Q-Back Posted July 17 Posted July 17 (edited) I find many discussions inconclusive about LeaperKim pedal dipping, let's make sure we understand it correctly. I'm talking mostly about the Lynx because I own one and I do experience this. What is pedal dipping It's the phenomenon that in certain situations, mostly during turning the wheel tilts forwards or backwards. How to reproduce it (Check the attached picture and feel free to edit it in excalidraw https://excalidraw.com/#json=kuLhSfdc9Wovnh_PQE_t0,pLINXtuQtl_8IHc62e85Og) 1. Do the hard acceleration (hitting at least 70% PWM) in straight line. 2. Then brake hard immediately after accelerating 3. Don't come full stop. Instead of this, when you're rolling low speed already start going in the circle At step 3 you should experience that when you face e.g. north then your pedals tilt forwards, but when you face south they tilt backwards. Very similar to bad calibration, but it does go away after a moment. Do this at night with your headlight turned on to see it even more clearly. It goes away after a moment when you do one of these - Just sustain the same speed for few seconds - Stop the wheel for few seconds - When wheel is tilted forward - brake - When wheel is tilted backwards - accelerate Why do we experience it? How EUC perform balancing The EUCs perform balancing using: - at least 2 gyroscopes (1st for forwards/backward tilt, 2nd for left/right tilt) - at least one accelerometer The gyroscopes point right below the EUC when standby - that's due to gravity. But once you accelerate brake or turn the gyroscopes start to drift as gravity is no longer the primary force affecting them. That's why you bring accelerometer to the action - you measure how much do you accelerate or turn right now and you calculate how to compensate the gyroscope drift. Why the pedal dip happens I believe the algorithm used for balancing is not sophisticated enough to handle turning properly at 100% of cases. You must take into account so many variables like: - gyroscopes drift - accelerometer being affected by both acceleration and turning - bumps on the road Therefore the algorithm has some drawbacks and you experience pedal dipping. It's corrected once the gyroscopes are not actively drifting but they hold their pointing position for a moment. My question to you Do you experience the pedal dipping as well on your EUCs (especially Lynx)? I do experience it and I believe it's not caused by the calibration. Edited July 17 by Q-Back Do this at night tip Quote
timmytool Posted July 17 Posted July 17 The filtering of the gyro, and the pid control etc is the cause as far a I know. All data can have spikes and other anomalies. Leaperkim have a filter with a decay of old data to iron out spikes etc thats why you can get a pedal dip on start up if starting on the kickstand and give it a good torque within a few seconds of balancing starting. I cant say this with absolute certainty but my experience and comments from more technically experienced people then me hint to this cause. 1 Quote
Skampster Posted July 18 Posted July 18 Both my Lynx and Sherman S do it minimally, so much so I’m not bothered by it, but my 16x is more obvious. The Falcon is similar to the Leaperkims. Quote
Q-Back Posted July 18 Author Posted July 18 It's also worth mentioning that you must give it a proper acceleration. If you ride it smoothly the pedal dip may not occur. Yesterday I went for chill 70km ride on the Lynx with offroad parts. I didn't notice any pedal dipping at all. But every time when I accelerate hard I clearly get pedal dipping. And by hard acceleration I mean that I'm forcefully leaning forward to get maximum possible acceleration and I'm usually hitting ~80% PWM by accelerating like this with ~70% battery. 5 hours ago, timmytool said: Leaperkim have a filter with a decay of old data to iron out spikes etc That sounds interesting. It would explain why some pedal dip scenarios seems to "random". I mean - sometimes when I accelerate and turn at the same time it occurs and sometimes when I try to do it the same way there's no dipping. Quote
noonewantstobepeterchris Posted July 18 Posted July 18 8 hours ago, timmytool said: Leaperkim have a filter with a decay of old data to iron out spikes etc Love it. That is my go to move when tuning a loop before actually touching any PID terms, especially when using velocity equations. Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.