Hummina Shadeeba
10 MW
The vesc’s internal wattmeter isn’t so accurate but if I clamped a meter to the battery wires could I adjust settings to get lower iron and copper losses?
tune vesc for best efficiency
- Thread starter Hummina Shadeeba
- Start date
jetpackjbd
100 W
Not quite sure what you’re asking, I don’t believe you’ll be able to reduce power losses from the resistances of different metals much if at all with settings
scianiac
10 kW
- Joined
- Oct 24, 2018
- Messages
- 542
Maybe? It's not that you could as much reduce the iron and copper losses as it's possible to improve the efficiency by reducing error. That is the VESC has an ideal of what the motor it's controlling is and is trying to control it based on that. If that is wrong it may not be driving it ideally. Although this effect could be pretty small.
Adjusting the ZVF and MTPA if your motor has enough salience will probably make a bigger difference. I actually was just testing a motor with vesc to see what the no load losses where at various RPMs with and without MTPA (it's an IPM motor, you should be careful with trying MTPA with a non salient motor, careful as in you probably shouldn't.)
But also remember that meter you use needs to accuracy read the average current, which will be pulsing quite fast. The VESC internal meter should be fairly accurate assuming the shunts are accurate. Although the real question is is it inaccuracte or inconsistent, if it's in accuracte but consistent it doesn't matter if you just want to increase the efficiency.
And beyond that you'll make a much larger improvement by adjusting the gear ratio/KV/wheel size to match the average power and motor. I'm working on this now. So if you capture a typical ride log you can figure out how long you spend at different RPMs and power levels, based on that, the iron losses of the motor which could be tested by just spinning the motor at various RPMs and recoding the no load power. Then the resistive losses can be calculated by the resistance. Will take some math but hopefully I should at least get a rough idea what my gear ratio should be for effiicney given that I already have more torque and speed than I can ever use.
Adjusting the ZVF and MTPA if your motor has enough salience will probably make a bigger difference. I actually was just testing a motor with vesc to see what the no load losses where at various RPMs with and without MTPA (it's an IPM motor, you should be careful with trying MTPA with a non salient motor, careful as in you probably shouldn't.)
But also remember that meter you use needs to accuracy read the average current, which will be pulsing quite fast. The VESC internal meter should be fairly accurate assuming the shunts are accurate. Although the real question is is it inaccuracte or inconsistent, if it's in accuracte but consistent it doesn't matter if you just want to increase the efficiency.
And beyond that you'll make a much larger improvement by adjusting the gear ratio/KV/wheel size to match the average power and motor. I'm working on this now. So if you capture a typical ride log you can figure out how long you spend at different RPMs and power levels, based on that, the iron losses of the motor which could be tested by just spinning the motor at various RPMs and recoding the no load power. Then the resistive losses can be calculated by the resistance. Will take some math but hopefully I should at least get a rough idea what my gear ratio should be for effiicney given that I already have more torque and speed than I can ever use.
Hummina Shadeeba
10 MW
I want to run the motor as I normally would and just have the commutation as efficient as possible. I read of methods to get the motor running smoothly but not necessarily efficiently
I think figuring the iron losses with different settings would be easy enough just seeing the no-load current draw at different speeds. Maybe copper losses could be compared between settings with a standardized ride and fixed throttle setting. What are all the settings that would be worth adjusting?
I think increasing timing is more heat in the esc but sends a better wave to the motor and therefore more efficiency in the motor.
I think figuring the iron losses with different settings would be easy enough just seeing the no-load current draw at different speeds. Maybe copper losses could be compared between settings with a standardized ride and fixed throttle setting. What are all the settings that would be worth adjusting?
I think increasing timing is more heat in the esc but sends a better wave to the motor and therefore more efficiency in the motor.
i don't know these settings, a controller doens't contribute that much to efficiency loss as the motor does itself.
You'll probably want an oscilloscope, some way to load the bike ( dyno ) to do the measuring part of the tuning.
You maybe able to change the efficiency by 0.5% or so, i don't expect sizeable gains from this tuning.
i don't know what parameters to look for because i always focus on bigger gains and don't bother with high hanging fruit optimizations.
You'll probably want an oscilloscope, some way to load the bike ( dyno ) to do the measuring part of the tuning.
You maybe able to change the efficiency by 0.5% or so, i don't expect sizeable gains from this tuning.
i don't know what parameters to look for because i always focus on bigger gains and don't bother with high hanging fruit optimizations.
scianiac
10 kW
- Joined
- Oct 24, 2018
- Messages
- 542
I don't think you will see any result from tuning settings unless you have the testing very controlled, the change will probably be very small unless the controller is running the motor poorly. Honestly I would first make sure everything looks stable. When I tune a motor I run the motor connected to a tablet mounted to my handlebars so that I can see the real time and sampled data. Adjust settings and see the result, making sure current and duty cycle are stable. I'm not doing it for efficiency though, rather for power tuning and as you increase power things generally get more unstable and inefficient of course.
If the motor is running fine thought it's probably setup fine and gains will be very small. You have to remember this isn't some old trap controller that just hammers through a waveform whether the motor likes it or not. FOC is trying to drive the motor as efficiently as possible. Timing is a concept from those old controllers that we not trying to match the sine wave from the motor.
You can change the zero vector frequency, in theory this has some effect on switching losses although it also depends on ERPM to some extent, lowering it too much causes other issues though and you can't raise it that much before running out of processor speed, like 40khz is the highest I've seen commonly used.
Most motor losses are not affected by the controller and you are better off on doing some rolling resistance, aero, gearing, motor cooling, etc.
If the motor is running fine thought it's probably setup fine and gains will be very small. You have to remember this isn't some old trap controller that just hammers through a waveform whether the motor likes it or not. FOC is trying to drive the motor as efficiently as possible. Timing is a concept from those old controllers that we not trying to match the sine wave from the motor.
You can change the zero vector frequency, in theory this has some effect on switching losses although it also depends on ERPM to some extent, lowering it too much causes other issues though and you can't raise it that much before running out of processor speed, like 40khz is the highest I've seen commonly used.
Most motor losses are not affected by the controller and you are better off on doing some rolling resistance, aero, gearing, motor cooling, etc.
scianiac
10 kW
- Joined
- Oct 24, 2018
- Messages
- 542
Yes or maybe no. In theory lower frequency does reduce switching loss but it can also introduce other losses and if you go too far instability. It's a tricky thing to tune and probably has very little effect in practice. You have to remember that modern FETs switch much faster with much lower Rds than older FETs so there are just less losses overall. I'm running my VESC Leaf at 16khz but that's because it made it run more stable at the high phase amps I'm running.
Similar threads
