turbo2ltr
Well-known member
You guys are really going to make me prove this, huh?
Acceleration pedal calibration problems ?
- Thread starter adric22
- Start date
Help Support My Nissan Leaf Forum:
Coffee_Slurry
Well-known member
If you don't intend to accelerate (go faster), then don't press the pedal on the right.
Meh. Mine works.
Meh. Mine works.
UkrainianKozak
Well-known member
I feel that in D Leaf actually delivers different power, I feel like it depends on the speed, I don't know why they did it, perhaps to try to mimic ICE behavior (when you are driving at high speed acceleration is less then the same throttle position at low speed), so in D acceleration is much more responsive at slow speeds.
I think by trying to mimic ICE behavior Nissan actually over-deliver... in some conditions they really nailed it, but in some driving scenarios I totally agree that the car is behaving.... amm.. unique... not like old electric cars, and not like ICEs either...
I wish they would do it simpler, or at least have some configurable options so each driver can tune in the throttle response to whatever is more pleasant individually.
I think by trying to mimic ICE behavior Nissan actually over-deliver... in some conditions they really nailed it, but in some driving scenarios I totally agree that the car is behaving.... amm.. unique... not like old electric cars, and not like ICEs either...
I wish they would do it simpler, or at least have some configurable options so each driver can tune in the throttle response to whatever is more pleasant individually.
DaveinOlyWA
Well-known member
ok suppose that depressing pedal 20% will get you 50 mph and 30% gets you 60 mph and so on but 15% only gets you 25 mph, etc.
so you take off gently as always, turn on to the main street and as you pull out of the turn, you hit the 20% level and suddenly the power level jumps up quite a bit because *maybe* the pedal response was say 2 mph per degree of compression but now you are in the "zone" where its 5 mph per degree.
so your power output jumps up to very high and slowly bleeds off as you approach your target speed?
this is basically my impression of the throttle response on the Leaf. it takes position of the pedal, then the delta of the set speed with the current speed and applies power sufficiently to attain set speed at a fairly brisk pace (very brisk at speeds under 50 mph)
now, i might be way off and maybe i am not, but i do notice that throttle response in D mode is HUGELY dependent on your current speed.
so you take off gently as always, turn on to the main street and as you pull out of the turn, you hit the 20% level and suddenly the power level jumps up quite a bit because *maybe* the pedal response was say 2 mph per degree of compression but now you are in the "zone" where its 5 mph per degree.
so your power output jumps up to very high and slowly bleeds off as you approach your target speed?
this is basically my impression of the throttle response on the Leaf. it takes position of the pedal, then the delta of the set speed with the current speed and applies power sufficiently to attain set speed at a fairly brisk pace (very brisk at speeds under 50 mph)
now, i might be way off and maybe i am not, but i do notice that throttle response in D mode is HUGELY dependent on your current speed.
evnow
Well-known member
I've changed to title to avoid misinterpretation by media.
turbo2ltr
Well-known member
It's not like I had anything else to do on a friday night, so...
Here's proof. It happens fast but the proof is in the CAN (messages)
http://www.youtube.com/watch?v=SG66sI0GRqo
You can see throttle value and a value I'm calling torque. I'm not sure it's torque. Its more likely current. Either way, the number naturally goes up with increased pedal, but also goes down with increased speed. So low speed and a lot of throttle produces the highest number. I'm using this value as a reference for "torque" for this demonstration as it has repeatable numbers that represent power output (though it's a slightly complex relationship). The point is it's repeatable which is what we need here. A value for x at y speed is always the same amount of torque output.
Throttle positions is also shown. 0 to 99.5% and accurately represents the position of my foot.
I start at a stop in D.
I mash the throttle as evident of the 99.5%
The car accelerates and you see a very high torque value, then is starts dropping I'm guessing because the motor is above it's base speed, but this is not because it's putting out less power.
At this point I have full power dots.
Then I start letting off the throttle, you can see I let off all the way down to 15% at about 8 second in. At that point the TQ is at 6352. The dashboard still has full power dots. This is the 'heart' of the problem. If I let off from 100% to 15% throttle, there is NO reason I should still have full power. There is no way this is intentional behavior. Because of the complex TQ number, you can't see it, but the power will actually continue to climb as you let off, if you let off before the ramp rate limit lets the power go to 100%.
