Crazy Idea / Question.

[adamlwvdc36]

Hi everyone, I have been reading in the "shadows" for sometime now, and decided it was time to speak up.

Where my head is at ATM:

Has anyone ever entertained the notion of hijacking the electronics on a phantom to drive larger ESC's and thus create a larger QC? In theory, you could use all of the onboard electronics and even utilize DJI's firmware and the Go app. I'm thinking it wouldn't be too much more than tapping into the PWM to each "native" ESC.

Anybody?

 

[solentlife]

Why larger ESC's ?

ESC's max out does not decide anything ... the motors max demand does.

You can have a max demand motor of eg 30A .... put a 100A ESC ... it will still only have to deliver 30A.

Nigel

 

[Numone]

What exactly would you be trying to achieve?

 

[adamlwvdc36]

Why larger ESC's ?

ESC's max out does not decide anything ... the motors max demand does.

You can have a max demand motor of eg 30A .... put a 100A ESC ... it will still only have to deliver 30A.

Correct, but if I wanted to put larger motors on it.... to be able to lift more weight

 

[adamlwvdc36]

What exactly would you be trying to achieve?

I'm just brainstorming here. Let's say to design a drone sized where the motors form a square that would be 3 or 4 feet square, for instance. Something bigger. To replace the main board on a Phantom 3 Pro for example is about 130 dollars. This and my own frame with new ESC's, motors, and props. Of course I would need to get the Vision module and a P3P gimbal. A power circuit for different batteries.

Like I said, I am not going out and buying parts... yet. Just trying to throw it out there for concept.

 

[Numone]

I'm just brainstorming here. Let's say to design a drone sized where the motors form a square that would be 3 or 4 feet square, for instance. Something bigger. To replace the main board on a Phantom 3 Pro for example is about 130 dollars. This and my own frame with new ESC's, motors, and props. Of course I would need to get the Vision module and a P3P gimbal. A power circuit for different batteries.

Like I said, I am not going out and buying parts... yet. Just trying to throw it out there for concept.

I can only imagine that the system in its entirety is completely designed around the size/weight of the Phantom.

 

[adamlwvdc36]

I thought that too... Until I took mine apart and attached an oscilloscope to the test points and found that it adjusts motor speed to maintain level flight in Atti mode, GPS Location in any GPS mode. Of course, DJI could have a value in it's software that says for this much pitch, change pulse width by this much, then adjust accordingly. I don't really know, this is why I am asking if anyone else has done it.

 

[Numone]

I thought that too... Until I took mine apart and attached an oscilloscope to the test points and found that it adjusts motor speed to maintain level flight in Atti mode, GPS Location in any GPS mode. Of course, DJI could have a value in it's software that says for this much pitch, change pulse width by this much, then adjust accordingly. I don't really know, this is why I am asking if anyone else has done it.

I suspect that DJI would've protected their system from cannibalisation.

 

[With The Birds]

I thought that too... Until I took mine apart and attached an oscilloscope to the test points and found that it adjusts motor speed to maintain level flight in Atti mode, GPS Location in any GPS mode. Of course, DJI could have a value in it's software that says for this much pitch, change pulse width by this much, then adjust accordingly. I don't really know, this is why I am asking if anyone else has done it.

You didn't need a scope to work that out (the changing motor speeds). I don't understand the weight/size issue- motor speeds must vary to maintain position and change flight attitude regardless of AC mass.

What test point are you talking about?

 

[happydays]

You'd be better off with a 550 or similar. Larger props - that's the Phantom's limiting factor.

 

[adamlwvdc36]

You didn't need a scope to work that out (the changing motor speeds). I don't understand the weight/size issue- motor speeds must vary to maintain position and change flight attitude regardless of AC mass.

What test point are you talking about?

The raw PWM connection from the controller to the ESC circuit

 

[solentlife]

OK ... so now we have a bit more info.

First - the Flight Controller with DJI is more than just a FC ... it also has the power section to supply the ESC"s... so if you isolate and replace the ESC's with larger - you will probably hit max that the FC power section can supply.

The flight characteristics of hover, level flight etc./ are based on the GPS / VPS and barometer reporting back to the FC - so that's no problem regardless of how big you go motors / props ... the FC will adjust the RPM of the motors to balance out for desired command.

Summary ... yes of course its possible - BUT can the power section of the FC take such power demand to give an increased lifting capability ?
It will also mean significant reduction of flight time... and whether the DJI battery can suffer the increased load ?

Nigel

 

[With The Birds]

OK ... so now we have a bit more info.

