2005 USDM STi yaw sensor answer...

[dayofpain]

its like this

yaw = pitch in relation to the horizon.

g's = measurement in "gravitys" of inertia in any direction.

2 different sensors measuring different things.

a car moving faster pitches into corners at different g's and differnt yaw rate as compared to a car moving at a slower rate of speed. also changes in suspension change things differently as well. now have two "almost" redundant sensors makes things awesome as one backs up the other and so forth. the two sensors cooerlate and decide which torque split to give the wheels. these only work in "auto" mode. in any other mode both sensors are defeated. two sensors being better because well lets face it, we all know more is better right? we are americans. ;P

maybe clears things up.

[sperry]

Originally Posted by dayofpain

its like this

yaw = pitch in relation to the horizon.

g's = measurement in "gravitys" of inertia in any direction.

2 different sensors measuring different things.

a car moving faster pitches into corners at different g's and differnt yaw rate as compared to a car moving at a slower rate of speed. also changes in suspension change things differently as well. now have two "almost" redundant sensors makes things awesome as one backs up the other and so forth. the two sensors cooerlate and decide which torque split to give the wheels. these only work in "auto" mode. in any other mode both sensors are defeated. two sensors being better because well lets face it, we all know more is better right? we are americans. ;P

maybe clears things up.

Your description of yaw and G's are wrong, but you're right about there being a difference between latteral G force and yaw rate.

If you imagine a 3-axis system where X points down the nose of the car, Y point out the passenger door, and Z point up through the roof... yaw is defined as rotation about the Z axis. Pitch would be rotation about the Y axis and roll would be rotation about the X axis.

Latteral G Force is force, not inertia. Well, technically it's a "pseduoforce" that's generated by the car traveling around a circle during cornering. The force's direction is towards the center of the circle the car is traveling around, and it's magnitude is what we call "latteral G's", and it's measured in multiples of the force of gravity on Earth (aka 9.8m/ss).

The DCCD uses lateral G to decide how to bias the torque between the front and rear wheels in Auto mode. The VCD (aka traction control) uses yaw rate to determine what wheels to brake to keep the car from spinning out or pushing.

So yes, two sensors are potentially better, but in this case they're actually being used for different purposes... one's for torque split, the other's for traction control.

[tysonK]

"let the speduoforce be with you."

[sperry]

Originally Posted by tysonK

"let the speduoforce be with you."

...fixed it.

[dayofpain]

Originally Posted by sperry

Your description of yaw and G's are wrong, but you're right about there being a difference between latteral G force and yaw rate.

If you imagine a 3-axis system where X points down the nose of the car, Y point out the passenger door, and Z point up through the roof... yaw is defined as rotation about the Z axis. Pitch would be rotation about the Y axis and roll would be rotation about the X axis.

Latteral G Force is force, not inertia. Well, technically it's a "pseduoforce" that's generated by the car traveling around a circle during cornering. The force's direction is towards the center of the circle the car is traveling around, and it's magnitude is what we call "latteral G's", and it's measured in multiples of the force of gravity on Earth (aka 9.8m/ss).

The DCCD uses lateral G to decide how to bias the torque between the front and rear wheels in Auto mode. The VCD (aka traction control) uses yaw rate to determine what wheels to brake to keep the car from spinning out or pushing.

So yes, two sensors are potentially better, but in this case they're actually being used for different purposes... one's for torque split, the other's for traction control.

ok so aparently we are going to get really technical, i forogt how retarded online forums can be.

your description is just as fallable as mine.

there are two sensors in the car

1. measures the inertial force IE: g-force (www.websters.com) of the car

2. measures the "direction the car is tipped" or the "degree" in relation to its horizontal axis.

the sensors correlate between each other and make the decision of which tires get torque and which do not.

lastly, i thought this thread was about how noone knew what they did, you posted the correct answer right off the bat.

i was just trying to help.

ALSO as point of reference. car makers tend to call shit anything they please whether its correct or not. IE: yaw sensor not measuring degrees in relation to vertical axiss but measuring degress against horizontal axiss. cause they think they are cool and innovative. they make up terms and use terms INCORRECTLY all the time. i took one such sensor apart and found it couldnt possibly measure the degree relation in which it was named for. due to its inner workings it could only measure based off of the horizon. UNLESS my sensor was mounted incorrectly (impossible due to bracket shape).

[sperry]

Originally Posted by dayofpain

ok so aparently we are going to get really technical, i forogt how retarded online forums can be.

your description is just as fallable as mine.

there are two sensors in the car

1. measures the inertial force IE: g-force (www.websters.com) of the car

2. measures the "direction the car is tipped" or the "degree" in relation to its horizontal axis.

the sensors correlate between each other and make the decision of which tires get torque and which do not.

lastly, i thought this thread was about how noone knew what they did, you posted the correct answer right off the bat.

i was just trying to help.