So then at 9 seconds, I get completely off the throttle, and you see the car go into regen with negative TQ values amd 0 throttle.
I then hit the throttle again. I go back to approximately the same throttle position I was just at that I had full power (I'm not looking at the screen so I'm doing it from muscle memory..I didn't get it perfect 15% but I got close) and what TQ number do we get? At 17% throttle, we get a TQ reading of 1264. That's 5 times less power than we JUST HAD at that same throttle position a second earlier.
I would love to hear what possible rational you guys can come up with for this behavior. I understand they remapped the eco mode throttle to have 0 to 40% power applied during the 0 to 75% of throttle travel, and 40-100% power applied during the last 25% of throttle travel. It makes sense and it works. In Eco, x throttle always = y power.
No, this is not "unintended acceleration" in the typical sense of "my car took off and I couldn't stop". Power ALWAYS goes to zero when you get completely off the throttle. But there is just no way that you should have 100% power output at 15% throttle at one point and 20% power at that same throttle position a second later.
Here's proof. It happens fast but the proof is in the CAN (messages)
http://www.youtube.com/watch?v=SG66sI0GRqo
You can see throttle value and a value I'm calling torque. I'm not sure it's torque. Its more likely current. Either way, the number naturally goes up with increased pedal, but also goes down with increased speed. So low speed and a lot of throttle produces the highest number. I'm using this value as a reference for "torque" for this demonstration as it has repeatable numbers that represent power output (though it's a slightly complex relationship). The point is it's repeatable which is what we need here. A value for x at y speed is always the same amount of torque output.
Throttle positions is also shown. 0 to 99.5% and accurately represents the position of my foot.
I start at a stop in D.
I mash the throttle as evident of the 99.5%
The car accelerates and you see a very high torque value, then is starts dropping I'm guessing because the motor is above it's base speed, but this is not because it's putting out less power.
At this point I have full power dots.
Then I start letting off the throttle, you can see I let off all the way down to 15% at about 8 second in. At that point the TQ is at 6352. The dashboard still has full power dots. This is the 'heart' of the problem. If I let off from 100% to 15% throttle, there is NO reason I should still have full power. There is no way this is intentional behavior. Because of the complex TQ number, you can't see it, but the power will actually continue to climb as you let off, if you let off before the ramp rate limit lets the power go to 100%.
So then at 9 seconds, I get completely off the throttle, and you see the car go into regen with negative TQ values amd 0 throttle.
I then hit the throttle again. I go back to approximately the same throttle position I was just at that I had full power (I'm not looking at the screen so I'm doing it from muscle memory..I didn't get it perfect 15% but I got close) and what TQ number do we get? At 17% throttle, we get a TQ reading of 1264. That's 5 times less power than we JUST HAD at that same throttle position a second earlier.
I would love to hear what possible rational you guys can come up with for this behavior. I understand they remapped the eco mode throttle to have 0 to 40% power applied during the 0 to 75% of throttle travel, and 40-100% power applied during the last 25% of throttle travel. It makes sense and it works. In Eco, x throttle always = y power.
No, this is not "unintended acceleration" in the typical sense of "my car took off and I couldn't stop". Power ALWAYS goes to zero when you get completely off the throttle. But there is just no way that you should have 100% power output at 15% throttle at one point and 20% power at that same throttle position a second later.
turbo2ltr
Well-known member
When playing the video, hit pause, then hit the number 4 (on your keyboard) and you'll see 17% at 6864, hit 8 and you'll see 18% @ 1360.
garygid
Well-known member
Three possible reasons, probably all of these:
1. the torque is probably a function of both the pedal position and the rate of pedal depression, to give what appears to be a "snappy" (quick) response, in spite of #2.
2. the pedal values are smoothed over a few seconds, with a faster increase-response and a slower decrease time constant so that the motor cannot be ripped apart by tapping "Flight of the BumbleBee" on the accelerator.
