Hitting rev limiter with supercharger

[sw20kosh]

your making too much power with that JRSC lol

[wparsons]

Originally Posted by CSG Mike

You do NOT want to be hitting the rev limiter with a HBP JRSC.

What's the reasoning? Over spinning the SC with the high boost pulley?

[Toyota86.ir]

Originally Posted by xsnapshot

I have a BRZ automatic with the JRSC and high boost pulley on E85.

When I floor it in D, I bounce off the rev limiter in 1st and 2nd before it has time to shift. I think it's just simply revving too fast. 3rd and onward seem ok.

Is there a way to program the transmission to shift sooner or am I SOL and have to use manual mode?

AT gearbox limit is 300bhp and JR HBP make 360-380bhp , i can not imagine that power level on AT!!!

Did you drived hard with no problem?

[CSG Mike]

Originally Posted by wparsons

What's the reasoning? Over spinning the SC with the high boost pulley?

1. you don't want to overspin the SC with an unintentional overrev

2. What do you think shock loads like that do at 100,000 RPMs? That's how fast the Supercharger is spinning (technically about 95,000-96,000 at redline...)

[wparsons]

Originally Posted by CSG Mike

1. you don't want to overspin the SC with an unintentional overrev

2. What do you think shock loads like that do at 100,000 RPMs? That's how fast the Supercharger is spinning (technically about 95,000-96,000 at redline...)

Oh no doubt, just based on your post that specifically mentioned the HBP I was curious if it was more about over spinning it, or the bouncing off the limiter part.

[CSG Mike]

Originally Posted by wparsons

Oh no doubt, just based on your post that specifically mentioned the HBP I was curious if it was more about over spinning it, or the bouncing off the limiter part.

Both really...

[xsnapshot]

Originally Posted by CSG Mike

I think this would be a better place to go, as my ability to explain this is somewhat limited.

https://en.wikipedia.org/wiki/Jerk_(physics)

A slightly longer version, using an example:

- At a constant vehicle velocity going forward, you don't feel any force.
- At a constant acceleration going forward, you feel the force, but the force itself is "constant" and easy to adapt to.
- With a variable acceleration going forward, you'll get thrown around, because you're trying to adapt to a constantly changing force. This is Jerk (change in acceleration). It's also why a lot of people prefer a turbo over a sc; that rapid climb in torque output as the turbo spools, causes a large jerk, while the linear torque output of a SC/NA car has a small jerk.

As acceleration is change in velocity over time, jerk is change in acceleration over time.

Now, lets translate this back into our example.

When you hit the rev limiter, the force causing the acceleration in RPMs is instantly taken away. However, the rate of climb in RPMs, is still present...

I have read through all of your replies.

I still am having trouble seeing how this is happening due to how you claim it is happening. Jerk is the derivative of acceleration. However (under a non-changing moment of inertia system) acceleration once a force is removed is zero. That seems to pose an issue.

The link you provide here says exactly that....

I also understand math and am actually a mechanical engineer as well.

What you are essentially saying is that if I have an electric motor with flywheel on the end and I spin up that motor such that it's velocity is exponentially increasing...(think constant jerk, linearly increasing acceleration) and then I turn that motor off.....you are saying that motor will continue to spin faster AFTER that torque is removed for a short while. This is not possible.

I see that data and postulate that something is going on with an inertia change of the rotating assembly being decoupled from the transmission. I don't see how Jerk can have anything to do with it since the force torque is zero.

What would be interesting is to see what would happen if you left the car in gear whilst hitting the rev limiter. Your RPM doesn't increase until you clutch in. My guess is RPM would NOT increase past the rev limiter in that scenario.

Also what mods are on this car you've presented?...just curious

[xsnapshot]

Originally Posted by CSG Mike

Both really...

So then what is one to do? No one addressed my initial question of whether or not the AT can be programmed to start it's shift sooner.

I had this same problem in my 2004 WRX AT, and the only solution I had for that vehicle was to raise redline from 6900 to 7050 (I think...it's been a while. I remember I had to raise it a small amount). That gave the AT enough time to complete it's shift before I hit the rev limiter.

In this vehicle however, I don't want to increase my redline past 7400 RPM.

[xsnapshot]

Originally Posted by Toyota86.ir

AT gearbox limit is 300bhp and JR HBP make 360-380bhp , i can not imagine that power level on AT!!!

Did you drived hard with no problem?

