[singhg5]

There were 3 MIL codes stored in my G5 - TPS, Temperature and Crank Position. However, when rollover sensor was turned over, there was no new code ! Only the previously stored codes were repeated !

Shouldn't new codes be displayed ?

https://www.youtube.com/watch?v=6I8ipAiZFGs

[Kevin Mahoney]

roll over sensor does not produce a code

[mattsz]

My B5 manual states that the "rollover sensor circuit malfunctioning" will produce a 1-long, 5-short MIL blink code, and that the engine will crank, but not start.

But, turning the sensor over, thus activating it and preventing the engine starting, isn't really a malfunction, is it?  So maybe sensor activation shouldn't throw a code...

[gashousegorilla]

No fault in the circuit or sensor, no code.  I bet if it was unplugged or one of those wires were snipped, you would get a code. When the bike stalled, the mil light did come on......

[singhg5]

No fault in the circuit or sensor, no code.  I bet if it was unplugged or one of those wires were snipped, you would get a code. When the bike stalled, the mil light did come on......

No.  There was no MIL light.

Service Manual states that for a Rollover Sensor circuit malfunction "MIL will glow continuous. Engine will NOT Start but will crank".

In my G5, half the statement is true and other half not so.

The idea of having MIL blinks is to quickly determine where is the source of engine stall or or not performing to its potential. Any circuit cut off - due to bad sensor or breakage of wires / connections - should produce a signal provided that sensor is sending/receiving input to/from ECU and is programmed to produce blinks. These MIL blinks are for detecting fault in circuit (incomplete) as well as sensor. (Essentially it is incomplete circuit when sensor is bad).

While riding if the rollover sensor screw comes loose and sensor tips over the bike will stall. How will the rider know where to look for the source of cut off ? MIL blinks would immediately point to the rider hey take off the seat and look for rollover sensor - something is wrong there ! That is just my thought. 

Anyone wants to confirm if your bike produces 1 Long and 5 Short blinks while the sensor is tipped over ?

[Craig McClure]

OH NO ANOTHER SWITCH!  I thought the Kill Button was for roll overs. Darn, is there anyway to deactivate this additional nanny state carbuncle?

[gashousegorilla]

No.  There was no MIL light.

Service Manual states that for a Rollover Sensor circuit malfunction "MIL will glow continuous. Engine will NOT Start but will crank".

In my G5, half the statement is true and other half not so.

The idea of having MIL blinks is to quickly determine where is the source of engine stall or or not performing to its potential. Any circuit cut off - due to bad sensor or breakage of wires / connections - should produce a signal provided that sensor is sending/receiving input to/from ECU and is programmed to produce blinks. These MIL blinks are for detecting fault in circuit (incomplete) as well as sensor. (Essentially it is incomplete circuit when sensor is bad).

While riding if the rollover sensor screw comes loose and sensor tips over the bike will stall. How will the rider know where to look for the source of cut off ? MIL blinks would immediately point to the rider hey take off the seat and look for rollover sensor - something is wrong there ! That is just my thought. 

Anyone wants to confirm if your bike produces 1 Long and 5 Short blinks while the sensor is tipped over ?

  Interesting Singh.  When you stalled the bike by turning over the sensor, The MIL light did not come on  before you reset the key?  And I agree with you , that it would be better to quickly narrow it down to one particular sensor in the event that is should loosen and pop off. Instead of having to look at all five other options in the book.  I have noticed with mine, if I have a Mil light on. When I hook up the wire to ground, and then turn the key back on, I wont get a Mil light glowing..... But it will go right to the codes. So I'm thinking, and you may have done this? ..... That with the sensor flipped over  and you tried to restart the bike in that state, Your bike would just crank and not start,and your Mil would glow.  Then you would need to go to the book, and go through the five options. 

[singhg5]

  .....

That with the sensor flipped over  and you tried to restart the bike in that state, Your bike would just crank and not start,and your Mil would glow. Then you would need to go to the book, and go through the five options. 

  BEFORE GROUNDING THE SINGLE POLE CONNECTOR

With the sensor flipped over and turning on ignition key - the MIL did NOT glow continuously. Instead MIL behaved exactly normal - came on for a few seconds for priming the pump and then turned itself OFF. Press the start button, the engine cranks and cranks but does NOT start.

It is shown in the video, may be you did not get a chance to see it fully.

