Ricky_nz
10 kW
This is a excerpt from my crash thread that i think really needs to stand alone rather than being burried in my longish thread. It might save someone some pain...
Crash thread here http://endless-sphere.com/forums/viewtopic.php?f=1&t=68791
Analysis of throttle response in my accident.
First if all to lift the front wheel i was getting up to speed (25 - 30kph) then rolling the throttle back to arround 500W to stop accelerating then i would give a sharp twist to get the front end up.
Hardware setup:
Direct connect cycle analyst v3 to 12fet controller and mac 10t rear motor woth magura pot twist throttle.
The cycle analyst was configured to allow use of the pot throttle with the controller expecting a hall throttle. It was also configured to make the throttle behave like a current/torque throtttle.
It was also configured so that below 5A battery current the current rise rate was set slow. I did this to attempt to always have the sprag clutch locked before applying real power as i had previously blown a clutch up at 1KW launching hard too many times and after going to arround 2.1KW I wanted some extra protection.
Above 5A i had the rise rate set as fast as possible to follow the throttle to allow fun.
i had the fall rates set fast so as to not have throttle hang after release.
This all worked well apart from one instance that my leg/knee wishes didnt happen.
The clutch was surviving well.
When i attempted the last wheelie I didnt watch the Watts but i believe i came off the throttle a bit too far dropping the current below 5A briefly which then had the effect of delaying throttle response for a short period (100s of ms) which I think caused me to grab too much throttle before the wheel came up. Basically behaved like turbo lag.
This lag sometimes and not others depending on how much i backed off the throttle made things a little unpredictable especially since I didnt realise that would happen before the crash.
The nonlinearaity and delay are two things that are bad in control systems.
The delay in feedback to be of the throttle response caused more throttle which then came on hard when the nonlinear rate change occurred.
This should act as a warning to be careful how you set your cycle analyst up because both delay and nonlinear/step changes are not the best.
I think i need to fine tune these settings to get the low power part to lock up the clutch quicker but still slow enough to save the clutch.
Of course for normal riding the setup was not a real issue. Rode it for months fine but the issue lay in wait for me when i pushed the envelope.
Crash thread here http://endless-sphere.com/forums/viewtopic.php?f=1&t=68791
Analysis of throttle response in my accident.
First if all to lift the front wheel i was getting up to speed (25 - 30kph) then rolling the throttle back to arround 500W to stop accelerating then i would give a sharp twist to get the front end up.
Hardware setup:
Direct connect cycle analyst v3 to 12fet controller and mac 10t rear motor woth magura pot twist throttle.
The cycle analyst was configured to allow use of the pot throttle with the controller expecting a hall throttle. It was also configured to make the throttle behave like a current/torque throtttle.
It was also configured so that below 5A battery current the current rise rate was set slow. I did this to attempt to always have the sprag clutch locked before applying real power as i had previously blown a clutch up at 1KW launching hard too many times and after going to arround 2.1KW I wanted some extra protection.
Above 5A i had the rise rate set as fast as possible to follow the throttle to allow fun.
i had the fall rates set fast so as to not have throttle hang after release.
This all worked well apart from one instance that my leg/knee wishes didnt happen.
The clutch was surviving well.
When i attempted the last wheelie I didnt watch the Watts but i believe i came off the throttle a bit too far dropping the current below 5A briefly which then had the effect of delaying throttle response for a short period (100s of ms) which I think caused me to grab too much throttle before the wheel came up. Basically behaved like turbo lag.
This lag sometimes and not others depending on how much i backed off the throttle made things a little unpredictable especially since I didnt realise that would happen before the crash.
The nonlinearaity and delay are two things that are bad in control systems.
The delay in feedback to be of the throttle response caused more throttle which then came on hard when the nonlinear rate change occurred.
This should act as a warning to be careful how you set your cycle analyst up because both delay and nonlinear/step changes are not the best.
I think i need to fine tune these settings to get the low power part to lock up the clutch quicker but still slow enough to save the clutch.
Of course for normal riding the setup was not a real issue. Rode it for months fine but the issue lay in wait for me when i pushed the envelope.