3. The programmer tried to accomplish conflicting functions, and there were un-compensated side effects.
1. the torque is probably a function of both the pedal position and the rate of pedal depression, to give what appears to be a "snappy" (quick) response, in spite of #2.
2. the pedal values are smoothed over a few seconds, with a faster increase-response and a slower decrease time constant so that the motor cannot be ripped apart by tapping "Flight of the BumbleBee" on the accelerator.
3. The programmer tried to accomplish conflicting functions, and there were un-compensated side effects.
turbo2ltr
Well-known member
Well I don't agree with 1 or 2.
There's no reason to apply more torque than you normally would if you have a high rate of change of throttle. If the user needs more torque, the apply more pedal.
Pedal values should NEVER be smoothed over seconds. Maybe a few 10ths of a second to get rid of noise, but should never be smoothed in the negative direction. If you get off the throttle, then tq should be reduced, period. There's no reason to smooth it. Sure you can smooth it in the positive direction but even then, if you hit the throttle, you want it to go now, not in a few seconds.
My guess is there is an edge case that doesn't properly handle getting off the throttle during the ramp rate limiting they do if you mash the throttle at a stop or very very low speed. This is why if you are rolling already at a few MPH and mash the throttle, this behavior does not occur.
There's no reason to apply more torque than you normally would if you have a high rate of change of throttle. If the user needs more torque, the apply more pedal.
Pedal values should NEVER be smoothed over seconds. Maybe a few 10ths of a second to get rid of noise, but should never be smoothed in the negative direction. If you get off the throttle, then tq should be reduced, period. There's no reason to smooth it. Sure you can smooth it in the positive direction but even then, if you hit the throttle, you want it to go now, not in a few seconds.
My guess is there is an edge case that doesn't properly handle getting off the throttle during the ramp rate limiting they do if you mash the throttle at a stop or very very low speed. This is why if you are rolling already at a few MPH and mash the throttle, this behavior does not occur.
Maybe the power dots are lies... kinda like the Guess-O-Meter "miles remaining".turbo2ltr said:
I can't speculate on what actually happens to "power" when letting off the throttle, but if it's true that one does not get 100% power immediately by flooring it because Nissan wants to avoid sudden maximum stress to the drivetrain, it's entirely reasonable that power would not go immediately to zero and into re-gen when suddenly letting go of the throttle.turbo2ltr said:
Your findings are interesting. They illustrate the non-linear relationship between "power" and throttle pedal position, which kinda the issue in the OP's scenario. But, the events are quite different: you floor it then ease up, whereas the OP's scenario is (more or less) constant throttle then a small increase, which resulted in "full power dots" and acceleration. Your scenario is reproducible; the original OP's is not.turbo2ltr said:
I don't know if there will be any answers any time soon; it'll be interesting to know what the story is if/when the OP has the car looked at by a dealer.
Agreed. Thanks for doing the logging - very informative. Some small amount of damping could be applied to throttle input - but eventually (where eventually is not very long) power output at the same throttle input (and perhaps speed?) should end up the same.turbo2ltr said:
Just one note - the bubble gauge is definitely not linear - one can get "full" bubbles at well less than full power. For example - it's pretty easy to see that 5 bubbles is around 25 kW.
I noticed something strange the other day coming around an undulating corner the other day which I'm not sure was caused by this throttle strangeness exasperated by the undulations causing me to be inconsistent with throttle - going around the corner which has two slight humps - when I hit the first hump power went to full or near full bubbles - was cut back down to desired bubble level of 3-4 coming off the undulation and back to full or near full again as I hit the second hump - saw a light flash on the dash in there as well (traction control kicking in?). I had intended to keep a a ~4 bubble power range through the corner but for some reason was not able to. In previous car never had this issue. Not sure how much of it was user error - road conditions don't often let me take this corner at the pace required for the undulations to jounce the car a bit to repeat. ECO mode was on should it should have been very difficult to get to full bubble power.
turbo2ltr
Well-known member
The bubble gauge may not be linear, but I don't think you'll ever see full bubbles without full power (or close to it). There's a difference between non-linear and repeatable. I do know there is a fair bit (my guess is about +/-7kw) of hysteresis which may make it look like the gauge shows more (or less) than actual power.