In one sense the rollover sensor is sending information to ECU as if bike is on its side and stops engine from starting. HOWEVER, there was no clue from MIL light that a sensor is creating this problem and which one is it. The rider may check all other switches or go in wrong direction seeking solution. He could still miss looking at the real cause and go around in circles for a long time.     

Rollover Sensor Code in Service Manual

[Arizoni]

There were no "roll over switch" fault codes because the roll over switch is working fine.

The fault codes are created during the boot up interval while the MIL is glowing steady after the ignition key is turned on.  If the computer detects a faulty switch or wiring connection during this boot up period it stores the error.

The roll over switch shutting off the engine when the motorcycle tips over isn't a fault.  It's doing what it is designed to do.

As for the need of such a switch, IMO, if it was an old carburetor style machine it wouldn't make much difference but with a fuel injected machine something is needed to immediately shut off the fuel pump if the bike goes down.
If the fuel line was broken or torn loose and there was not some sort of safety switch the pump would merrily continue to pump fuel out all over the motorcycle and possibly the rider.  Not good if it catches fire.

[mattsz]

With the sensor flipped over and turning on ignition key - the MIL did NOT glow continuously. Instead MIL behaved exactly normal - came on for a few seconds for priming the pump and then turned itself OFF. Press the start button, the engine cranks and cranks but does NOT start.

...with a fuel injected machine something is needed to immediately shut off the fuel pump if the bike goes down.
If the fuel line was broken or torn loose and there was not some sort of safety switch the pump would merrily continue to pump fuel out all over the motorcycle and possibly the rider.  Not good if it catches fire.

Arizoni's comment makes sense, but Singhg5's actual test shows something a bit different: if the sensor is turned over simulating a dropped bike, why is the fuel pump still working (you can hear it prime up in the video)?  When the bike stalls due the the rollover sensor activating, is that merely killing the ignition or spark, or is it shutting off the fuel?  I wonder if the pump would continue to run if, in this condition, there was a leak?

[singhg5]

The fault codes are created during the boot up interval while the MIL is glowing steady after the ignition key is turned on.  If the computer detects a faulty switch or wiring connection during this boot up period it stores the error.

Codes represent a circuit malfunction, which includes wiring, connection, switches....  When that circuit breaks, it is considered a fault (irrespective of component responsible for circuit breakdown). During the boot up period the ciruit WAS broken with upside down sensor in my case or it could have been broken wires at that time or any other component.

The current does not flow if there is a missing switch, or broken wire or switch is turned off (as upside down rollover sensor would do by its design). The computer is only detecting flow of current or lack of current. The computer does not know why current does not flow. The computer does not know if the switch is turned off or a wire is disconnected.
 
When the rollover switch is upside down, the current stops flowing and ECU knows that rollover sensor circuit has some problem and it stops engine in its tracks. So the ECU is behaving the way it is programmed to do.

The codes are there to point the rider in right direction so that he can find out which component caused the circuit breakdown. 

Do you want to do a test on your bike and let me know what you see ?

[singhg5]

Arizoni's comment makes sense, but Singhg5's actual test shows something a bit different: if the sensor is turned over simulating a dropped bike, why is the fuel pump still working (you can hear it prime up in the video)? 

When the bike stalls due the the rollover sensor activating, is that merely killing the ignition or spark, or is it shutting off the fuel? I wonder if the pump would continue to run if, in this condition, there was a leak?

You are right that in my G5 the fuel pump kept working BUT ignition was cut off.

However the Service Manual says that rollover sensor circuit malfunction will cut off fuel supply to injector AND the ignition.

Perhaps something may be amiss in my bike ! 

Only one thing worked - the bike did NOT start.

[Royalista]

On a side note: should the rollover sensor switch off immediately as Arizoni stated?

When I was in doubt whether mine was functioning I did some tests. It works all right, but it takes a noticeable amount of time (app. 20-30 sec) before it turns the engine off.

[singhg5]

On a side note: should the rollover sensor switch off immediately as Arizoni stated?

When I was in doubt whether mine was functioning I did some tests. It works all right, but it takes a noticeable amount of time (app. 20-30 sec) before it turns the engine off.

It took a few seconds to turn off engine.

Any details of what you did and what you saw.

What prompted you to do the test? Did you check with ignition key on and sensor tipped enough to stall engine if MIL light glow continuously at that moment ? 

[gashousegorilla]

  BEFORE GROUNDING THE SINGLE POLE CONNECTOR

With the sensor flipped over and turning on ignition key - the MIL did NOT glow continuously. Instead MIL behaved exactly normal - came on for a few seconds for priming the pump and then turned itself OFF. Press the start button, the engine cranks and cranks but does NOT start.

