RE: BMW M8 confirmed with new adjustable brake tech
Discussion
Haltamer said:
That's even cooler.
Nothing more impressive IMO than the miracles that can be performed with individual wheel braking; Agile handling assist; faux LSD's, ESC, ABS and the likes.
It’s certainly not the most rubbish job in the world! Nothing more impressive IMO than the miracles that can be performed with individual wheel braking; Agile handling assist; faux LSD's, ESC, ABS and the likes.
Getting to drive a lot of pre production stuff at some cool places around the world is one of the perks. Like every job, there’s plenty of ‘normal’ meetings, excel spreadsheets and admin work too, so it’s not all Nürburgring and Dubai!It’s an interesting job to work in though. We’re right at the end of the line for functionality, so it can be fairly entertaining trying to get cars to a point we can do our work. We end up having to have a pretty broad understanding of all the electronics, dynamics and powertrain and how it all works.
The control systems now are really impressive. With the advent of BEVs there’s a lot of opportunity for new ideas and new challenges. Some of it is really pushing what is actually possible with technology!
YEAAAHHH< lets put lots of pointless electric tech on this 2 tonne plus £90k whale of a car so it feels just like our £10k cars did 25 years ago!! Booooyaah
It does seem rather pointless, cars have moved on, so what's wrong with how the brakes feel on the M6 etc? They're not "sports cars" they're big burly sporty cruisers. Nowt wrong with that.
It does seem rather pointless, cars have moved on, so what's wrong with how the brakes feel on the M6 etc? They're not "sports cars" they're big burly sporty cruisers. Nowt wrong with that.
Edited by poo at Paul's on Thursday 9th May 20:34
RacerMike said:
The control systems now are really impressive. With the advent of BEVs there’s a lot of opportunity for new ideas and new challenges. Some of it is really pushing what is actually possible with technology!
I'm a big fan of the NSX for this reason:- With electric motors on the front wheels and modern brakes, you've got instantaneous "STOP" and "GO" on each front wheel independantly; It's very exciting for a future where every bit of traction can be fully utilised for turn in / power etc. in real time:- Whilst there's an appreciation for full analogue control, It's impressive when I feel the assistance systems on my fairly boggo Civic turn understeer into an instant turn in.I was discussing the topic of ESP with a friend the other day; Whilst it seems simple on the surface (And is relatively so for basic ABS / TC systems), but when given yaw, pitch, roll and wheelspeed sensors It'd take me a fair while to knock that into an ESP system.
Out of intrest, how are the systems "Tuned"? I'm guessing now they are largely software based, given the variable TC modes available on many cars. (Drift mode et al, Especially the McLaren Example)
It'd be intresting to see the programming behind it all.
anonymous said:
[redacted]
Whilst the pedal feedback is to some extent decoupled from the actual braking it is just proportional to the braking effort and uses springs and rubber bump stops to provide the 'feel'. When I've got some time I'll have a think/lplay about whether it could be frigged in some way to provide different feel for different 'settings' but I don't think it would.SBC won't slam on the brakes if you open the door whilst underway but will happily take the fingers off the unititiated in the workshop if they're working on it and someone opens a door.
Stability guys get all the glamorous trips
Krikkit said:
Does anyone build their own ABS systems these days?
Did they ever?Weren't the original (automotive) ABS systems Bosch? At least when they went mass market rather than some 50's curio. No doubt aero systems that came before them were made by a sub-supplier too. Something tells me Dunlop were involved?
Think PH did an article on it actually.
Haltamer said:
Out of intrest, how are the systems "Tuned"? I'm guessing now they are largely software based, given the variable TC modes available on many cars. (Drift mode et al, Especially the McLaren Example)
It'd be intresting to see the programming behind it all.
If you've ever had any experience of Matlab and Simulink, you'd have a pretty good idea of how the backend of most of them work. And indeed how most automotive software works. Usually a back end C type language which is then integrated into a system style input/output layer which allows various math channels, sensor inputs, state machines and outputs to be easily integrated. There's then a third calibration layer which enables modification of the PID controller behaviour and functional switches.It'd be intresting to see the programming behind it all.
The first layer is generally programmed (and closely guarded by) the supplier. This is where the majority of the IP lies. The model based layer is also pretty well tied down as it provides the strategy and methods for doing certain things. The third layer is then often what the OEM has access to, and it allows the way the various controllers behave to be 'tuned' in the various situations the car encounters.
The reality is there's a combination of all three things. Early on, logic changes may be needed to achieve what the OEM wants, and equally there may be new requirements that are added that require new logic.
