Potential brake problem R1200RT

[klondike]

I met a fellow R12RT rider while in the black hills of South Dakota. When I pulled into the parking place by him you could definitly tell the bike had been down,left bag valve cover and front turn signal were damaged. When I parked he came up and looked at the front end and said it was wired different. He then said his front brake sensing wire had been wrapped on the inside of the fork leg and was close enough to the rotor in that it rubbed and wore through and shorted out. He went from no brakes to locked up brakes amd dropped it. He was from Alberta Canada. I didn't get to talk to him very long,don't know if he uses this site. My wire is well secured to the brake line and fork leg and on the outside of the caliper, but I could see that if was put behind the caliper it could definitely rub the rotor. It would be a good idea to check yours.

[steveknapp]

I think there was a recall for this exact problem.

[ShovelStrokeEd]

Something else at play there, a shorted sensor wire would certainly trigger a fault code and might even cause residual braking to become paramount, I just don't see it triggering locked wheel or wheels. A basic tenent of brake design would be that the system be incapable of a control output not initiated by a control input. I kinda doubt you got the full story there.

[steveknapp]

Ed, if the sensor was intermittent, and the pulse train either stopped or dropped out here or there, the ABS would engage preventing what it sees a lockup. The rider might grab a fistful of brakes as they are not stopping. Then when the sensor starts functioning again the ABS sees wheel speed, and re-enages the brakes, now full force.

 

Basically a more violent example of what happens over some bumps.

 

Not sure, but an intermittent speed sensor is a hard fault to diagnose in the controller. Espically for brakes.

[ShovelStrokeEd]

Nope, not buying it, ABS sees wheel speed and acts like ABS, as soon as the wheel locks it should start cycling the brakes. Given the number of teeth on the sensor wheel, I would have a hard time believing even an intermittent short could successfully mimic a moving wheel anyway. I would be really surprised if there weren't a pretty good noise filter on those wheel speed sensors anyway that only admitted pulses of a certain shape and amplitude. My controls guys tell me that isn't all that hard to program nor does it require much overhead.

 

I can see a wheel speed sensor fault disabling the ABS and with it, the servo controls. Given the difficulty of modulating the residual braking, the guy probably thought he had no brakes, and, in his panic, tripled the force at the lever, locked his front wheel and fell, keep in mind ABS is disabled as well. Far more likely than the servo suddenly outputting full brake force with no corresponding control input other than a sensor short.

[Joe_Rocket]

Agreed. There is absolutely NO chance that a failed sensor would cause wheel lockup. The computer will have a sensor check based on feedback from the sensor itself, current draw, voltage AND will be specifically designed to be fault tolerant of sensor short or open... guaranteed.

 

It is likely that the person who dropped their bike either didn't know what happened and was making things up along the way... or they knew exactly what happend and was lying about it in the hopes that a dealer and or BMW would buy the story.

 

JW

 

 

Nope, not buying it, ABS sees wheel speed and acts like ABS, as soon as the wheel locks it should start cycling the brakes. Given the number of teeth on the sensor wheel, I would have a hard time believing even an intermittent short could successfully mimic a moving wheel anyway. I would be really surprised if there weren't a pretty good noise filter on those wheel speed sensors anyway that only admitted pulses of a certain shape and amplitude. My controls guys tell me that isn't all that hard to program nor does it require much overhead.

 

I can see a wheel speed sensor fault disabling the ABS and with it, the servo controls. Given the difficulty of modulating the residual braking, the guy probably thought he had no brakes, and, in his panic, tripled the force at the lever, locked his front wheel and fell, keep in mind ABS is disabled as well. Far more likely than the servo suddenly outputting full brake force with no corresponding control input other than a sensor short.

[steveknapp]

I would be really surprised if there weren't a pretty good noise filter on those wheel speed sensors anyway that only admitted pulses of a certain shape and amplitude.

 

VR sensors, as used for wheel speed on all the bikes and engine posistion tracking in the hexhead, vary amplitude greatly depending on frequency.

 

The only constant is if you integrate the waveform out of the sensor over a given amount of travel (angle in the case of an tone ring or toothed wheel) the area under the curve at any speed will be nearly equal. Variance in sensor to ring distance as the wheel rotates will introduce some variance. And there are some losses in the sensor and wiring that will introduce some small variance, but the theory holds true.