It is shown in the video, may be you did not get a chance to see it fully.

In one sense the rollover sensor is sending information to ECU as if bike is on its side and stops engine from starting. HOWEVER, there was no clue from MIL light that a sensor is creating this problem and which one is it. The rider may check all other switches or go in wrong direction seeking solution. He could still miss looking at the real cause and go around in circles for a long time.     

Rollover Sensor Code in Service Manual

   No , I did not see that in the video. I just checked again.  What I see is your bike stalling when you turn the sensor upside down. Then you hook up the wire, then I see the key off and you trying to start the bike.  My point was, or what I was wondering was..... before you hooked up the wire did the Mil light come on ?  When the bike stalled ?  AND before grounding out the wire with the key on ? And stay on ? And again... I agree it should be narrowed down and get a fault code when flipped. BUT, I'm thinking..... That is not how it is set up. It is not a two wire switch, where the contacts would break , and you would have an open circuit triggering a code.  It is a three wire sensor, so there may be a path.....

 If the bike goes down, it sound like it cuts the ignition.  So the bike stalls and the fuel pump shuts down.  No ignition, no rotation at the rotor seen by the pulsar coil as it stops.... No pump.  Unless one tries to start the bike laying on the ground with a ruptured tank ? The bikes still wound not start...and the pump would only run, while someone was pushing in the start button. But like you say, the book says otherwise ?  So something is wrong with the book or your bike?  My bet it's the book....  I'll see what mine does.

[singhg5]

GHG:

Go to 4:37 of the video, you will see that MIL behaves as if normal, even though sensor is upside down. This part of the test was done completely separately at different time than rest of video prior to that time. Between 4:37 to end, no ground wire was used. 

Think of it that film is actually starting from 4:37 and nothing happened prior to that. Though in video due to editing it appears as if everything was done together in sequence.

Interesting that you pointed out there are 3 wires on sensor. I had not looked at it that carefully, was too busy filming and counting codes  .  If it is not a ON/OFF switch it may be an angular sensor like a gyroscope - electronic micro gyroscope. There could be a power supply wire, a ground wire and an OUTPUT wire from the sensor that sends signal to ECU. The sensor is actually calculating angular rate several times during operation and has to reach certain angle to significantly alter the output signal. And its algorithm has to make a decision whether it is still in 'safe' angle or moving towards 'unsafe' angle. That could explain why it takes a few seconds for sensor to turn off engine.

There may be output signal detection limits that indicate to ECU whether sensor is upright or tipped.

If that is what is happening, then the bike can be set up to not create a fault code when the sensor is tipped because ECU is still getting a signal from output wire. It can also be set up to create a code when sensor output is outside the limits of upright position, which can be helpful for the rider if sensor comes off frame and tips over.

Now if someone's bike cranks but does not start and there is no continuous MIL glow AND they have checked all other possible known causes, it is time to add one more place to look - take off seat and check for loose rollover sensor screw !

[gashousegorilla]

GHG:

Go to 4:37 of the video, you will see that MIL behaves as if normal, even though sensor is upside down. This part of the test was done completely separately at different time than rest of video prior to that time. Between 4:37 to end, no ground wire was used. 

Think of it that film is actually starting from 4:37 and nothing happened prior to that. Though in video due to editing it appears as if everything was done together in sequence.

Interesting that you pointed out there are 3 wires on sensor. I had not looked at it that carefully, was too busy filming and counting codes  .  If it is not a ON/OFF switch it may be an angular sensor like a gyroscope - electronic micro gyroscope. There could be a power supply wire, a ground wire and an OUTPUT wire from the sensor that sends signal to ECU. The sensor is actually calculating angular velocity several times during operation and has to reach certain angle to significantly alter the output signal. That could explain why it takes a few seconds for sensor to turn off engine.

There may be output signal detection limits that indicate to ECU whether sensor is upright or tipped.

 OH !!   Got ya     Yes I see what you mean.  Ummmmmm.....  I know that on some sensors that can detect a "tilt" , there is a small ball bearing in a tube, so when moved or jostled, the ball rolls..... Wondering if this is similar ?   I'll see what mine does....

[Craig McClure]

Does anyone know how to disable this switch?  I don't want it!

[Royalista]

It took a few seconds to turn off engine.

Any details of what you did and what you saw.

What prompted you to do the test? Did you check with ignition key on and sensor tipped enough to stall engine if MIL light glow continuously at that moment ?

