30 replies to this topic

[gruntguru]

This is bound to become a very controversial topic yet this article (in revealing the technology has been adopted by Toyota for its bZ4X electric SUV) treats it in a very matter-of-fact way. https://www.designfa...rearticle&pn=04

 

 

A uniquely shaped steering wheel, which eliminates the need to change grip when steering, is part of the vehicle's steer-by-wire system. Using steer-by-wire technology does away with the mechanical connection through the steering shaft and results in a more direct response between steering operation and driving force. It also enables the vehicle to turn with less steering angle and more precision, but it could take some getting used to for some drivers.

 

Whether the steer-by-wire system makes it to the production model is not clear but it does seem the time is right. With more and more people prepared to accept hands-off driving or cars with no wheel at all, why not steer-by-wire?

[Greg Locock]

It would be a significant step forward in safety and cost. Well actually we'd probably accept the same level of safety and go for a bigger cost save. I'd have thought by now we'd have a good understanding of the failure rate of the epas/rack combination. First time I was driven in an SBW was in 1988, in the first car we built with ABS and TC and active suspension. Every single system failed within one lap, the very mechanical clunk as the steering column reconnected was reassuring.

[Ross Stonefeld]

Thanks, I hate it.

[Canuck]

This irritates my tinfoil hat a great deal. Of course I don't spent a lot of time thinking about the dangers of the steering column pointed at my chest, but I'd like to think^* that's been largely mitigated by smart design and airbags.

 

* I'm telling myself that this is true regardless of real-world information.

[kikiturbo2]

considering the lack of steering feel in conventional EPS steering, I really find nothing positive about SBW...

[desmo]

Has nobody really yet done a good job of EPS? 

[Greg Locock]

To be honest the aspirational target for EPAS is 'as good as HPAS' for steering feel. I like to think Ford do a good job at that, since it is part of my job, I'd be interested to hear of any reasonably modern Ford product with perceptible problems with the EPAS.

 

The fundamental issue is that the EPAS motor runs very fast, and is then heavily geared down (33:1 say) and drives the rack via a ballscrew. Backdriving that lot, which is essentially what the tie rod is trying to do to impart feel, is impossible, directly. So, electronically a balance is set up whereby the motor helps to overcome it's own inertia (N^2 relationship for inertia) and the friction in the ballscrew and so on, using feedforward and a lot of tuning. Trouble is, if you don't do enough of that compensation the system feels dead due to friction and referred inertia, and if you do too much then the system 'motors' itself, which is unnerving to say the least. Different vehicle manufacturers have different access to the tuning and strategy levels of the software. We're allowed to go deeper than most, and are learning all the time.

 

Also of course you lot insist we fit stupid low profile tires which create random inputs due to CP width that we mask. We could put a lot less masking into the system if we went back to bicycle tires. BMW used to have a 3 Hz low pass filter in their steering column. Vorsprung dur rubber.

[Chubby_Deuce]

What problem would this solve? FBW makes a lot of sense for airliners with long and complex control circuits that need constant computer intervention. Cars have a very short distance from driver to steering rack and EPS already does an adequate job of adding control inputs.

 

I feel like there's just a generation of designers that really, really want to drive with a Batmobile steering wheel.

[404KF2]

Yeah goodbye steering feel, although EPS even takes care of most of this on moderns.

 

A solution looking for a problem, but likely to create more.

Edited by 404KF2, 07 May 2021 - 21:50.

[desmo]

From my viewpoint and admitting they're unambiguously better than old cars, modern cars are in a lot of ways crazy over engineered.

[Fat Boy]

To be honest the aspirational target for EPAS is 'as good as HPAS' for steering feel. I like to think Ford do a good job at that, since it is part of my job, I'd be interested to hear of any reasonably modern Ford product with perceptible problems with the EPAS.

