Spring's here, time to die

[Ocean1]

Is that due to the practicality of a just-pre-lockup state being at a level of grip lower than what is ideal, or is there a different reason?

If you graph time/braking force up to where traction starts to drop you see traction spike just before lock-up, as already discussed. What's not obvious is that the steady braking force state that produces optimal STEADY traction is a fair bit lower than that spike. So ABS control strategy is to ramp up braking force to just pre-lockup, (but harder than the above steady pre-lockup force) then ease it and ramp it up again. I've seen tests that demonstrate that even manually pulsing in and out of that time/braking force spike produces better results, and an ABS PLC with a scan rate of a couple hundred hertz should make even better use of time in that spike.

[R-Soul]

If you graph time/braking force up to where traction starts to drop you see traction spike just before lock-up, as already discussed. What's not obvious is that the steady braking force state that produces optimal STEADY traction is a fair bit lower than that spike. So ABS control strategy is to ramp up braking force to just pre-lockup, (but harder than the above steady pre-lockup force) then ease it and ramp it up again. I've seen tests that demonstrate that even manually pulsing in and out of that time/braking force spike produces better results, and an ABS PLC with a scan rate of a couple hundred hertz should make even better use of time in that spike.

ABS strategy is to brake hard enough to lock up, detect lock up, release, and re brake hard in an effort to make average braking force higher than manually braking hard until lockup and relying on the rider having the presence of mind to pulse it manually and fast enough. There is no way for the system to detect "just pre lock up". What are you going use to sense that? Either there is wheel speed difference (i.e. locked up) or there is not.

According to my reckoning, this is why experienced racers can brake harder without ABS on track- they manage to keep the average braking force slightly higher by keeping the wheel on the edge of starting to slip by "feel" or experience - something sensors cannot do, at the point where the wheel is actually laying rubber down on teh road, but is still turning (so its not completely dynamic friction yet).

[R-Soul]

Normal differences of 30%? Has Drew been giving you riding lessons? I'd want ABS to come in at around 10% slip, in the 0-10% slip range is where max grip is found iirc. If it is pulsed (and I'd much prefer it to be a proportional control instead), the frequency should be high enough not to unsettle the suspension, and ideally high enough to barely be felt at all.
Wheelies are kind of a non-issue, not like you'll be counting on much stopping ability from the front during a wheelie.

In one of the last few Kiwirider magazines, the Japanese engineer for the ZX 10 (I think) was talking about how they put together their traction control and ABs systems, and I seem to recall a quote by him that normal speed difference (ie under normal acceleration and riding) between the front and back wheel were already in the order of 15% or higher. I recall this because I was quite surprised at the numbers. I used 30 % as a bit of a thumb suck, allowing for extreme acceleration, and possibly a buffer range before the system makes a final analysis of "lock up".

[bogan]

ABS strategy is to brake hard enough to lock up, detect lock up, release, and re brake hard in an effort to make average braking force higher than manually braking hard until lockup and relying on the rider having the presence of mind to pulse it manually and fast enough. There is no way for the system to detect "just pre lock up". What are you going use to sense that? Either there is wheel speed difference (i.e. locked up) or there is not.

As per my previous statement (#443), the max grip can be found at a small percentage slip, wheel speed sensors can detect small amounts of slip. A wheel will take time to lock up once the grip force is overcome, with good a enough ABS package, this can be detected and the lock up prevented.

In one of the last few Kiwirider magazines, the Japanese engineer for the ZX 10 (I think) was talking about how they put together their traction control and ABs systems, and I seem to recall a quote by him that normal speed difference (ie under normal acceleration and riding) between the front and back wheel were already in the order of 15% or higher. I recall this because I was quite surprised at the numbers. I used 30 % as a bit of a thumb suck, allowing for extreme acceleration, and possibly a buffer range before the system makes a final analysis of "lock up".

Sounds like something was lost in translation there. High performance ABS can't just rely on lockup detection, slip will come into it as well.

[R-Soul]

no grip when locked up , more grip when rolling

Something like that

Stephen

Yes that is the dofference between static and dynamic friction.