AT's are not usually given hp ratings (usually torque). IIRC this AT is sourced from the Lexus sport line and is coupled to a motor that makes significantly more torque than our stock FA20.

Plenty of people are running much more than 300 bhp through their AT's. I guess only time will tell how they hold up. So far mine has been fine but I don't have many miles on it yet.

[Ultramaroon]

Originally Posted by CSG Mike

Zoomed in to the highlighted spot. Note how the clutch is fully depressed, and the transmission is out of gear (on a very obviously grannied shift), and the gas pedal is completely off, yet the RPMs still momentarily keep rising.

I have an alternate interpretation of the data which hinges on a few assumptions.

1. The clutch switch recorded is the starter interlock, not the cruise cancel.

2. The sample rate for all but the tach is 10 Hz.

3. We acknowledge the latency between throttle command and servo response is a factor. I regularly hear a little extra "vroom" when I disengage the clutch and attribute this to DBW latency.

If the above are true, it appears that we are looking at a single bounce off of the rev limiter because the engine is unloaded while still on the throttle. I see the loss of gear indication before clutch disengagement. The only way to explain the below sequence of events is if the wrong clutch switch is being recorded.



Comments?

[CSG Mike]

Originally Posted by xsnapshot

I have read through all of your replies.

I still am having trouble seeing how this is happening due to how you claim it is happening. Jerk is the derivative of acceleration. However (under a non-changing moment of inertia system) acceleration once a force is removed is zero. That seems to pose an issue.

The link you provide here says exactly that....

I also understand math and am actually a mechanical engineer as well.

What you are essentially saying is that if I have an electric motor with flywheel on the end and I spin up that motor such that it's velocity is exponentially increasing...(think constant jerk, linearly increasing acceleration) and then I turn that motor off.....you are saying that motor will continue to spin faster AFTER that torque is removed for a short while. This is not possible.

I see that data and postulate that something is going on with an inertia change of the rotating assembly being decoupled from the transmission. I don't see how Jerk can have anything to do with it since the force torque is zero.

What would be interesting is to see what would happen if you left the car in gear whilst hitting the rev limiter. Your RPM doesn't increase until you clutch in. My guess is RPM would NOT increase past the rev limiter in that scenario.

Also what mods are on this car you've presented?...just curious

Mathematically, if you had constant jerk input for a linearly increasing acceleration, assuming the motor itself had zero internal drag, then would removing the jerk:

- Result in zero acceleration
- Result in negative acceleration

- What would the net result be in angular velocity on the flywheel itself?

The car from which this data is presented, is from a JRSC, in CARB trim


Some anecdotal data:

With a (hard) redline of 8400 on my Honda, this is what happens when approaching it at WOT:

- Hitting the redline in 1st gear, results in an overrev of approximately 8800
- Hitting the redline in 2nd gear, results in an overrev of of approximately 8750.
- Hitting the redline in 3rd gear, results in an overrev of about 8600
- Hitting the redline in 4th gear, results in ~8530.
- Hitting the redline, in 3rd gear, on a downhill grade of about 18 degrees, results in an overrev of 8700+

- All of these are with the assumption that the car is kept in gear, throttle pinned, and effectively "riding the limiter".
- Data is all taken from my personal S2000; this type of "abuse" is something I can't, in good conscience, do to someone else's car, unless specifically asked to (which is why I don't have this type of data with a FA20).
- Tune is Speed Density (as are all S2000)
- Testing was performed to determine what redline I actually wanted, so that I see the max rpm that I want to see, as there are places where I will actually want to ride the limiter, and will hit the limiter hard.

- Approaching the redline softly (part throttle, but open loop), results in it never going over 8430.



Other food for thought.

Engine output is not a constant, but a very rapid series of torques. Can the torque of an individual combustion cycle actually cause RPMs to rise that much? This isn't math I've actually done; just theory, as I have no idea how to compute the losses, but a quick google would indicate about 75% of the energy would be lost.

[CSG Mike]

Originally Posted by Ultramaroon

I have an alternate interpretation of the data which hinges on a few assumptions.

1. The clutch switch recorded is the starter interlock, not the cruise cancel.

2. The sample rate for all but the tach is 10 Hz.

3. We acknowledge the latency between throttle command and servo response is a factor. I regularly hear a little extra "vroom" when I disengage the clutch and attribute this to DBW latency.