 

It's easy to setup such a sensor for a fixed rate application. But much more difficult for something that must work from very low speeds, read low voltages, to high speed. Add in any electrical noise and it gets worse.

 

Another challenge is accurately placing the "edge." At low speeds, if you want a fast update rate, you might measure tooth-to-tooth, or a series of a few teeth. But your measurements can't drift, you need to measure from the same point on each tooth to be accurate. At low speeds the low frequency, low voltage waveform makes it hard to pick a good threshold that does not get tripped by noise but accurately places the edges at the same spot for each tooth. On the engine side you're looking at possibly having timing wander 1-2 degrees if not done right..

 

Also since the peak voltage varies with speed it creates issues not unlike a failing ignition component. Does the spark just quit? Or is it intermittent, firing at times, not at others? At some point when things fail enough it just quits.

 

Sorry if it's disjointed, but it's hard for me to remember which strategies for dealing with these buggers still have patent disclosures being prepared.

 

Far more likely than the servo suddenly outputting full brake force with no corresponding control input other than a sensor short.

 

Sure, but "full brake force with no coresponding control input" isn't what I wrote. Disagree with that all you want, but read my post again, it's NOT what I said.

 

Ever hit a bump while braking and have the ABS engage? That's more along the lines of the failure case I see.

[steveknapp]

The computer will have a sensor check based on feedback from the sensor itself, current draw, voltage AND will be specifically designed to be fault tolerant of sensor short or open... guaranteed.

 

Curious, but if your theory of diagnostics holds true, why does the ABS unit need to see some wheel speed (~3mph) each time you start the bike before turning off the brake failure lamp?

 

The reason is that it's much easier to check for a valid VR input once the bike rolls than it is to validate the sensor at a standstill. At a standstill it's just a coil of wire. The ABS unit applies no voltage to it. Nor is any current flowing until the sensor itself generates the signal.

 

Diagnostics for intermittent faults of the sensor while the sensor is operating is nearly impossible at a reasonable cost.

[PaulW42]

There is an incident documented on the UKGS website on a 1200GS where an incorrectly refitted sensor wire wore through and left the ABS permanently engaged and hence no (i.e. not even residual)brakes! So the opposite of the experience related here.

 

Check your wire routing folks.

 

Not sure which is worse.

 

 

Paul

[ShovelStrokeEd]

OK, I read your post again and see what you meant. I also read the OP's post about the story he got again. I still don't see my theory as wrong but can now see that yours may also be right. That is a really scary fault mode. NO BRAKES, YIKES!!

 

I can see what you meant, in light of the behavior of the sensors and the range of speed necessary and agree that integration would be the only way to 'filter' signal. Problem being, integration takes time. I was not aware that there was such a large amplitude variation. Shouldn't that have been converted into a square wave pulse train somewhere on the board? Getting a bit off topic, I know, but the why of things and the how have always fascinated me.

[Joe_Rocket]

Not knowing whether the sensor is hall-effect or not (in this particular application), I'm not sure of the parasitic nature of the sensor... I can however, tell you this:

 

1. There is some bias current or voltage which will tell you the sensor is alive; whether the bike is moving or not. When the bike isn't moving, you most likely have a steady voltage or current.

 

2. Having the sensor be alive and drawing the proper current DOES NOT mean the system is good. Keep in mind that something like a missing or damaged tone ring could cause a system failure even though the sensor is working, which would not be detected until the motorcycle was placed in motion.

 

I have to stick with my original analyis... that there is NO WAY for a sensor failure to cause the brakes to lock up.

 

I believe that someone mentioned that a sensor failure caused the brakes to fail completely (as in no brakes).... again; I say "no way". The argument (I guess) is that if the sensor fails in the "wheel is not turning therefore we are in a skid" mode, the ABS will permanently disconnect the brakes and your brake lever will simply bottom out with no brakes.

 

Utter hogwash.

 

If the ABS sensor fails in the above mode (assuming it even can), the ABS controller might release the brake for a short period BUT, it will very quickly see that even though it released the brakes, it is STILL not getting motion sensed from the sensor (The sensor is failed remember?)... so the computer says... I released the brakes but I didn't get a motion signal back from the sensor, therefore the sensor has failed and therefore I will go into residual braking mode.

 

Again... there is simply no way a sensor faiure will cause either full brakes or no brakes.

 

Joe West

 

The computer will have a sensor check based on feedback from the sensor itself, current draw, voltage AND will be specifically designed to be fault tolerant of sensor short or open... guaranteed.