ALSO as point of reference. car makers tend to call shit anything they please whether its correct or not. IE: yaw sensor not measuring degrees in relation to vertical axiss but measuring degress against horizontal axiss. cause they think they are cool and innovative. they make up terms and use terms INCORRECTLY all the time. i took one such sensor apart and found it couldnt possibly measure the degree relation in which it was named for. due to its inner workings it could only measure based off of the horizon. UNLESS my sensor was mounted incorrectly (impossible due to bracket shape).

Sorry, not trying to be retarded, just precise.

First, I just did a search for "inertial force". I've never heard the term before, but it turns out that it's a synonym to pseudoforce. Basically a perceived force that's a result of inertia (the tendancy of an object at rest to stay at rest). For example: centripital force isn't a real force it's a pseudoforce that you feel because you're moving in a straight line but the car is pushing on you as is goes around a turn, and that is indeed the force that's measured by a latteral G sensor.

However your 1st statement was:

"g's = measurement in "gravitys" of inertia in any direction."

The unit of measurement for inertia is actually the same unit as mass, which is either grams or pounds. Not's G's. G's are actually a measurement of acceleration (1G = 9.8m/ss), and in the case of "G-Force" we factor out the mass, and consider it a unit of Force, which is normally measured in Netwons or pounds-force. Your statment confused me, since you actually ment "inertial force" which was a term I hadn't heard before.

As far as yaw... yaw has absolutely nothing to do with deviation from the horizon. Pitch and roll can be deviations from the horizon, but yaw is rotation about the verticle Z axis. Attemping to measure pitch or roll in conjunction with a system designed to keep you from spinning or pushing wouldn't work. Imagine your right side brakes coming on every time you drove down a hill, or hit a pot hole! Perhaps the reason the yaw sensor you took apart looked wrong is because you're thinking of roll and not yaw.

This is yaw:



Also, I don't know how old the sensor was that you took apart, but all the sensors used in cars these days would be solid state, and look simply like computer chips. Mechanical "yaw sensors" that you might be able to look at and figure out what they're measuring are called gyroscopes and can be found in older military aircraft, which due to their cost would never show up in a production vehicle.

The way a yaw sensor would be able to tell if your car is plowing of spinning is by comparing the wheel speed sensors (part of the ABS system) to the current yaw rate from the yaw sensor. In a stable turn there is a known speed difference between the inside and outside wheels. At a certain wheel speeds, if the yaw rate is too high for that turn the computer can deduce that the car is oversteering, and apply the brakes to the outside wheels, thus generating a yaw moment to counter the spin. If the yaw rate is too low, the car computer can deduce that the car is pushing and apply the brakes to the inside wheels and generate a yaw moment that will help the car turn.

At no point is there any torque split occuring. The EVO's AYC uses torque split left to right to control yaw instead of brakes, but in the STi it's all part of the ABS system... the DCCD and the latteral G sensor aren't used at all. That's a completely seperate system designed to balance the control aspects of a RWD car with the traction advantage of AWD.

[dustinr]

Well said !

[dustinr]

Well I just received my Subaru Performance mag Version 1.2 and it has a good article about the DCCD and what it does etc. In an earlier post in this thread I said that I thought the car's Auto DCCD transferred torque forward during hard cornering so as to pull the car out of the corner and both Ban Suv's and Sperry said that it did not, that if anything it moved the torque backwards to full unlock 35/65. Not knowing for sure I didn't argue the point. Well in the article they explain that it's full unlocked for braking and that it gradually increase the torque forward thru a corner till it goes full lock 50/50 at corner exit so as to pull the car out of the turn and counter-act any oversteer or wheel-slip. The article even has a cool 3-D graph which shows what the DCCD does given later-G's, throttle position, and corner position. And they reiterate the fact that the '05 has a Yaw-Rate sensor.

[sperry]

Originally Posted by dustinr

Well I just received my Subaru Performance mag Version 1.2 and it has a good article about the DCCD and what it does etc. In an earlier post in this thread I said that I thought the car's Auto DCCD transferred torque forward during hard cornering so as to pull the car out of the corner and both Ban Suv's and Sperry said that it did not, that if anything it moved the torque backwards to full unlock 35/65. Not knowing for sure I didn't argue the point. Well in the article they explain that it's full unlocked for braking and that it gradually increase the torque forward thru a corner till it goes full lock 50/50 at corner exit so as to pull the car out of the turn and counter-act any oversteer or wheel-slip. The article even has a cool 3-D graph which shows what the DCCD does given later-G's, throttle position, and corner position. And they reiterate the fact that the '05 has a Yaw-Rate sensor.