If the above are true, it appears that we are looking at a single bounce off of the rev limiter because the engine is unloaded while still on the throttle. I see the loss of gear indication before clutch disengagement. The only way to explain the below sequence of events is if the wrong clutch switch is being recorded.



Comments?

Some more data for you

The hard pivots are from the data being recorded at a slower rate; you're correct in that many channels are 10 and 5 hz. They're the ones I usually don't really need to worry about, so are polled lower.

The data here is mostly at a higher poll rate, but unfortunately it's super old before I started logging everything at 50+ hz. It's from the same data set, but zoomed in more.

[xsnapshot]

Originally Posted by Ultramaroon

I have an alternate interpretation of the data which hinges on a few assumptions.

1. The clutch switch recorded is the starter interlock, not the cruise cancel.

2. The sample rate for all but the tach is 10 Hz.

3. We acknowledge the latency between throttle command and servo response is a factor. I regularly hear a little extra "vroom" when I disengage the clutch and attribute this to DBW latency.

If the above are true, it appears that we are looking at a single bounce off of the rev limiter because the engine is unloaded while still on the throttle. I see the loss of gear indication before clutch disengagement. The only way to explain the below sequence of events is if the wrong clutch switch is being recorded.



Comments?

I could see that. There has to be more to this problem. I've contemplated things like

1.) the throttle plate closing causing motor to quickly see a vacuum thus causing a change in the inertia of the system. (doesn't seem like a strong enough effect).

2.) something happening with the decoupling of the drivetrain from the motor

I don't really have a good answer for this.

Originally Posted by CSG Mike

Mathematically, if you had constant jerk input for a linearly increasing acceleration, assuming the motor itself had zero internal drag, then would removing the jerk:

- Result in zero acceleration
- Result in negative acceleration

- What would the net result be in angular velocity on the flywheel itself?

The car from which this data is presented, is from a JRSC, in CARB trim


Some anecdotal data:

With a (hard) redline of 8400 on my Honda, this is what happens when approaching it at WOT:

- Hitting the redline in 1st gear, results in an overrev of approximately 8800
- Hitting the redline in 2nd gear, results in an overrev of of approximately 8750.
- Hitting the redline in 3rd gear, results in an overrev of about 8600
- Hitting the redline in 4th gear, results in ~8530.
- Hitting the redline, in 3rd gear, on a downhill grade of about 18 degrees, results in an overrev of 8700+

- All of these are with the assumption that the car is kept in gear, throttle pinned, and effectively "riding the limiter".
- Data is all taken from my personal S2000; this type of "abuse" is something I can't, in good conscience, do to someone else's car, unless specifically asked to (which is why I don't have this type of data with a FA20).
- Tune is Speed Density (as are all S2000)
- Testing was performed to determine what redline I actually wanted, so that I see the max rpm that I want to see, as there are places where I will actually want to ride the limiter, and will hit the limiter hard.

- Approaching the redline softly (part throttle, but open loop), results in it never going over 8430.



Other food for thought.

Engine output is not a constant, but a very rapid series of torques. Can the torque of an individual combustion cycle actually cause RPMs to rise that much? This isn't math I've actually done; just theory, as I have no idea how to compute the losses, but a quick google would indicate about 75% of the energy would be lost.

The way I understand it, mathematically, when you remove the torque from the motor, you would accordingly have zero Jerk and Acceleration. It still has an inherent velocity of course. If you had drag on the system then you would have a deceleration upon removal of the torque. (and whatever corresponding jerk would be due to that drag force)

I can't see any single combustion event raising RPM's as much as you are seeing. That's quite a rise...400 RPM.

According to your data it's happening for 1/10th of a second. At ~7000 RPM that means the motor is accelerating for 11 revolutions beyond the limiter.

What kind of rev limiter do you have in the S2000? A hard rev limit? Ignition cut? Fuel cut? Soft rev limit?

Would you happen to know if a soft rev limiter would fix this problem? I had that NA and on the rare occasion I hit the limiter, it sure seemed a lot "nicer" on everything than the hard limiter.

[Ultramaroon]

Originally Posted by CSG Mike

Some more data for you

The hard pivots are from the data being recorded at a slower rate; you're correct in that many channels are 10 and 5 hz. They're the ones I usually don't really need to worry about, so are polled lower.

The data here is mostly at a higher poll rate, but unfortunately it's super old before I started logging everything at 50+ hz. It's from the same data set, but zoomed in more.

Thanks, Mike. I'll delve into it tonight after work.