Happily exploring a forest road it grew adventurous by the bend. After passing a point of no return it became mud, muddier, then very slippery. The bike got down once, then we both got down, and finally we got down real good. Fuel leaked out the filler cap, but the engine didn't stop until i hit the kill switch.

So I thought: what's with the roll-over sensor?

First test: tilt the bike 60° to left, to right. Nothing happens. Sensor bad?
Second test: lay a tire on the floor and put bike down on the tire. Now the engine did stop, after some time. I have not timed the delay, but it was considerably long. Did this test left and right with equal result.

Conclusion: sensor works but it takes some time to have effect.
Does it work by starving engine from fuel?

I have not noticed, nor checked, whether the MIL light did come on while laying down.
After uprighting the bike and putting it on the center stand the engine was fired up in the normal way. There was no warning from MIL nor is their an error code stored (hence roll-over sensor circuit considered in order).

[mattsz]

Now if someone's bike cranks but does not start and there is no continuous MIL glow AND they have checked all other possible known causes, it is time to add one more place to look - take off seat and check for loose rollover sensor screw !

I have found that screw loose twice in 2 years and 5000 miles.  Because it's a rubber mount, I haven't cranked down on it like I would on a screw I really didn't want to come undone...

[gashousegorilla]

 I tried mine and I got the exact same results as Singh and Royalista..... NO MIL  light.  I found that the bike did cut out at about half way around or so, and relatively quickly?  It sounded like after the ignition was cut, it just finished up a cycle or so.  So if the bike goes down, it will kill it. And picking it back up will automatically reset it.  And I do hear SOMETHING rattling around inside the sensor. So I guess that tells me , that whatever is in there is free to move. And perhaps make and brake the needed contact.

 However ...... I now have a new code stored in my collection,    At the end of all the others. I now have a 1-5 code.  Unplugging the sensor will give a MIL light, and store a code.  With the sensor removed , it will cut ignition and fuel... no pump.  I didn't want to snip a wire, sooo it was the next best thing to simulating a fault in the circuit.    


 So it does work, a little different then the book. But it does work... and pretty well in my case.  And self resetting.  BUT, I can see where maybe ?.... If  say the sensor were to fall off it's mount, you may be a little confussed for a bit, until you saw it hanging there ? But as Singh points out, IF the bike were fall over or the sensor were to fall off AND whatever is in the sensor was to not reset.... You may pull your hair out looking for the cause.  Shaking the sensor, and listening for movement MAY help figure out whats going on? No movement, or rattle MAY be an indicator of a stuck sensor.....

[Royalista]

However ...... I now have a new code stored in my collection,    At the end of all the others. I now have a 1-5 code.  Unplugging the sensor will give a MIL light, and store a code.  With the sensor removed , it will cut ignition and fuel... no pump.  I didn't want to snip a wire, sooo it was the next best thing to simulating a fault in the circuit.    

If I get this right that would mean no more riding, yes?
I find that a little upsetting.
I hope I have that wrong.

[gashousegorilla]

  No...NO..... no  All is fine !  The code is just stored. And once you plug the sensor back in, your good.   Bike starts, no more MIL light, no worries.

[singhg5]

I tried mine and I got the exact same results as Singh and Royalista..... NO MIL  light.  I found that the bike did cut out at about half way around or so, and relatively quickly?  It sounded like after the ignition was cut, it just finished up a cycle or so.  So if the bike goes down, it will kill it. And picking it back up will automatically reset it.  And I do hear SOMETHING rattling around inside the sensor. So I guess that tells me , that whatever is in there is free to move. And perhaps make and brake the needed contact.

 However ...... I now have a new code stored in my collection,    At the end of all the others. I now have a 1-5 code.  Unplugging the sensor will give a MIL light, and store a code.  With the sensor removed , it will cut ignition and fuel... no pump.  I didn't want to snip a wire, sooo it was the next best thing to simulating a fault in the circuit.    


 So it does work, a little different then the book. But it does work... and pretty well in my case.  And self resetting.  BUT, I can see where maybe ?.... If  say the sensor were to fall off it's mount, you may be a little confussed for a bit, until you saw it hanging there ? But as Singh points out, IF the bike were fall over or the sensor were to fall off AND whatever is in the sensor was to not reset.... You may pull your hair out looking for the cause.  Shaking the sensor, and listening for movement MAY help figure out whats going on? No movement, or rattle MAY be an indicator of a stuck sensor.....