I've had the stability system completely cut the throttle on me on several occasions. It seems to get triggered if you hit a sharp bump while going through a turn "spirited". (Drive it like you stole it!) I believe the system is detecting excessive and sudden yaw due to the momentary loss of grip caused by the bump. The light on the dash will light and will reduce power, and in some cases will completely cut the power (I've had my foot on the floor with zero power) for a few seconds.
Car companies should hire me to find edge cases (though I'm not implying the above is an edge case, I believe it's 100% intended operation of the stability control system).
I've had the stability system completely cut the throttle on me on several occasions. It seems to get triggered if you hit a sharp bump while going through a turn "spirited". (Drive it like you stole it!) I believe the system is detecting excessive and sudden yaw due to the momentary loss of grip caused by the bump. The light on the dash will light and will reduce power, and in some cases will completely cut the power (I've had my foot on the floor with zero power) for a few seconds.
Car companies should hire me to find edge cases (though I'm not implying the above is an edge case, I believe it's 100% intended operation of the stability control system).
The times I've tested I seem to get full bubbles around 60 kW - hard to watch both the bubbles and dial when going that fast unless up a long straight hill.turbo2ltr said:
I think this may have also affect my "strange" throttle response over the undulations as described above...turbo2ltr said:
Agreed!turbo2ltr said:
DaveinOlyWA
Well-known member
definitely could be better attuned to driving needs, but it sounds like simply taking foot off pedal and re applying "reset" the acceleration demand.
but i agree that the initial request for acceleration does seem to change the parameters
but i agree that the initial request for acceleration does seem to change the parameters
planet4ever
Well-known member
Sounds to me like you were sloshing the electrons around in the battery, and the fuel pump was having trouble picking up a steady flow. :lol:drees said:
Ray
adric22
Well-known member
turbo2ltr said:
Thank you for this.. You've nailed it. Now I know I'm not crazy.
ttweed
Well-known member
Just curious--when you did the Youtube test, was TC on or off? Have you tried it both ways? Is the result any different?turbo2ltr said:
Thx,
TT
turbo2ltr
Well-known member
It was on. While I don't expect any difference, I can't say I specifically tested it. I'll try it next time I'm out.
DaveinOlyWA
Well-known member
turbo2ltr said:
makes a lot of difference in snow. i tried it last winter when we had a thin layer of snow over ice. i pulsed the acclerator and was getting speed bumps to the 20's and 30's. turned it off and it pulsed to the 80's and 90's so yeah, the accelerator is a bit "touchy"
http://www.youtube.com/watch?v=fNvC3BG4Syg
u see the first time i hit it, it jumped to 90 and i really did not hit it that hard. the 2nd time it went to 30-40 mph with depressing throttle maybe an inch
garygid
Well-known member
The sudden 90 mph just shows that instantaneous wheel speed, even a small spin/slip, gets translated into a visible mph.
That would be interesting to watch some CAN Logs of such an event. Now, combined 3-CAN logging is easy to do with one custom-end on an OBD cable, a Windows PC (running CAN-Do), and about 4 x $55 of off-the-shelf equipment. The 3 AVR-CAN cards (by Olimex) need to be "flashed" with our "matching" firmware, and a low-cost "USB-AVR-JTAG" programmer combined with the free AVRdude software will flash the boards.
To make investigating such events easier:
I will try and improve the CAN-Do Log viewer to show only a user-selected time period, and to have a playback "speed-control" for watching variables on its "Dashboard".
That would be interesting to watch some CAN Logs of such an event. Now, combined 3-CAN logging is easy to do with one custom-end on an OBD cable, a Windows PC (running CAN-Do), and about 4 x $55 of off-the-shelf equipment. The 3 AVR-CAN cards (by Olimex) need to be "flashed" with our "matching" firmware, and a low-cost "USB-AVR-JTAG" programmer combined with the free AVRdude software will flash the boards.
To make investigating such events easier:
I will try and improve the CAN-Do Log viewer to show only a user-selected time period, and to have a playback "speed-control" for watching variables on its "Dashboard".
Similar threads