 

The fundamental issue is that the EPAS motor runs very fast, and is then heavily geared down (33:1 say) and drives the rack via a ballscrew. Backdriving that lot, which is essentially what the tie rod is trying to do to impart feel, is impossible, directly. So, electronically a balance is set up whereby the motor helps to overcome it's own inertia (N^2 relationship for inertia) and the friction in the ballscrew and so on, using feedforward and a lot of tuning. Trouble is, if you don't do enough of that compensation the system feels dead due to friction and referred inertia, and if you do too much then the system 'motors' itself, which is unnerving to say the least. Different vehicle manufacturers have different access to the tuning and strategy levels of the software. We're allowed to go deeper than most, and are learning all the time.

 

Also of course you lot insist we fit stupid low profile tires which create random inputs due to CP width that we mask. We could put a lot less masking into the system if we went back to bicycle tires. BMW used to have a 3 Hz low pass filter in their steering column. Vorsprung dur rubber.

I have a 2020 Ford with EPAS. I agree the steering feel is quite good, even if the ratio is a little slow. I _wish_ there were more sidewall on the tires. I would be quite happy with an extra 30mm or so of sidewall. I have absolutely zero interest in a FBW steering system, regardless of any theoretical positives.

 

Since we're talking street cars, do you ever play with the magnetic dampers? I have a stand-alone controller for mine and have been playing quite a bit with the tuning. You can do some very interesting tricks, but there are also some shortcomings, particularly larger displacement, low-frequency freeway undulations. It seems to work well in high-G situations, but the daily driving bit is more tricky to tune. I did speak to an OE tuner and he looked at the software for my controller. He said it was about 25% of the software he uses.

[Ross Stonefeld]

From my viewpoint and admitting they're unambiguously better than old cars, modern cars are in a lot of ways crazy over engineered.

 

I had to ferry some people to doctor's appointments and hopped between a 1989 Saab 900 and a 2019 Subaru Outback. The latter is nice, albeit takes a few moments of "uh..is this thing done loading?" before I feel comfortable taking off. My problem is I was regularly well over the speed limit without noticing it, there's no sense of speed/weight/any physicality to the vehicle.

 

The Saab has some really clever early-version 'smart driving'. Round about 100kph/60mph the engine sags so you have to make a conscious effort to poke through the speed limit on the highway. I think there's a crack or harmonic happening somewhere but it's not unuseful and mildly amusing.

[Lee Nicolle]

To be honest the aspirational target for EPAS is 'as good as HPAS' for steering feel. I like to think Ford do a good job at that, since it is part of my job, I'd be interested to hear of any reasonably modern Ford product with perceptible problems with the EPAS.

 

The fundamental issue is that the EPAS motor runs very fast, and is then heavily geared down (33:1 say) and drives the rack via a ballscrew. Backdriving that lot, which is essentially what the tie rod is trying to do to impart feel, is impossible, directly. So, electronically a balance is set up whereby the motor helps to overcome it's own inertia (N^2 relationship for inertia) and the friction in the ballscrew and so on, using feedforward and a lot of tuning. Trouble is, if you don't do enough of that compensation the system feels dead due to friction and referred inertia, and if you do too much then the system 'motors' itself, which is unnerving to say the least. Different vehicle manufacturers have different access to the tuning and strategy levels of the software. We're allowed to go deeper than most, and are learning all the time.

 

Also of course you lot insist we fit stupid low profile tires which create random inputs due to CP width that we mask. We could put a lot less masking into the system if we went back to bicycle tires. BMW used to have a 3 Hz low pass filter in their steering column. Vorsprung dur rubber.

I am never going to demand low profile tyres, ride is crap, more feed back through the steering and try to carry a load and the tyres get very hot.

60 profile 16s is the way to go.

And since I often fit tyres to these rims most are shaped every way but round

[Greg Locock]

FatBoy - no Ford Australia is pretty much Ranger/Everest only. The rheo dampers are on Mustang only so far as i know in the Ford world. I'd like to use them at some point, they are very neat tech.

[Fat Boy]

FatBoy - no Ford Australia is pretty much Ranger/Everest only. The rheo dampers are on Mustang only so far as i know in the Ford world. I'd like to use them at some point, they are very neat tech.

 

The ability to dynamically cross-weight the car is a powerful tool.