FINALLY ...  we are getting somewhere !
 
I guess you had unplugged the wire from sensor to ECU and that gave you a code ?

I have been trying to understand the construction of rollover sensors and how they work. Not much information easily available - companies want to keep that to their chest. So far it seems there are at least two different designs (may be even more). One of them is like a Tuning-Fork design that oscillates and creates current. The other is perhaps a disc type - details are sketchy and hard to get it. There is a change in the current as the sensor is tipped because of the kind of material used to make them, which includes certain ceramic or crystal materials of such elements as LiNbO3, Lithium niobium salts or LiTaO3, Lithium tantalum salt or some trade secret stuff.

Good going GHG.

PS - This is NOT an ON/OFF switch and we can at least figure that much out  .

[gashousegorilla]

  You know I'm always down for a good experiment     And yes Singh, I got the code because the sensor was unplugged WITH the key on.  You can unplug it to your hearts content , AS LONG as you don't turn on the key.... Or you will get a code. Unplugging it WILL NOT defeat it.. IMHO ,that is a good thing! And how  MOST safeties work    Nothing wrong with the bike shutting down should you drop it !  I have NO Idea exactly WHAT is in there, and I agree about them keeping such things close. BUT..... I have seen the little ball bearing somewhere 

[Royalista]

So, if the sensor malfunctions or if the circuit is broken somewhere and it cannot be repaired at the spot, then it's over with riding?

[gashousegorilla]

So, if the sensor malfunctions or if the circuit is broken somewhere and it cannot be repaired at the spot, then it's over with riding?

 YES ..... IF there was an issue with the circuit,  You would get a code.  NO.... you would not get a code if the sensor does not reset.  Either way...  riding is over if you cant fix it on the spot. And you would not KNOW if the sensor was stuck... because you would not get a code. And that is IF that sensor can even get stuck ?   BIG IF ....

[Bulletman]

I have found that screw loose twice in 2 years and 5000 miles.  Because it's a rubber mount, I haven't cranked down on it like I would on a screw I really didn't want to come undone...

Same case here, I've found the sensor almost fallen over, just snagged up under the seat, otherwise I just tightened it up and left it...it's hasn't budged since...

[Kevin Mahoney]

If you put the bike down and the sensor activates, you have to buy a new one at a cost of $350 and it can only be installed by a dealer...just kidding.
You put the bike back upright. turn the key off and then cycle it on and go on your merry way.

[ROVERMAN]

Nice one Kevin, now maybe we can get this mountain of a thread back to a molehill and move on.
Roverman.

[Craig McClure]

If you put the bike down and the sensor activates, you have to buy a new one at a cost of $350 and it can only be installed by a dealer...just kidding.
You put the bike back upright. turn the key off and then cycle it on and go on your merry way.

Can the Switch be disabled or removed? If I fell outrunning a Rhinoceros, it would be a lot more convenient not to have to wait around. How can I get rid of this Switch? (I am over 21, & bike is out of warrantee) Thanks,

[Kevin Mahoney]

No comment, If it were me I would also install a deadmans switch. That petrol pumps puts out 40 lbs of pressure. Sort of like a small firehose squirting gasoline. MIght be an interesting science experiment though! (kidding, do not sue me, do not do this at home)

[singhg5]

GHG:

Did the same as you - disconnected the rollover sensor and got same results. MIL displayed the previously stored 3 older codes and then the LAST code was 1-Long and 5-short which represents rollover sensor.

This means a person will have to go through all the older codes before he sees the latest code - when trying to find out why the bike is not starting at the moment. 

I then proceeded further to get some more information. Once it is taken off camera, I will post it here. More discussion to follow .

[Arizoni]

Well, its nice that something good came out of this study.

We now know, the last trouble code is the latest one.

That could be useful for the motorcycles that have several old codes hanging around in the memory.

[gashousegorilla]

 OH !  .... I have done the experiments .  You most definitely want it to shut off !  Like it was mentioned before... There ain't no float bowl on a throttle body.  The bike would probably run upside down, without something to kill the ignition and fuel !   Resets immediately... no waiting needed.

GHG:

Did the same as you - disconnected the rollover sensor and got same results. MIL displayed the previously stored 3 older codes and then the LAST code was 1-Long and 5-short which represents rollover sensor.

This means a person will have to go through all the older codes before he sees the latest code - when trying to find out why the bike is not starting at the moment. 

I then proceeded further to get some more information. Once it is taken off camera, I will post it here. More discussion to follow .