First - the Flight Controller with DJI is more than just a FC ... it also has the power section to supply the ESC"s... so if you isolate and replace the ESC's with larger - you will probably hit max that the FC power section can supply.

The flight characteristics of hover, level flight etc./ are based on the GPS / VPS and barometer reporting back to the FC - so that's no problem regardless of how big you go motors / props ... the FC will adjust the RPM of the motors to balance out for desired command.

Summary ... yes of course its possible - BUT can the power section of the FC take such power demand to give an increased lifting capability ?
It will also mean significant reduction of flight time... and whether the DJI battery can suffer the increased load ?

Nigel

The flight controller isn’t driving any ESC directly, it simply outputs a reference signal you are likely familiar with, a typical servo drive signal. The waveform at this point is the same as you will see from most RC receivers driving an ESC or normal servo (1ms =0 2ms= 100%). This signal, along with a voltage reference from the output mosfets, is used by a speed controller SOC to generate a PWM reference for a three phase gate driver IC which drives the power mosfets feeding the motors. You could probably upgrade the mosfets but would likely need to reflash the speed controller ROM, this would def be the case with a motor swap, principally due to motor inductance values. The current mosfets are good for 15A RMS (23A peak).

 

[With The Birds]

The raw PWM connection from the controller to the ESC circuit

So between the speed controller soc and mosfet driver? The flight controller output would be a 50-500hz square wave.

 

[N017RW]

No, the frequency is constant and somewhat irrelevant in a PWM system. Don't confuse this with it being random though. Most common r/c stuff uses a 50hz refresh rate.

The PWM is pulse WIDTH modulation.

Don't confuse this with duty-cycle either.

 

[solentlife]

The flight controller isn’t driving any ESC directly, it simply outputs a reference signal you are likely familiar with, a typical servo drive signal. The waveform at this point is the same as you will see from most RC receivers driving an ESC or normal servo (1ms =0 2ms= 100%). This signal, along with a voltage reference from the output mosfets, is used by a speed controller SOC to generate a PWM reference for a three phase gate driver IC which drives the power mosfets feeding the motors. You could probably upgrade the mosfets but would likely need to reflash the speed controller ROM, this would def be the case with a motor swap, principally due to motor inductance values. The current mosfets are good for 15A RMS (23A peak).

You mistook my point ..

ALL ESC's are 'governed' by signal ....

My point is the actual power for motors passes through a power section on the board ... to the ESC's ... then to the motors. That would likely be maxed out by any increase in motor and demand.

Nigel

 

[solentlife]

Just to show what can be done with a 'dead' P3 :

"I have a full Phantom 3S with Controller and batterys. It took a swim in the river and never flew again.

The choice was to revamp her as a P3S but that would only be sharing space with my better P3Pro.
The idea discussed with my pal who is a dab hand at FPV and Quads - was to rip out the dead DJI gear and replace with common Flight Controller and parts.

As far as we can tell - the motors are fine .. so that's one item we can leave. The shell is good enough .... that's another. That also allows us to use the battery holder built into the shell ...

The TX and Radio gear is to be my 9xr and FrSky DJT / V8JT ... so will be PWM based.

Parts ordered (Banggood) :

PWM to PPM encoder for on-board conversion from FrSky Rx to FC :

https://www.banggood.com/PWM-To-PPM-...p-1070269.html

Flight controller :

https://www.banggood.com/BETAFLIGHT-...p-1162891.html

4x ESC :

https://www.banggood.com/Racerstar-R...p-1105108.html

GPS with Compass :

https://www.banggood.com/UBLOX-NEO-M...-p-971082.html

I already have the combined Tx and camera from Hobby King :

https://hobbyking.com/en_us/fx718-6-...-play-kit.html

and I ordered adaptor for Rx to Android as a tester before going for full Goggles :

https://www.banggood.com/Eachine-ROT...p-1147692.html

Total price $98 .... a far cry from the ~$200 just to replace with DJI gear just to fly !

Hopefully this should get me started ... and on my way !"

Nigel

 

[With The Birds]

No, the frequency is constant and somewhat irrelevant in a PWM system. Don't confuse this with it being random though. Most common r/c stuff uses a 50hz refresh rate.

The PWM is pulse WIDTH modulation.

Don't confuse this with duty-cycle either.

50hz (20ms) refresh rate was standard with the now rare PPM (pulse position modulation). Most servos will happily work with a refresh rate between 40 and 200hz. As you probably know duty cycle and refresh rate is largely irrelevant for servo control, pulse width is what matters with most servos moving to neutral positon with a 1.5ms pulse. I was interested where the OP was reading the PWM waveform, the flight controller output is a standard servo control signal. You should see a true PWM waveform on the line that feeds the mosfet gate driver IC from the motor speed controller SOC.