 

Curious, but if your theory of diagnostics holds true, why does the ABS unit need to see some wheel speed (~3mph) each time you start the bike before turning off the brake failure lamp?

 

The reason is that it's much easier to check for a valid VR input once the bike rolls than it is to validate the sensor at a standstill. At a standstill it's just a coil of wire. The ABS unit applies no voltage to it. Nor is any current flowing until the sensor itself generates the signal.

 

Diagnostics for intermittent faults of the sensor while the sensor is operating is nearly impossible at a reasonable cost.

[steveknapp]

Problem being, integration takes time. I was not aware that there was such a large amplitude variation.

 

You get the same energy, but in a shorter timeframe. Hence the amplitude change. Low speeds are a real challenge.

 

The integration can be done real-time, and it's one approach. But for example, the peak voltage can be quite high, and may need to be clipped/limited...which is all good, except it throws off the integration.

 

If you are really curious, check out what you can find online, for example this ON semi part has a nice description of it's operation.

http://www.onsemi.com/pub/Collateral/AND8149-D.PDF

Or:

http://www.onsemi.com/pub/Collateral/NCV7001-D.PDF

 

Just reading that stuff makes my brain hurt.

 

Shouldn't that have been converted into a square wave pulse train somewhere on the board?

 

It is done in the controller, not in the sensor itself. And there are many ways to approach that conversion. GM and Ford each have their own.

[steveknapp]

Not knowing whether the sensor is hall-effect or not (in this particular application), I'm not sure of the parasitic nature of the sensor... I can however, tell you this:

 

For someone who isn't sure of the type of sensor or it's properties you seem very confident to say things are "impossible" or "hogwash" You don't even take the time to read what's posted here. Like the fact that it's a VR sensor, or that no one is saying "a sensor failure to cause the brakes to lock up".

 

Most all of your assumptions are based on the idea that the sensor fails hard, not intermittent. Which probably isn't the case when the sensor wiring is being rubbed against a moving brake rotor. Intermittent faults are the hardest to detect and deal with.

 

IF you make your assumption that it was a sudden hard failure, a nearby tree ripped the wire off the bike for example, then your statements hold true.

 

Consider the scenario instead where the computer has wheel speed data, but it's incorrect and somewhat random. In this case your proposed diagnostic algorithim won't function. And the behavor of the brakes would be unusual to say the least.

 

For what it's worth, my bud Drew's '96 Windstar has a bad front-left wheel speed sensor. On each stop as you drop below about 30mph the ABS engages on that wheel. The other three stop the van OK. Last I checked there was no ABS failure lamp lit. Either that or it burnt out!

 

There is some bias current or voltage which will tell you the sensor is alive; whether the bike is moving or not. When the bike isn't moving, you most likely have a steady voltage or current.

 

I posted two links for Ed. Read at least the first one, the section on diagnostics. The only bias current present is placed there by that IC expressly for the purpose of diagnostics. You'll note that the diagnostics have a special mode that must be engaged and from the looks of it can't be done on the fly.

[steveknapp]

My wire is well secured to the brake line and fork leg and on the outside of the caliper, but I could see that if was put behind the caliper it could definitely rub the rotor. It would be a good idea to check yours.

 

BTW, theory aside, checking this wiring is a good thing.

 

Or does anyone disgree with that?

[Joe_Rocket]

hehe... checked mine when I heard about the recall; there must ber 10 pounds of zip ties holding mine in place

 

JW

 

 

My wire is well secured to the brake line and fork leg and on the outside of the caliper, but I could see that if was put behind the caliper it could definitely rub the rotor. It would be a good idea to check yours.

 

BTW, theory aside, checking this wiring is a good thing.

 

Or does anyone disgree with that?

[PaulW42]

I believe that someone mentioned that a sensor failure caused the brakes to fail completely (as in no brakes).... again; I say "no way". The argument (I guess) is that if the sensor fails in the "wheel is not turning therefore we are in a skid" mode, the ABS will permanently disconnect the brakes and your brake lever will simply bottom out with no brakes.

 

Utter hogwash.

 

That would have been me. Well it may be hogwash but have a look at this thread:

 

http://www.ukgser.com/forums/showthread.php?t=51568

 

I know the author and he knows what he is talking about and is pretty levelheaded (well OK he takes his GS around the Nuburgring regularly but APART from that he is pretty level headed ).

 

Paul