My setup, for me super stable at 35mph+, and excellent turning
Well, BLDC is name of the setup (BrushLess Direct Current), even the paper calls:
- Six-Step for BLDC Motor Drive
- DTC for BLDC Motor Drive So even they specify that six-step and DTC are de-facto driving approach, while BLDC is a 3 phase motor with DC -> AC inverter setup.
Now, to the inverter controller chip DRV8302 … I’ve had a deeper dig into it and unless TI botched another data-sheet (not the first time) there is not flux sensing on this chip, so it’s a pure switching driver. Fine by me, so the flux sensing is done somewhere else … for example STM ? Now, on vecs website there is a statement: " All of the hardware is ready for sensorless field-oriented control (FOC). Writing the software is the remaining part. However, I’m not sure if FOC will have many benefits for low inductance high-speed motors besides running a bit quieter. Sensored and sensorless FOC is fully implemented since FW 2.3. " which more or less indicates that there has to be flux sensing provided on the PCB … without it how will you do FOC ? Which brings us to your data-sheet where is states that for DTC mode you need flux and torque estimation, all of which to me looks like it’s done in vesc in mode called “BLDC”
If this inverter would be a pure six-step you would only have a 100% torque or nothing … unless you implement voltage regulator / current limiter before six step inverter.
The current is simply controlled by pwm and current sense feedback. Simple 6-step commutation, as stated in the paper I gave you.
Here you can see the commutation: http://vedder.se/2014/08/startup-torque-on-sensorless-bldc-motors/
It really is as I told - believe me, I wouldn’t tell if I weren’t sure 
DTC is ab early approach, and FOC is the “full distance”. Benjamin did this, and implemented all features (current sensing for 2 phases) beforehand, knowing that there is the possibility of using it.
The cortex m4F is quite a powerful controller, and enables this method of driving. It was limited for vesc 4 in maximum RPM, and is now further expanded (as many already hit the limit).
If you buy a BLDC controller, and nothing else is stated, then it is 6 step 120° commutation. Easiest driving method, straight forward. Works considerably well with low effort. Think of Brushless cooling fans!
DTC is used in cheaper controllers, and will do the job well enough for most cases - I think you know it well. FOC on the other hand is the next step, but need more expensive controllers, and may be not worth the higher effort and costs.
Best regards
Markus
1 Like
Well, most of stuff you say makes sense … still it sound as a crying shame that this controller was not started as DTC in the first place … a straight jump to FOC seems like a jump from inventing the wheel to 4 wheel drive with torque vectoring (bypassing the invention and refinement of differential).
[edit] that’s my lame “you are right” [/edit]
Side note: actually it’s the first time I heard six step, all the time for that time of control term used was “soft commutation” or just “switcher”.
Now in regards of FOC on vesc … most of folks state not to use it … what is the problem from your point of view ?
Side note2: I’ve seen around here that lot’s of hub motors and motors in general here have problems with heat rejection, at the beginning I was a bit surprised, but if all of those run in six-step (switcher) mode than there is lots of power being converted into heat inside the motor … which make more sense now.
The problem with FOC is the higher speed of switching! This causes a lot of noise on the supply lines. This wouldn’t be a problem at all, if the safe voltage range for the DRV was a bit higher. Due to very high power, people tend to use the higher battery voltages to reduce (battery) current. But this reduces the distance between the absolute maximum voltage and the supply voltage. With the higher switching speed, and high RPM more and more voltage spikes hit the DRV, which slowly (sometimes faster) degrades.
Problem is that this remaining voltage difference is so small, that crow-bars won’t work any more. There aren’t so sharp ones available (in small packages/designs). What helps, and these are the more “reliable” vescs, is adding higher capacities here and there, especially in the DC link. These should be high capacity AND low ESR one’s, that they can react fast enough to smooth out the spikes. The more the manufacturer keeps on the original design (maytech for example), the more likely it fails. Other manufacturers just use higher rated parts. But the most problematic one stays (DRV). So putting in better parts only shifts the limit, but only in small quantities :-/
The VESC hits a limit here, and unfortunately the DRV is not available in a more voltage tolerant variant 
20 V more would do the job!
And regarding the hub motors - it’s not the problem of six step, it just creates more torque ripple. It’s just the absurd amount of current you put in 
These are airodrive motors, they are intended to be run with a large “fans” in front of them. Forcing air through the windings! And now we only rely on the circulation by turning the motor. No wonder it heats up! Even six step is quite efficient.
