Page 28 of 31

Re: ESC32 Q&A

PostPosted: Tue Nov 10, 2015 9:30 am
by Manu
Pixhawk outputs @400Hz:
Idle 1042-1043us
Armed: 1110us (or above)
max: 1950us

Esc32 config:
PWM_HI_VALUE 1950
PWM_LO_VALUE 1042
PWM_MIN_START 1100
PWM_MIN_VALUE 900
PWM_RPM_SCALE 5200

I am a bit concerned about what will happen if temperature changes if we need to get absolutely same PWM timings between autopilot and ESC. Changing PWM_LO_VALUE 1042 to something lower make ESCs never get the solid green led fixed, which I find very disturbing. Is there a way to check the source code?

I also removed the safety switch on the pixhawk which cut PWM signal to the esc in order to be sure that we always send something to the ESCs. I had some strange startups when using antispark plugs, maybe it could be useful to delay the initialization of ESCs to be sure that autopilot was powered before?

Re: ESC32 Q&A

PostPosted: Tue Nov 10, 2015 12:12 pm
by Manu
Ok, so I increased PWM_LO_VALUE to 1045 and now it is ok. ESC cannot start if throttle is above PWM_LO_Value. I was expecting a sort of deadzone between PWM_LO_Value and _Min_Start, and do not exactly understand the purpose of PWM_Min_Valid if PWM_LO_Value overrides it.

1042 vs 1045us is less than 0.3% change, I would not expect PWM settings to be so tight...

One other point regarding Close vs Open Loop. I made a calibration with success and it seems ok on my thurst stand.
ESC32_v3.3.1_cal_R2V_15A.pdf
(389.37 KiB) Downloaded 354 times

Then, in my office I removed props to check every motor/esc direction and verify PWM inputs (read above :D). Without props of course Close Loop is off, but it make ESC fall into error (red blinks, bad motor behaviour). This close loop mode looks sexy but more details on what will go on in flight could make users more confident :)


Last point: current measurement offset. This is a very nice thing, and I believe that hardware is good. However when ESC is running at low loads, I see negative current compared to the 0.00A at initialization with motors stopped. It is a bit weird and not a big issue but maybe it can be corrected in firmware (dunno the hardware schematics, but maybe a voltage drop somewhere). If it helps, jumper is set to Vbat.

As you can see, Vbat is sensed at 28.68V which is far from 24.2V. Other ESC readings: 24.1V - 24.23 - 24.35V.

My 2 cents ;)

Re: ESC32 Q&A

PostPosted: Mon Nov 23, 2015 10:50 pm
by Manu
Any comment to this? Is there something wrong with voltage readings, a resistor to check somewhere maybe?

Re: ESC32 Q&A

PostPosted: Wed Jan 13, 2016 12:36 pm
by TobyOctane
Hiho,

I am Tobias Schwering from High Octane Motorsports (http://www.octanes.de). We are building a racecar each year for the Formula Student competition and use a quite elaborated CAN bus system with distributed nodes.
As weight is crucial in this competition, we use RC motors in our cooling system for 3 years now and switched to a fitting propeller last year. But so far we need a selfdesigned controller to convert CAN messages to the PWM signals the standard ESC needs. To eliminate this one (again, it weighs something!) we would like to use the ESC32.
But as our system is far simpler than a multicopter, the fan power should only be controlled by the coolant temperature. So your complex CAN protocol is actually far too much. So we would like to reduce it to its basics, would it be possible to give us some access to the code of a v3? Or are there still some v2 left?
And in case we have to use the closed source v3, are the messages the same as in the github code available?

Re: ESC32 Q&A

PostPosted: Thu Jan 14, 2016 7:41 pm
by aBUGSworstnightmare
TobyOctane wrote:Hiho,

I am Tobias Schwering from High Octane Motorsports (http://www.octanes.de). We are building a racecar each year for the Formula Student competition and use a quite elaborated CAN bus system with distributed nodes.
As weight is crucial in this competition, we use RC motors in our cooling system for 3 years now and switched to a fitting propeller last year. But so far we need a selfdesigned controller to convert CAN messages to the PWM signals the standard ESC needs. To eliminate this one (again, it weighs something!) we would like to use the ESC32.
But as our system is far simpler than a multicopter, the fan power should only be controlled by the coolant temperature. So your complex CAN protocol is actually far too much. So we would like to reduce it to its basics, would it be possible to give us some access to the code of a v3? Or are there still some v2 left?
And in case we have to use the closed source v3, are the messages the same as in the github code available?


Hi Tobias,
It you not want Do use a small microcontroller (to measure the temperature, connect its UART to the ESC32Vx and use CLI unterface to controll RPM of the fan motor) use CAN; V2 messages are valid for the V3 as well. --> you can send a CAN message to the note to change it's RPM.
Joerg

P.s. it you have a spare UART ob your host you can use it too.

Re: ESC32 Q&A

PostPosted: Fri Jan 15, 2016 10:39 am
by TobyOctane
aBUGSworstnightmare wrote:It you not want Do use a small microcontroller (to measure the temperature, connect its UART to the ESC32Vx and use CLI unterface to controll RPM of the fan motor)

The Problem is that all of this has to survive reliably in a racing car. So we need an enclosure, sealed connectors usw for the microcontroller, so the current solution adds up to more than 100g. And as a team member constructed new brake calipers to save 50g compared to a bought solution, this is quite a lot! ;)
Thanks for the info, then we will try to implement the autoquad can protocoll in our system.

aBUGSworstnightmare wrote:P.s. it you have a spare UART ob your host you can use it too.

thanks, but with an ignition system nearby i would not really trust uart, CAN was always reliable so far.

Re: ESC32 Q&A

PostPosted: Mon Jan 18, 2016 1:33 pm
by rput
Hi,

I have a quick question: Is it possible to control the ESC32 using pwm, yet still read out RPM info from CAN at the same time?

Re: ESC32 Q&A

PostPosted: Mon Jan 25, 2016 8:22 pm
by rput
anyone?

Re: ESC32 Q&A

PostPosted: Mon Jan 25, 2016 8:28 pm
by sandmen
no. It's not possible.

Re: ESC32 Q&A

PostPosted: Mon Jan 25, 2016 9:26 pm
by rput
ok, thanks