The DRV chips are a sort of “all-in-one” gate driver chip. In fact the vesc basically started out as a DRV8302 development board with an STM32 processor and shunts added so it could be more precisely controlled.
The older generation includes a 60v to 5v buck converter to help run the peripherals and logic stuff using a second 5v to 3.3v converter. Coincidently they always seemed to blow a hole right where that buck converter is.
The new ones like the DRV8323RS have an adjustable buck and current sense amps all built in. They are also nearly impossible to get right now.
What most “3rd gen” vescs seem to be moving towards are independent gate drivers. Three individual gate drivers to switch the mosfets, a separate buck and separate current sense amps (on designs that need it).
I think there are some cost savings using one chip vs several, but it also locks you in to something that may not be available forever. The independent gate drivers seem to be more reliable too, as I have yet to see one fail, but time will tell.
DRVless means you have to replicate all the functions of the DRV with separate components. So the functions of gate drivers, current sense amps, and buck converter have to be figured out separately. A common way of doing this would be to use 3x half bridge gate drivers, 3x current sense amps like the INA181, and buck converters of choice to create 10V to 15V for the gate drivers and then another stage to create 5V. This is not technically “discrete” since it still involves using ICs rather than a bunch of individual resistors, caps, and transistors. All of this will be outlined in the CFOC3 schematic
Yeah the supply issues can be the primary reason although you can get better performance in my opinion. One can choose better individual components to serve all the functions that DRV used to. It also means component replacement is easier since replacing an individual half bridge driver or buck IC is easier than replacing an entire DRV.
Yeah…but going DRVless renders this mute. Now the ICs involved are easier to deal with an replace that the DRV ever was.
Gunna be getting back to development on this soon. I’m involved in a lot of other projects that I don’t necessarily let everyone in on. Hence why eta for this is undefined
Something I always kind of wanted to do with the CFOC2, but never did, was to put together a sort of flow-chart for debugging failed units. Maybe that’s something we could do with the CFOC3. Maybe a CFOC3 wiki?
Put together another prototype since I pretty much trashed the first one doing all kinds of surgery to it.
Works decent but need to tune the gate drive a bit more. Mosfet turnoff is just a tad too aggressive.
While I could do all of these bodges to the rest of the prototypes, I think I need to go ahead do another run. I’d rather not give beta testers anything too janky.
Tuned the gate drive stuff a bit better and now get appropriate overshoot. There’s always gunna be some amount of it and that’s why you spec MOSFETs with a voltage rating about 25% above your operating voltage. These MOSFETs are 100V rated.
In the turnoff waveform, you see a spike up to about 68V which is an 8V spike. That’s just fine imo. I like anything less than +10V for up to the max rated motor phase current.
I did up 100A with success. I don’t want to go higher until I have round 2 of these prototypes. But things are looking really good!
The next round of prototypes are done and about to ship to me. Then I can continue progress.
FYI, another secret project of mine that uses similar architecture and components just passed a road test at 140A motor current with a 16S battery on my ebike. I expect similar performance for the CFOC3
Just a matter of gate drive tuning now. Was hitting some good milestones on the bench. Tweaking Rgs and Cgs values to get those good rise/fall times. Stability is priority!
MOSFETs exhibit slight ringing on the gates during turnoff. Not used to seeing this specific occurrence. Might have something to do with the older SGT architecture of the MOSFETs. It’s not beyond control but just interesting to me.
Hey, this thing looks exactly like something we would need right now!
Looks awesome!
I have two questions: Is it going to have support for uart to bluetooth? Pwm input at the same time?
Everything is going to be controlled via vesc tool?
So it’s like most VESC based controllers. It has the same features as as your typical VESC but without built in bluetooth. You’ll have to add your own BT module
Hey there, can you already tell me what FET’s and Fet driver you are using for the cheap focer 3? I want to redesign the version two because I need a little more power. Or do you guys have any recommendations for 20s >100A controllers?
