I was able to take a second, very short, test run today.
A combination of bad ABS and poor slicing choices made my rims very fragile, I will fix this as soon as I get more plastic.
Any way…
On this iteration, I finished (and triple checked the math) the calculations for differential correction, placed the batteries in series and installed independent capacitor buffers for each motor.
Check the video, I don’t think you will notice much difference, since it’s very short.
Applying correctly the algorithm for differential correction took a while, fine tunning the response was not as trivial as I expected. But now the car is turning reeeeaaallllllyyy sharp now!!
I was using the batteries in parallel mainly because I didn’t know better.
The specification for the motor said 2-3S and so I did. But when reading the fine letters, I found out this is only for Lipo batteries and mine are NiMH.
For NiMH I should be using 6-9S, so the top 16V of my two batteries should be fine and I have no more problems of the batteries being drained too fast or the car being too slow.
I didn’t know what to expect from the capacitor buffers and without a oscilloscope or some sort of telemetry, it is impossible to take conclusions. But I noticed the motors are responding much faster on any change on throttle and the PWM signals generated by the Arduino are fluctuating less.
What didn’t happen was better torque on low speeds…this was disappointing but after some research I understand why.
The capacitors are just working as a filter, helping to avoid drops on tension during fast changes in throttle and stopping noise from generating in the circuit duo to the electronic properties of the motors.
I have recently discovered Tom Stanton on YouTube and his impressive Electric Bicycle project, check it out: https://www.youtube.com/watch?v=9rIIJbDkpN8&list=PLj3Bh6Krv9CWdRwVFwF1xS_-IQH9GBHyo
What he found is that sensorless motors respond very poorly in low rpms, the ESC is doing it’s best to guess the state of the motor but it just can’t keep up.
Using direct drive is definitely not helping with this situation and a sensored ESC is considerably more expensive. So I don’t see a solution for this problem soon.
I started to install current sensors now, this should help improve response on different situations and help to avoid some others. I’ll explain this better on the next post, after they are in place and well tested.
What is really annoying me for now is not being able to break (just reverse).
But unless someone knows how to get in touch with the guy from BlHeli to get the firmware code, there is nothing I can do about that!
https://www.youtube.com/watch?v=HqTf5WZde5s&index=3&list=PLp-wk6da92n7OjE7TEQDDB73mX-id44l1