  Very cool Singh     Looking foward to what you have found....  It's a good thing to look deaper into these things. So maybe it helps someone down the line ?  You know... Like all them Video's maybe? 


 

[mattsz]

No comment, If it were me I would also install a deadmans switch. That petrol pumps puts out 40 lbs of pressure. Sort of like a small firehose squirting gasoline. MIght be an interesting science experiment though! (kidding, do not sue me, do not do this at home)

But isn't Singhg5's fuel pump priming when his sensor is upside down?  I would think it shouldn't operate at all...

[singhg5]

The operating voltage of rollover sensor is 5V - just like throttle position sensor. I think they both have common source of power.

The OUTPUT signal voltage of rollover sensor is 0.6V at normal upright riding position. It increases to 3.9V when tipped.

I had expected it to be a real hi-tech electronic device with microelectronics and elaborate angular calculations / algorithms that compute mega data in seconds on the chip.

But it appears to be a simple device - TRUE to its RE heritage - simple inexpensive something that does a job . Indian 'jugaad'. It is a 'narrow band' sensor, that alters output signal voltage.

What can I say, it is a 'mule' - my bike and also the sensor - kind of switch kind of sensor !!

BTW - GHG for you   http://www.youtube.com/watch?v=VITrkCT3hKM

[Royalista]

Thanks for the good research and the  pictures.

[gashousegorilla]

The operating voltage of rollover sensor is 5V - just like throttle position sensor. I think they both have common source of power.

The OUTPUT signal voltage of rollover sensor is 0.6V at normal upright riding position. It increases to 3.9V when tipped.

I had expected it to be a real hi-tech electronic device with microelectronics and elaborate angular calculations / algorithms that compute mega data in seconds on the chip.

But it appears to be a simple device - TRUE to its RE heritage - simple inexpensive something that does a job . Indian 'jugaad'. It is a 'narrow band' sensor, that alters output signal voltage.

What can I say, it is a 'mule' - my bike and also the sensor - kind of switch kind of sensor !!

BTW - GHG for you   http://www.youtube.com/watch?v=VITrkCT3hKM

  HAHAHA !!  No... no complicated calculus or Chinese arithmetic needed !  Excellent work Singh !    See that.... Now we all know what the readings should be if we suspect a problem with the rollover sensor. Outstanding !  No Voltage change when flipped.... bad sensor.  Pretty cool. What do you think Singh ?  Some kind of variable resister in there , Like the map sensor ?  Like I said, I could hear movement of some sort in there, did you hear it as well ?

[Royalista]

I hear that too. Kind like a ball rolling, I thought it was actuating a switch.
Is the increase in voltage linear or direct?

[singhg5]

  HAHAHA !!  No... no complicated calculus or Chinese arithmetic needed !  Excellent work Singh !    See that.... Now we all know what the readings should be if we suspect a problem with the rollover sensor. Outstanding !  No Voltage change when flipped.... bad sensor.  Pretty cool. What do you think Singh ?  Some kind of variable resister in there , Like the map sensor ?  Like I said, I could hear movement of some sort in there, did you hear it as well ?

I hear that too. Kind like a ball rolling, I thought it was actuating a switch.
Is the increase in voltage linear or direct?

It did not show linear change in voltage proportional to change in angle. Instead it jumped from one reading to another - from lo voltage to high voltage and vice-a-versa. It only senses two positions - like the narrow band oxygen sensor on RE.

On the other hand TPS and MAP sensors respond progressively with change in throttle position and air pressure in manifold, respectively, producing linear response curves.
 
There were construction crew noises around me that drowned any low sound and I was turning it slowly to see gradual angular changes in voltage so did not hear rattle inside.  But from everything we know and from your observations, there has to be something moving from one side to another that alters its output voltage.           

PS - Anything RE, it has to rattle !

[Royalista]

Aye, aye, sir.
 

[gashousegorilla]

 I'm thinking it could be something like this.  Very low pitched rattle when shaken right next to your ear.

http://img.alibaba.com/img/pb/100/556/651/651556100_529.jpg


 If it is something like that, it WOULD be very simple and reliable.  Similar to those old round Honeywell thermostats, with the mercury "ball switch"..... Simple and last forever.

[cstorckiii]

If only I had paid attention to this thread last week, I wouldn't have found myself in the parking lot at work, sitting on the ground next to my bike, with the tank in my lap trying to check the fuel pump screen for paint blockage. Not at all embarrassing.

But, thanks guys. Good stuff.