[Ross Stonefeld]

Now in GT3 cars(racing ones)

 

https://www.motorspo...er-bmw/6506753/

 

The 'tuning out the kerb strikes' thing is...that's a benefit? Sure you don't want your arms knocked off but that sounds like less feedback? I dunno, maybe it's like F1 power steering is probably pretty good and nothing like the low-end-road-car stuff I've driven all my life.

[gruntguru]

I am wondering what sort of fail-safe systems would be required? The nightmare scenario is to be driving down the higway at high speed when one or both front wheels suddenly becomes independant of the steering wheel.

[Greg Locock]

I'm wondering if the ESC could be used to steer the car, or at least bring it to a halt in a straightish line. 

 

Then there's this

 

https://www.schaeffl...lion kilometers'%20worth%20of%20journeys.

Edited by Greg Locock, 16 May 2021 - 23:22.

[Greg Locock]

Better explanation

 

https://www.schaeffl.../steer-by-wire/

 

Well that's that  then!

[gruntguru]

The racing versions would no doubt be subject to a raft of rules. If you think about the F1 rules around throttle by wire - just to avoid the possibility of electronics helping the driver control the car. SBW takes that potential to another level.

 

At a more basic control level, SBW creates the oportunity to steer each wheel independantly to optimise the slip angle at each tyre and not be bound by the monotonic toe relationship dictated by a mechanical steering system.

[Greg Locock]

Yup, although active toe control can get you that.

[Tenmantaylor]

I havn't driven a good one yet. All remove any feel that hydraulic systems had in terms of immediacy and textural feedback. Also think many modern systems fake 'feel' with faked resistance and weight that isn't linked to what the tyre and wheel is actually experiencing. By the time they have actually worked out how to replicate this feeling in an electronic system the lag (even if only 50ms) will never be eradicated and the system will be far more complex and subject to more possible failures.

 

End result. Driving as an experience is being slowly killed by electronics.

[404KF2]

Mercedes had a boatload of fun with their "all electronic" braking system in the E Class a few generations ago.  Recalled 'em all and put normal brakes in.

 

Again, this is a solution looking for a problem.

[gruntguru]

There is a long list of benefits.

[desmo]

Safety conspicuously not among them I expect.

[Ross Stonefeld]

Multiple people in multiple government agencies are rubbing their hands with glee at the potential applications. Triple-redundancy security you say. Well okay.   

[Ross Stonefeld]

Do you trust technology enough to use digital steering for 24 hours racing at the Nordschleife?

 

 

http://www.dailyspor...mn-beckons.html

Edited by Ross Stonefeld, 04 June 2021 - 13:16.

[desmo]

If it's good enough for fighter pilots... of course there's less to crash into in the sky.

[Ross Stonefeld]

Fly by Wire in a plane allows them to do things the pilot can't reasonably achieve though, due to the massive inherent instability of a fighter jet to get the performance they need. 

 

Plus, I dunno, what's the failure rate of a fighter vs a GT3 car in endurance racing? Planes get constant maintenance but bits are usually not falling off and locking up and etc. And tons of redundancies.

 

(didn't FBW, or at least the way it was setup, contribute to the Air France crash in the Atlantic a few years ago? Pilot was pushing down, copilot was pulling up, the controls weren't synced on an Airbus. But are on a Boeing iirc.)

[MatsNorway]

Honestly a bad electrically assisted steering feels dangerous, Toyota Avensis company car had it, no feel and hard on light touch, move it with some force and suddenly its too light with no feel there either.

 

I think no servo should do a comeback. Best feel i ever had was from a no H servo car. You really should drive with driving gloves tho to be able to "match" the speed and ease of H or Electric assist can provide in some situations.

[GreenMachine]

(didn't FBW, or at least the way it was setup, contribute to the Air France crash in the Atlantic a few years ago? Pilot was pushing down, copilot was pulling up, the controls weren't synced on an Airbus. But are on a Boeing iirc.)

 

I think the standard (at least now) is that left seat overrides right seat.

 

I recall an incident of pilot/copilot conflict - I don't know what happened to the call 'my controls'     , and I can't remember the circumstances.

 

At least in most cases, cars only have one set of controls    so that problem is unlikely to arise.