Nah it still has performance issues with algorithm at low speeds. I also really hate the control mode transitions, there is no cross-checking between speed estimators.
I’m currently trying to figure out the details of getting the most accurate current samples to avoid those same issues.
I just get HELLA triggered when people use IR garbage.
The Trampa VESC 6 enclosure has its own thermal resistance case to ambient of around 10°C/W, Both of these MOSFETs’ thermal resistance junction to case is below 1°C/W. The difference in thermal conductivity is negligible.
From my preliminary tests, the VESC 6 seems to have similar thermal performance to the Flipsky 6.6 with enclosure at 30kHz switching FOC 80A motor. Not sure why you would’ve had worse performance with Onsemi FETs. P/N NTMFS5C628NL.
EDIT: Here’s one of the reasons why the IR FETs SUCK
Idk if I am doing something wrong, but I can’t even get the thing to turn off, the LEDs stay on.
I used the newest firmware version, set the timeout to 30s, left it unplugged to anything except the battery and let it sit for an hour. LEDs stayed on. The green LED was dimly lit, which either means it never turned off, or the MCU is periodically waking up from sleep mode just to toggle the PWM on the LED. The green LED is not connected to any PWM capable output pin on the STM32F405RG.
Now, if the LEDs stay on even if the ESC has shutoff, well, then that means the current draw is really high.
I got my mk3 and tested with 15V power supply and 30V power supply, and I gotta say, impressed!
With 15V the current went to 16uA within one minute of disabling Vesc, with 30V it took 3 minutes to get under 20uA, but as soon as Vesc goes to sleep it jumps from around 200mA to 30uA.
I’ve been meaning to test this but school has been kicking my butt. Thanks for this, I’ll try to verify it this weekend.
Still not sure if I want to cut the rubber covering to get a look at the bottom side, but I can probably guess what they did on the power supply. That shutdown current will go up with age, as the electrolytic capacitor leakage current increases, and I don’t think every single unit will make the cut, since manufacturing variation will have some with higher leakage currents as well as higher buck converter shutdown currents.
The VESC 6+ schematic is not the mk3 schematic, I’m guessing they have some sort of SR-latch circuit that drives the buck enable pin. MCU drives set, off-chip comparator circuit drives reset.
So I finally got around to testing it on mine. It stabilized to around 20uA at 36V after a few minutes. Guess Eletrolytic leakage rarely gets close to maximum datasheet specification? (every datasheet specifies worst case being in the hundreds of microamps at these voltages).
It appears that the POWER_SWITCH input directly controls the power state of the ESC, meaning you need a latching pushbutton and it does NOT have smart turn on and turn off. I’m guessing this input is wired directly in an open drain fashion to the buck converter enable pin.
(cue latching push button horror stories).
This also means the positive and negative rails are still live inside of the ESC.
When I get around to implementing shutdown on my ESC, I’m putting the MCU in stop mode for fast wake-up and smart features. I personally get annoyed by the long boot time.
“Discovered the answer to my question…the push button for powering on should be a momentary button.
The overall problem I was having was my MK3 came loaded with VESC 6 “60” bootloader so the firmware wasn’t updating with the correct “60_MK3” bootloader. Now the momentary button starts it right up.”