 

[adamlwvdc36]

The flight controller isn’t driving any ESC directly, it simply outputs a reference signal you are likely familiar with, a typical servo drive signal. The waveform at this point is the same as you will see from most RC receivers driving an ESC or normal servo (1ms =0 2ms= 100%). This signal, along with a voltage reference from the output mosfets, is used by a speed controller SOC to generate a PWM reference for a three phase gate driver IC which drives the power mosfets feeding the motors. You could probably upgrade the mosfets but would likely need to reflash the speed controller ROM, this would def be the case with a motor swap, principally due to motor inductance values. The current mosfets are good for 15A RMS (23A peak).

my idea was to use the PWM (what you call a servo signal) pulled directly off of the main board (4 per board, obviously) and have them run to completely separate ESC's.. each which has it's own feedback sensing. The power isn't really an issue because I had the thought of having one battery for the controller and possibly a dedicated battery for each ESC/motor. Maybe I'm crazy here, but to me, it seems that the controller shouldn't care what ESC it is connected to. It outputs a constant frequency signal to drive a motor controller which in turn creates a "variable" frequency (which is also the same frequency all of the time, just varying duty cycle and total "phase time" to create the same effect as different frequencies to drive the MOSFET's. The only place the frequency changes is after the MOSFET's (the effect of the VFD circuitry is a differing signal consisting of pulses of varying width to create the effect of a rising and falling differential in voltage).

So between the speed controller soc and mosfet driver? The flight controller output would be a 50-500hz square wave.

No, between the flight controller and the motor controller IC. Which is a constant frequency... However, the duty cycle changes, hence PWM (Pulse Width Modulation)
Flight Controller>---(test here)--->motor controller>---(32 bit communication)--->3 Phase driver>---(first "variable" frequency)--->MOSFET's>------>Motors

For all intents and purposes, an ESC is nothing more than a VFD. It creates a "sine wave" depending on the frequency dictated by the driver IC, through the motor controller IC, and originating at the flight controller. At low speed or throttle, there is a lot of pulses per half wave, but they have shorter duty cycles. When driving an inductive load, this has the same effect as a low frequency / low amplitude sine wave. The duty cycle always starts off low, rises up, and then falls again before each individual winding is driven the same again but in reversed polarity.

Then, when at full "throttle" this all happens again but faster with longer pulses (but fewer of them since the actual pulse frequency stays the same) creating the effect of a high frequency / high amplitude sine wave thus driving the motor faster.

I believe that if we were to hijack the PWM signal before it goes into the motor controller IC and route it to standalone ESC's, it would/should work. performing a simple battery mod would allow us to supply power to these circuits. Hell, if we parallel extra batteries to the intelligent batteries of the Phantom, we could even have the intelligent battery circuitry monitor overall voltages for us through it's UART interface.

 

[adamlwvdc36]

Just to show what can be done with a 'dead' P3 :

"I have a full Phantom 3S with Controller and batterys. It took a swim in the river and never flew again.

The choice was to revamp her as a P3S but that would only be sharing space with my better P3Pro.
The idea discussed with my pal who is a dab hand at FPV and Quads - was to rip out the dead DJI gear and replace with common Flight Controller and parts.

As far as we can tell - the motors are fine .. so that's one item we can leave. The shell is good enough .... that's another. That also allows us to use the battery holder built into the shell ...

The TX and Radio gear is to be my 9xr and FrSky DJT / V8JT ... so will be PWM based.

Parts ordered (Banggood) :

PWM to PPM encoder for on-board conversion from FrSky Rx to FC :

https://www.banggood.com/PWM-To-PPM-...p-1070269.html

Flight controller :

https://www.banggood.com/BETAFLIGHT-...p-1162891.html

4x ESC :

https://www.banggood.com/Racerstar-R...p-1105108.html

GPS with Compass :

https://www.banggood.com/UBLOX-NEO-M...-p-971082.html

I already have the combined Tx and camera from Hobby King :

https://hobbyking.com/en_us/fx718-6-...-play-kit.html

and I ordered adaptor for Rx to Android as a tester before going for full Goggles :

https://www.banggood.com/Eachine-ROT...p-1147692.html

Total price $98 .... a far cry from the ~$200 just to replace with DJI gear just to fly !

Hopefully this should get me started ... and on my way !"

Nigel

Interested. Let me know how it goes.