Are we still arguing this?

First, the DCCD *never*, and I mean *never*, changes the torque split of the center diff. It is always 35f/65r (I was under the impression initially that it was 50/50, but 35/65 seems to be the correct split). That split is defined by the physical characteristics of the differential. Unless the DCCD is somehow swapping out internal parts in the diff during a turn, the torque split is fixed.

What the DCCD changes is the locking bias. (Yes, I've explained this before):

Originally Posted by sperry

The power/torque from the motor is fixed to the wheels (I thought it was 50/50 but I guess it's actually 35/65).

What's varied is the amount of speed difference is necessary before the center diff locks...

When the DCCD is set to full forward mode, the center diff is locked, i.e. the front and rear wheels have to turn at the same speed... so while the rears are getting 65% of the power, they are still locked with the fronts. If you jump on the gas you'll break all 4 loose.

When the DCCD is set to full rear, the center diff requires a much higher speed at the rears before they lock to the fronts. So those rear tires are getting 65% of the power and are free to spin wildly while the fronts grip and pull with their 35% of the power. At full rear, the car behaves more like a RWD car because jumping on the throttle will break the rears loose w/o losing the front traction.

Auto mode uses a lateral G sensor to decide where to set the split. I doubt it will ever go all the way forward, because a locket center diff is only usefull on gravel or snow or in the wet, otherwise the wheels will bind up. My guess is that the diff sits at middle ground normally, but will throw the bias rearward under high latteral G's because having the rears uncoupled lets throttle modulation turn the car... i.e. auto mode helps the car to turn more like a RWD car, but gives you additional traction when driving in a straight line.

To reitterate, the whole point of the DCCD (in Auto Mode) is to blend the advantages of AWD with the control of RWD. If you've ever driven at the limit, you know that "throttle steering" is very useful to control the car. AWD inherrently counter-acts this ability. The DCCD is designed so that in the corner, the car throttle steers, but in a straight line (i.e. when exiting the corner) the locking bias of the center diff moves towards 50/50 meaning that rear wheel spin translates into front wheel traction more easily, helping the car to pull out of the corner (yet limiting the car's ability to turn, which is less necessary on exit).

As far as the whole "shifting bias to the rear during cornering"... that *is* what's happening. According to the chart, when you brake and turn in the car has shifted to full rear bias. As you apply throttle for the 1st half of the corner, the bias stays rear-oriented to give the driver a RWD-like throttle steer ability. Once the latteral-G's begin to unload (signifying that the car is through the apex of the turn and is accelerating out to the exit), the bias is ramped forward, adding traction.

Getting back to the yaw sensor. I'm still missing what it's used for that's somehow different from the . It's only ever mentioned in conjunction with the latteral-G sensor. I guess that yaw sensor data combined w/ latteral-Gs might result in more accurate DCCD mdulation, but the addition of a yaw sensor doens't make any fundamental difference in the way the DCCD works. In fact, I would propose that a good driver that's skilled at managing their traction circle would gain no advantage from the addition of a yaw sensor, since the car's yaw and latteral-G's should always match because the driver's not tossing the car around and drifting, thereby making the latteral-G sensor alone sufficient.

I'm much more interested in hearing about the yaw control built into the ABS system. I've seen the tech articles and I understand how it's supposed to work, but is it actually in the '05 STis? I haven't seen any marketing literature that mentions the STi having "traction control" or "yaw control"... you'd think that stuff would get more press.

[Kevin M]

Let's set aside this argument over the nuts and bolts of DCCD operation and discuss why it does what it does. In non-DCCD, OE Subaru AWD systems, you have 50/50 torque split all the time. The biggest disadvantage of that is the added workload of the front tires when turning and accelerating reduces available grip. By moving more of thetorque rearwards, there is more grip available up front. The reason for the locking characteristics of the DCCD are what sperry just posted- it's to use wheelspin in the rear to force the front tires to carry more acceleration load as the lateral load is reduced as you unwind the wheel coming out of the turn.

[dustinr]

Originally Posted by sperry

Are we still arguing this?

First, the DCCD *never*, and I mean *never*, changes the torque split of the center diff. It is always 35f/65r (I was under the impression initially that it was 50/50, but 35/65 seems to be the correct split). That split is defined by the physical characteristics of the differential. Unless the DCCD is somehow swapping out internal parts in the diff during a turn, the torque split is fixed.

