005]

Personally, I find it quite unfortunate to go after complex, heavy mechanical solutions, when an active system can mimic any mechanical system you care to throw at it, and do more interesting things, while still being lighter and smaller.

And at lower cost