The friction factor changes between the two. In mechanical engineering we learnt that if you have a block of material stitting stationary on a floor, and you start pulling it, then you will need a certain amount of force (F1) to start moving it. This force is equal to the static friction factor (because it is at rest) multiplied by its weight. Once it starts moving, then you will need a force to keep it moving at a constant speed (so no additional forces from acceleration then). The force (F2)required to keep it moving at a constant speed will be equal to the dynamic friction factor multiplied by its weight. F2 will always be smaller than F1. If I remember correctly (and this changes form material to material) F1 is typically in the order of 1.2 times F2 (although it hs been twenty years).

Not sure why this is. Perhaps at rest there are small intermolecular bonds that form between atoms, that dont have time to form at speed?
When a tyre is rolling, the outside of that tyre is at rest relative to the road where it touches that road at the contact patch. So static friction applies. When the tyre locks from hard braking and starts sliding, then dynamic friction applies, reducing traction (the friction factor as the weight remains the same) of the tyre on the road, and hence braking force on the bike. Releasing the brakes will allow the periphery of the tyre to start rolling again, re-establishing static friction.

[Ocean1]

ABS strategy is to brake hard enough to lock up, detect lock up, release, and re brake hard in an effort to make average braking force higher than manually braking hard until lockup and relying on the rider having the presence of mind to pulse it manually and fast enough. There is no way for the system to detect "just pre lock up". What are you going use to sense that? Either there is wheel speed difference (i.e. locked up) or there is not.

Yes, I just said that. The fact that it's a viable control strategy doesn't mean it's not very effective.

According to my reckoning, this is why experienced racers can brake harder without ABS on track- they manage to keep the average braking force slightly higher by keeping the wheel on the edge of starting to slip by "feel" or experience - something sensors cannot do, at the point where the wheel is actually laying rubber down on teh road, but is still turning (so its not completely dynamic friction yet).

You might have been right a few years ago, but more recent results I've seen show new generations of ABS are both more effective than every unassisted rider tested and more effective than steady-state remote test rigs that maintain the wheel on the edge of starting to slip better than riders can. I've heard of a car version that actually modulates the pulse frequency depending on speed, to make best use of that peak in traction. Bike ABS is now detecting rear wheel lift and reducing front braking force to keep it down and weighted to the level that produces best combined brake force ratios.

Coefficients of friction. Slippery fuckers.

[R-Soul]

As per my previous statement (#443), the max grip can be found at a small percentage slip, wheel speed sensors can detect small amounts of slip. A wheel will take time to lock up once the grip force is overcome, with good a enough ABS package, this can be detected and the lock up prevented.

Slip is measured by detecting the difference, between the wheels (i.e. of one wheel relative to the other), of speed at the periphery of the wheels. The point is that there is already, during normal riding, significant difference in speed. This could be because of the driven wheel slipping under acceleration, the one wheel following a slightly longer radius around a corner, or just wheel wear, where one wheel wears faster than the other, which could make the system, which has been calibrated at a particular radius of the wheels, think there is more slip. Where actual slip due to lock up of one wheel comes in is really a matter of testing and where the engineeers set the threshold. So they say "given all the other things that could be occurring, normal slip could mean a difference in speed between teh wheels of up to X under normal reasonable conditions. We add in a buffer of safety (as we dont want ABs going off willy nilly), and therefore we set the absolute limit of where we think it is conclusive that there is lockup at Y% difference". Each manufacturer probably has different threshold settings according to what their own tests show.

High performance ABS just tries to set the threshold more accurately. And others, like Honda's C-ABS, have a few more add-ons.

Sounds like something was lost in translation there. High performance ABS can't just rely on lockup detection, slip will come into it as well.

Same thing. At least - both lockup and slip are measured by teh same thing - relative difference in wheel speed.

[Banditbandit]

This thread has ...




Turn off the computers - get the bikes out of the shed and just ride them ..

[bogan]

Slip is measured by detecting the difference, between the wheels (i.e. of one wheel relative to the other), of speed at the periphery of the wheels. The point is that there is already, during normal riding, significant difference in speed. This could be because of the driven wheel slipping under acceleration, the one wheel following a slightly longer radius around a corner, or just wheel wear, where one wheel wears faster than the other, which could make the system, which has been calibrated at a particular radius of the wheels, think there is more slip. Where actual slip due to lock up of one wheel comes in is really a matter of testing and where the engineeers set the threshold. So they say "given all the other things that could be occurring, normal slip could mean a difference in speed between teh wheels of up to X under normal reasonable conditions. We add in a buffer of safety (as we dont want ABs going off willy nilly), and therefore we set the absolute limit of where we think it is conclusive that there is lockup at Y% difference". Each manufacturer probably has different threshold settings according to what their own tests show.