There's now more of a shift to OEM owned IP. It's where many see their competitive advantage coming from in future. Tesla for instance, I believe, use a domain based control with their own software for things like traction control. It makes absolute sense they do it like that as they're a tech firm primarily, and owning their own IP is probably the key to their survival as a company. Porsche I believe also take ownership of some or all of their software logic. They effectively just buy the hardware from the supplier and the electronics merely host the software. That software sits as a black box controller inside the ABS hardware which coverts the requests made by the OEM owned logic into solenoid and pump actuation to deliver wheel torque and reductions in engine torque.
To actually tune the 'feel' of it....it's a case of putting the car in an unstable situation, feeling what the car does, taking a trace, analysing the trace and then working out which part of the controller is doing what. And changing it! Good example would be doing a big lift off oversteer through a fast wet corner on a wet handling circuit, letting the oversteer build and then putting a delayed amount of corrective lock in. It's up to us to then decide whether we think it conforms to the attribute we expect, tune it and then eventually sign it off. It's a bit like colouring in a big picture book of dynamic behaviour in all the different conditions a car will see.
Must be hard to get reliable results out of track work, you can't just drive the same corner over and over while keeping all parameters the same. We're seeing a little bit coming through to rig testing these days, not sure how data's analysed though, hard to judge how it feels if the car's not moving and no one's in it..
scarble said:
Must be hard to get reliable results out of track work, you can't just drive the same corner over and over while keeping all parameters the same. We're seeing a little bit coming through to rig testing these days, not sure how data's analysed though, hard to judge how it feels if the car's not moving and no one's in it..
Biggest issue for sim or rig work currently is the lack of a tyre model that works post limit. It's monumentally non linear, and varies hugely depending on the tyre, so trying to tune a post limit system with a tool that can't actually simulate post limit behaviour is currently impossible.Some aspects are more objective. There's clear metrics for stopping distance and controllability for ABS stops, but trying to judge whether the amount of engine torque down coming out of a hairpin on the road was too much, too little, too harsh, held on too long etc etc has to come down to an individuals experience. Most people come from a dynamics/motorsport/racing background generally so the actual repeatability of the driving is reasonably high. And with the good analysis tools you have (a direct connection into the software with the ability to log all 20,000 plus signals that are calculated/modelled and measured), you can pretty confidently see what part of the controller causes the issue you feel. But of course....it's not actually always the ABS that causes the issue. It could be transmission, powertrain, E-Diff or any one of any of the other systems that can influence engine or wheel torque both positively or negatively.
I know in theory we can do simulation of slip in the direction of rotation, think it uses Pacjeka model.. Whatever that is, but not used it myself. Think we have a way to come before everyone is using and trusting this though. Have no reason to trust it myself I guess, never seen it corrolated to real world.
Obviously inertia/rotation simulation needs to be done for any vehicle system that takes it as an input and that can get a bit tricky.
Obviously inertia/rotation simulation needs to be done for any vehicle system that takes it as an input and that can get a bit tricky.
scarble said:
I know in theory we can do simulation of slip in the direction of rotation, think it uses Pacjeka model.. Whatever that is, but not used it myself. Think we have a way to come before everyone is using and trusting this though. Have no reason to trust it myself I guess, never seen it corrolated to real world.
Obviously inertia/rotation simulation needs to be done for any vehicle system that takes it as an input and that can get a bit tricky.
The models are there for the car no problem, but its the tyre behaviour which is not currently possible to model. You can get some relatively realistic results in some fringe cases, but actually trying to simulate a tyre over slip at the moment, and make tuning which would correspond to the real world isn't really possible. Much in the same way that ride tuning for a car always needs some degree of real world work as ride is not just a function of the basic forces. It's as much about the interaction of many different complex systems that gives the drive a 'feel' for how the car is riding.Obviously inertia/rotation simulation needs to be done for any vehicle system that takes it as an input and that can get a bit tricky.
Generally engineers hate problems that can't be given precise metrics as the assumption is 'it's just physics'. However, where attributes are involved, there is always a degree of the unknown.
I love this debate.. You're right, I want to boil it all down to first principles. Anything can in theory be simulated right?
As far as I know the non linear friction is simulated and so is weight transfer, steering angle and "attitude" I think is the phrase for vehicle angle vs. tangent of the turn? But lateral slip isn't really, I'm sure it could be given as an input to the force on the simulated tire, not sure if our system does that or not, but obviously there's no real lateral loading. But why would that matter?
