Tweet
Quote:
To understand why low speed throttle is bad, you need to know a little about how ESCs work.
In general, with a constant load a motor's RPM changes as a result of the average voltage it sees. Since a battery is a (relatively) fixed voltage, we need some way to vary voltage seen by the motor. A microcontroller is a digital device - it can generate only "on" and "off" conditions. How does one take a fixed input (battery) and an on/off control and make an analog (variable) output?
The answer is by going on and off very quickly, varying the relative amount of on and off pulses. The longer you are "on" and the shorter "off", the more average voltage the motor sees. This is called "Pulse Width Modulation" or PWM. The on-time is called "duty cycle" and is repeated at a certain rate (times per second) called "frequency." The duty cycle is controlled by the transmitter, the frequency is controlled by the speed control.
In a brushed control, this goes directly into the two motor leads, with the coils and mass of the motor acting to average out the pulses into a more-or-less linear value.
In a brushless control, the same thing is going on, with an added twist. Not only are the pulses varying the voltage, but the voltage must be moved between the 3 different wires in order to simulate the plates of the commutator (comm) on the armature.
Now, why is low speed bad? The on/off switching is done with a device called a FET. A FET is a low-resistance switch when fully turned on. When it's not fully turned on, it has a resistance which varies depending on exactly how turned-on it is. This is called the "linear mode." When in linear mode, a FET is like a big resistor - you know, those things inside your slot car handle that get HOT, or like the ceramic thing in your old mechanical speed control. The problem is, FETs don't like to be in linear mode - as they warm up, they get more and more resistive. But the motor at low RPM is less and less resistive, meaning it will draw more current.
What you have is a situation that can quickly degenerate into what we call "Thermal Runaway" - the heating of something causes it to heat itself even more.
So why don't ESC designers not run FETs in linear mode? Well, we try real hard not to, but it's a matter of physics - you can't completely eliminate it. If we make the linear mode time very short, it adds bigger capacitors to the board to provide the necessary energy, and people complain about size. They also complain about glitching their radios, as the speed at which the FETs are switched then generates radio frequencies. If we spend more time in linear mode, we can eliminate the glitching in the radio - but then we run into potential heat problems. Engineering always is a balance of competing problems and solutions.
What happens at low throttle is that the FET goes into the linear mode for a larger percentage of the on-time. That means the FET spends more time heating than it should - and hot FETs have more resistance, meaning next time they are in linear mode, they will heat up even hotter. At the same time, the motor is at lower RPM and therefore pulling more amps (a stalled motor will draw the maximum current).
You can get a lower RPM (ie, a lower average voltage) at maximum duty cycle only by lowering the input voltage. The speed control will run better with a lower input voltage even though it may be passing more amps simply because there will be less time for the FETs to be in linear mode. It is not an exaggeration to say that the FET may have 10X or more resistance when in linear mode. Heat generated is directly related to resistance.
Andy Kunz
John
To understand why low speed throttle is bad, you need to know a little about how ESCs work.
In general, with a constant load a motor's RPM changes as a result of the average voltage it sees. Since a battery is a (relatively) fixed voltage, we need some way to vary voltage seen by the motor. A microcontroller is a digital device - it can generate only "on" and "off" conditions. How does one take a fixed input (battery) and an on/off control and make an analog (variable) output?
The answer is by going on and off very quickly, varying the relative amount of on and off pulses. The longer you are "on" and the shorter "off", the more average voltage the motor sees. This is called "Pulse Width Modulation" or PWM. The on-time is called "duty cycle" and is repeated at a certain rate (times per second) called "frequency." The duty cycle is controlled by the transmitter, the frequency is controlled by the speed control.
In a brushed control, this goes directly into the two motor leads, with the coils and mass of the motor acting to average out the pulses into a more-or-less linear value.
In a brushless control, the same thing is going on, with an added twist. Not only are the pulses varying the voltage, but the voltage must be moved between the 3 different wires in order to simulate the plates of the commutator (comm) on the armature.
Now, why is low speed bad? The on/off switching is done with a device called a FET. A FET is a low-resistance switch when fully turned on. When it's not fully turned on, it has a resistance which varies depending on exactly how turned-on it is. This is called the "linear mode." When in linear mode, a FET is like a big resistor - you know, those things inside your slot car handle that get HOT, or like the ceramic thing in your old mechanical speed control. The problem is, FETs don't like to be in linear mode - as they warm up, they get more and more resistive. But the motor at low RPM is less and less resistive, meaning it will draw more current.
What you have is a situation that can quickly degenerate into what we call "Thermal Runaway" - the heating of something causes it to heat itself even more.
So why don't ESC designers not run FETs in linear mode? Well, we try real hard not to, but it's a matter of physics - you can't completely eliminate it. If we make the linear mode time very short, it adds bigger capacitors to the board to provide the necessary energy, and people complain about size. They also complain about glitching their radios, as the speed at which the FETs are switched then generates radio frequencies. If we spend more time in linear mode, we can eliminate the glitching in the radio - but then we run into potential heat problems. Engineering always is a balance of competing problems and solutions.
What happens at low throttle is that the FET goes into the linear mode for a larger percentage of the on-time. That means the FET spends more time heating than it should - and hot FETs have more resistance, meaning next time they are in linear mode, they will heat up even hotter. At the same time, the motor is at lower RPM and therefore pulling more amps (a stalled motor will draw the maximum current).
You can get a lower RPM (ie, a lower average voltage) at maximum duty cycle only by lowering the input voltage. The speed control will run better with a lower input voltage even though it may be passing more amps simply because there will be less time for the FETs to be in linear mode. It is not an exaggeration to say that the FET may have 10X or more resistance when in linear mode. Heat generated is directly related to resistance.
Andy Kunz
John
Andy have an email? Some of the things he explains in the quote supplied by John didnt makes sense regarding FET operation states, maybe it was the wording?
Nortavlag Bulc 
Comment