1 Like
… hmmmmmm I actually checked the interweb and there are not many integrated drivers with reasonable voltage that quickly pop up in search. How about designing a custom driver & inverter stage ?
sied note: if voltage is the issue, we can always go for “unrealistically overkill transistors” and just drive the current curve for longer before we have a white flag day and jump higher in voltage.
side note 2: What about the controller chip, the stuff on vesc is stated somewhere in data-sheet that it can do dual FOC control at ~45kHz (STM32F4xx). This could surely drop the final price for driving whole axle (and using common ground for 2 independent motors does help limit whacking the battery with high current noise)
We - a team of work colleagues - work on just such a controller. As we try to drive 10 kW, we had the white flag day in theory. Up to 250 A at 60 V are the goal. It just consists of a beefier power stage, but these high currents are quite a challenge! Every little resistance in the path kills voltage, creating a source of high thermal energy…
Cooling transistors with such high currents is a big challenge. Even small pieces of not super nice conducting material rises the temp to a point of no good. So we are on the way. Watercooling seems to be a good thing. Even for the motors. If you want to have high peak performance, you have to cool excess heat. So we can go with smaller motors, and peak high torque 
Regards Markus
none of inverters that I work with come without water cooling 
thou on electric skateboard it’s a no go. Just a random shot, but have you guys considered potting it ? we were looking into potting some of power electronics to make it dissipate over the large surface of the box but we managed do fmea out of it 
(but to be fare it was not motor power switching, just some other power control with constant ON states - like a solid state relays, very little loss) as a result we’ve found some potting resin material from 3M with K factor of ~7 (if my memory server me right) … it’s far from perfect but damn better than air !.
goin back to your solution … 250 apms … I’m not sure whenever this is your every day skate board material … good thing about keeping below 100A is that wires are still acceptable for miniature application and you actually have an option of fusing the damn thing … beyond 100 apms you are entering the crazy world of gigavac etc and as much as I hate to put HV into places where it shouldn’t be, he I would say HV is your only option.
side note: of course you could be really mad and want to go that current, so for inverter stage I would suggest using split pi … I came across this guys some time ago … unfortunately they did not had any inverters that we wanted but a very interesting bit of kit that they have
it won’t be a e-skate, but it uses some parts of it! I’m here because of news for the speed controller and experience for wheels.
Because actually we’re going for 4 wheels drive. The front will consist of 2 hub motors - pretty easy to drive - and the back with 2 stronger ones. So actually we’re going to have two of those motors!
Batteries are ready for 500 A continuous, and we have some stuff already for all the security, such as high current fuses 
We have background from our job for such stuff, so it was rather easy to get it.
Our goal was to overshoot a bit on the power side - and I think we managed that. Better safe than sorry!
Even though it is not a eskate - I can show the final solution when we are finished! Basically all things could be part of an e-skate, a powerful one 
Was just on Carvon’s website and you can’t but the motors by themselves anymore currently… Does anyone know if this will be available again… If not, does anyone have an extra set they would be willing to sell? I have been saving for these for awhile and now they’re gone I’m feeling crushed…
Hey @uigiroux there are no plans of selling the motors by themself atm. I am unsure if Jerry will open up a group buy again since we just had one on December. You can ask him through instagram @carvonskates and check with him.
… god damnit… I was looking forward to getting one of these more than anything this whole last year for my build… of course right before I can buy one they are no longer on sale. I wish there had been a post saying that they will only be available til ____. ;(
lol I should get a free set for all the friends of mine who I ranted and raved about how great these things are, if they had one of those refer a friend and get $100 I could have bought enough for a 4WD!
I am getting a 4wd XL drive am going to assemble my own motors and Everything, for future reference so I don’t have to bug Jerry all the time. @uigiroux it’s a one man army Jerry allways busy.
wait until the preorders are delivered. some may have extras, myself included. just be patient.
1 Like
I have an XL set on order that I am not even sure what I’m planning on doing with it!
2 Likes
Ok cool, well once yall get them let me know if you wanna sell them I’d really appreciate it! Thanks guys!
were you looking for the speed drive or speed drive xl?
1 Like
Either one would be ok for me, probably the SD, but I’d be happy with the XL just as much in sure, just hate playing another extra $100 for nearly the same thing… Lol.
Upgrade with lower kv so more torque