There's definitely a point though where the effort and cost of an increasingly complex simulation ends up outweighing the cost of real world testing, especially for smaller OEMs, particularly when it's always got to be checked in the real world anyway.
As far as I know the non linear friction is simulated and so is weight transfer, steering angle and "attitude" I think is the phrase for vehicle angle vs. tangent of the turn? But lateral slip isn't really, I'm sure it could be given as an input to the force on the simulated tire, not sure if our system does that or not, but obviously there's no real lateral loading. But why would that matter?
There's definitely a point though where the effort and cost of an increasingly complex simulation ends up outweighing the cost of real world testing, especially for smaller OEMs, particularly when it's always got to be checked in the real world anyway.
Some great posts Racer Mike.
I started in tyre dev., then moved to subjective testing, then I tuned ABS, traction and stability control for a few years. Then moved into vehicle dynamics with a generous side helping of roll over prevention. Now I am mainly involved with adaptive damping suspension tuning, but soon to start again with a detailed look at hardware effects on yaw damping.
I started in tyre dev., then moved to subjective testing, then I tuned ABS, traction and stability control for a few years. Then moved into vehicle dynamics with a generous side helping of roll over prevention. Now I am mainly involved with adaptive damping suspension tuning, but soon to start again with a detailed look at hardware effects on yaw damping.
te, the last thing I want to have adjustability built in is braking. I dread where this is all going...foxhounduk said:
I MIGHT BE MISSING SOMETHING BUT CARS ALREADY HAD A "GOOD" BREAK-FEEL IN THE PAST- GO BACK 20 YEARS. I LOVE HOW THEY KEEP ADDING GUBBINS WITH MORE SENSORS AND ACTUATORS TO MAKE IT FEEL LIKE HOW A GOOD BREAK PEDAL USED TO FEEL,WHEN EVERYTHING WAS MUCH SIMPLER.
Gubbins? GUBBINS? WE DON'T WANT GUBBINS!!anonymous said:
[redacted]
Most of the automotive measurement, calibration and software world is sewn by by Vector and ETAS (which I believe both grew out of Bosch!), and they do a lot of webinars, conferences and e-learning.ETAS:
YouTube: https://www.youtube.com/user/etasgroup
Events: https://www.etas.com/en/regPage/germany_de/en/3744...
Twitter: https://twitter.com/ETAS_global
Vector:
Events: https://www.vector.com/int/en/events/overview/
E-Learning: https://elearning.vector.com/
There is an industry magazine, although it's pretty rare that I see it: https://www.automotivetestingtechnologyinternation...
As was said earlier by Racer Mike, the braking system is pretty much what is in the Alfa. BMW are now telling the world they are adapting it too. However they seem to have gone further by offering some sort of stability control function.
I've been driving my Alfa QF for approx 5000 miles now across Europe as well as the heavy UK traffic. I used to enjoy the brake pedal feel and ability to modulate the pressure on some older servo equipped cars. On rally cars you needed harder pressure, but once the system was warmed up it was perfect. What I have noticed with the Alfa system is that at higher speeds; above say 50mph the pedal can feel nice and progressive. At lower speeds say around 25-30 it can bite firmly, even with the gentlest pressure.
I've been driving my Alfa QF for approx 5000 miles now across Europe as well as the heavy UK traffic. I used to enjoy the brake pedal feel and ability to modulate the pressure on some older servo equipped cars. On rally cars you needed harder pressure, but once the system was warmed up it was perfect. What I have noticed with the Alfa system is that at higher speeds; above say 50mph the pedal can feel nice and progressive. At lower speeds say around 25-30 it can bite firmly, even with the gentlest pressure.
mwstewart said:
lol. More incremental introduction of technologies that can facilitate a restricted level of control by the driver, but marketed in such a way to make it sound like progress. No thanks.
Not entirely true in this case. IPB/Brake By Wire isn't really taking any control away from the driver, and it's introduction isn't to do so. It's primary drivers are a reduction in weight (good) and complexity (good), and improvement in performance (needed to deliver some of the safety features needed by incoming legislation) and cost.Before worrying too much, I'd recommend having a go. Honestly, it feels like a big step on from the semi by wire systems currently around, and it highlights as well that no brake system for the last 30 years has really had any actual brake feel. I'd argue that you don't actually get any 'brake' feel with any servo assisted pedal as the force variation is way too small to make it through the assistance. My Caterham race car arguably did have the potential to feel when a wheel was about to lock, but I can't honestly say if that was entirely through the pedal. It was probably a combination of everything.
Gassing Station | General Gassing | Top of Page | What's New | My Stuff