High performance ABS just tries to set the threshold more accurately. And others, like Honda's C-ABS, have a few more add-ons.

Modern microcontrollers can do a lot with memory and modelling, the obvious options are; learning the slip ratios during normal riding, learning the slip ratios and adjusting accordingly during a lock up. The later is probably the way to go, you might have a very large slip to begin with, then pulse around the slip value giving the most grip.
A "buffer of safety" is perhaps a misnomer, as when ABS activates the safest option is to stop the bike as quick as possible, adding extra slip to make the slip easier to detect is more of a safety compromise.

Same thing. At least - both lockup and slip are measured by teh same thing - relative difference in wheel speed.

Lockup is binary, either its locked up, or its rotating, slip is detected across a continuous scale. These are most definitely not the same thing.

[imdying]

The Honda C-ABS is quite cunning, and does away with using the pressure generated at the master cylinder, and thus does away with the quick on/off/on/off you get with older systems, which is felt as a shudder by the rider. It's actually a brake by wire system, and as a result it uses a motor driven high pressure pump to supply the calipers, and that allows them to modulate the pump to vary the pressure. It used to be thought that on/off/on/off was a good thing... and it was for the level of tech/development, but then, as Bogan explains, tech moved on (primarily the manufacture of low cost high speed micro controllers) and this has enabled them to surpass that by actually remaining more in control of the brake pressures, and thus it lets them live on the ragged edge of the wheel speed/split/lockup equation.

There are some who expect that similar attributes will be applied to the EFI systems at some stage, allowing even finer granularity of the fuel injectors (and thus doing away with horrible multi-injector per cylinder setups), saving more money and decreasing emissions and fuel consumption. Things like this come very slowly though... I expect Honda's introduction of a brake by wire system went through pretty rigorous testing, and even then they had some problems which forced some recalls. As far as fuel metering goes, legislation has a lot to be thanked for the current crop of smoothly fueling EFI bikes (although on the other hand, it was also responsible for the mid 90s injection which was just a little premature for the tech on bikes).

[MrKiwi]

According to my reckoning, this is why experienced racers can brake harder without ABS on track- they manage to keep the average braking force slightly higher by keeping the wheel on the edge of starting to slip by "feel" or experience - something sensors cannot do, at the point where the wheel is actually laying rubber down on teh road, but is still turning (so its not completely dynamic friction yet).

Most of the riding public are not experienced racers (and I doubt even experienced racers can do what you claim), so when you need to brake hard and keep rolling ABS helps. It's not the saviour and can't counter for stupid, but it is something for riding on tarmac that I think riders should choose to have, in my opinion.

[Crasherfromwayback]

Most of the riding public are not experienced racers (and I doubt even experienced racers can do what you claim), so when you need to brake hard and keep rolling ABS helps. It's not the saviour and can't counter for stupid, but it is something for riding on tarmac that I think riders should choose to have, in my opinion.

That's the problem though. YOU think we should choose to have. I CHOOSE not to thanks. But people like you...will end up having our CHOICES taken away from us.

[imdying]

As the tech matures, that claim will go by the by. They will deliver, inevitably, greater more consistent performance than the best late brakers in the world can manage, and they'll do it day in day out with minimal maintenance. Give a racer so much as a crook tummy and you'll degrade their performance after 25 laps without a doubt... the computer will however march on tirelessly. What is needed is better quality data (sensors), the ability to analyse that data in more complicated ways (ECUs and their algorithms), and most importantly a financial or legislative motive. Good ABS systems are now at a point where they will suffice for everybody on the road, and it will not be long before the same applies to track riding.

[imdying]

That's the problem though. YOU think we should choose to have. I CHOOSE not to thanks. But people like you...will end up having our CHOICES taken away from us.

There will (probably quickly) come a point where it is moot... the system will provide unfettered access to the entire range of performance offered by the braking system, only activating when it really needs to (because they are always a little premature at the moment, and do not have the adjustability that you get with traction control systems).

[Crasherfromwayback]

There will (probably quickly) come a point where it is moot... the system will provide unfettered access to the entire range of performance offered by the braking system, only activating when it really needs to (because they are always a little premature at the moment, and do not have the adjustability that you get with traction control systems).

As long as I always have the option of locking a wheel or wheels up I don't give a shit what the future brings. I don't do fluro vests either.