What the DCCD changes is the locking bias. (Yes, I've explained this before):

To reitterate, the whole point of the DCCD (in Auto Mode) is to blend the advantages of AWD with the control of RWD. If you've ever driven at the limit, you know that "throttle steering" is very useful to control the car. AWD inherrently counter-acts this ability. The DCCD is designed so that in the corner, the car throttle steers, but in a straight line (i.e. when exiting the corner) the locking bias of the center diff moves towards 50/50 meaning that rear wheel spin translates into front wheel traction more easily, helping the car to pull out of the corner (yet limiting the car's ability to turn, which is less necessary on exit).

As far as the whole "shifting bias to the rear during cornering"... that *is* what's happening. According to the chart, when you brake and turn in the car has shifted to full rear bias. As you apply throttle for the 1st half of the corner, the bias stays rear-oriented to give the driver a RWD-like throttle steer ability. Once the latteral-G's begin to unload (signifying that the car is through the apex of the turn and is accelerating out to the exit), the bias is ramped forward, adding traction.

Getting back to the yaw sensor. I'm still missing what it's used for that's somehow different from the . It's only ever mentioned in conjunction with the latteral-G sensor. I guess that yaw sensor data combined w/ latteral-Gs might result in more accurate DCCD mdulation, but the addition of a yaw sensor doens't make any fundamental difference in the way the DCCD works. In fact, I would propose that a good driver that's skilled at managing their traction circle would gain no advantage from the addition of a yaw sensor, since the car's yaw and latteral-G's should always match because the driver's not tossing the car around and drifting, thereby making the latteral-G sensor alone sufficient.

I'm much more interested in hearing about the yaw control built into the ABS system. I've seen the tech articles and I understand how it's supposed to work, but is it actually in the '05 STis? I haven't seen any marketing literature that mentions the STi having "traction control" or "yaw control"... you'd think that stuff would get more press.

I'm not arguing I'm trying to learn... Most of the articles and even Subaru's materials are mis-leading. They make it sound like the Torque distribution changes when in actuality it's the amount of wheel spin that it will allow before transfering the torque? Not trying to beat a dead horse I just like to understand how things work and operate...I'm a gear whore.

[sperry]

Originally Posted by dustinr

I'm not arguing I'm trying to learn... Most of the articles and even Subaru's materials are mis-leading. They make it sound like the Torque distribution changes when in actuality it's the amount of wheel spin that it will allow before transfering the torque? Not trying to beat a dead horse I just like to understand how things work and operate...I'm a gear whore.

I didn't mean argue in the "quarrel" sense of the word, just in the "debate" sense of the word.

I'll agree that Subaru's own literature is the worst on this subject, which is why I didn't even believe there was a true yaw sensor on board, since calling the latteral-G sensor a "yaw sensor" is just the sort of thing a marketing department would do to take advantage of the missing AYC on the US EVOs.

As it is, I still don't see a real reason for a yaw sensor, since the DCCD really only needs a lateral-G sensor, unless they also added some sort of anti-spin traction control system that would require a yaw sensor, then decided to work it into the DCCD as well since if it's already on board they might as well use it to help DCCD accuracy.

[tysonK]

There's an article in the new "Drive Performance" Subaru magazine on the DCCD, but I think you guys are way beyond any of that information.

That stupid little magazine pisses me off anyways. It advertises having a performance car will give you cool options to personal the perforamcen of your car. And does articles on "performance" installing a STi short shifter. When we all know as soon as we breathe on our cars the warranty is gone.

Actually I would not be suprised if I installed my own STi factory short shifter it would void my warranty!

[sperry]

Originally Posted by tysonK

Actually I would not be suprised if I installed my own STi factory short shifter it would void my warranty!

Actually, if the dealer installed the STi short shifter, it would void your warranty.

[Kevin M]

Originally Posted by sperry

Actually, if the dealer installed the STi short shifter, it would void your warranty.

It would void your shifter's warranty, and certain problems you could develop with your gearbox. Under the Magnusson-Moss Warranty Act, the rest of it is intact.

[sperry]

Originally Posted by BAN SUVS

It would void your shifter's warranty, and certain problems you could develop with your gearbox. Under the Magnusson-Moss Warranty Act, the rest of it is intact.

Unless Subaru just doesn't feel like honoring the law. You know like, how my exhaust made them install the wrong grease in the CV joint during manufacturing. According to the M-M Act, they had to prove that my exhaust caused the leak in my axle... and they were "very confident in their decision" to deny my warranty claim based on my exhaust and the now infamous "B" sticker.

If you want to keep your warranty, never touch a thing on your own car. Bring it to the dealer for everything. Granted, you'll spend 4 times as much money as you need to, and the car will probably be more jacked up than if you did all the work right the 1st time yourself, but at least that all important warranty woud still be intact. I'm through w/ warranties. Unless you've got access to a pro-bono lawyer, they're unenforcable.

[doubleurx]

ahhh...................W (what is 2+2)