The difference between the two designs is mainly phase shunts (75/300) vs low side shunts (100/250)… and in a perfect world where all other problems are either solved or the same, the phase shunt versions should be better. It’s also worth mentioning that those are reference designs. Companies like Spintend, Flipsky and MakerX all put their own spin on it… with different components and layouts… and that can make or break a controller.
Turns out reliable VESC hardware design is pretty complex. A motor controller is a simple concept, but there is a crazy amount of small details needed to make these things work, especially when running that much current through a relatively small device. Lots of seemingly innocuous choices that the designers make can have huge unintended consequences down the road. Then you have to factor in all the things the controller will experience in it’s life. Frayed and shorted phase wires? Sensor wires shorting against the motor windings? High voltage traveling across the canbus? Power wire touching a heatsink?
It’s really interesting watching the hardware builders on the VESC discord. They blow up a lot of controllers, but learn a ton in the process and come up with some really cool solutions. There are also a lot of really talented designers throwing out ideas along the way. Unfortunately the normal big name makers don’t share their VESC build logs with it’s successes and failures, so it’s hard to know what is going to work well, and what isn’t. Usually all we get is a few pictures and a promise.
I think this question ultimately leads to the real question - which VESC is not going to blow up on me? Truth is I don’t know. The DRV wizard said it best. “From Experience I can positively affirm to all of you that I’ve seen on my bench for repair 99% of all type of Vesc base controller!”. That has been my experience too. They all blow up in different ways. Some of it is firmware related, but most of it is small flaws in the hardware design.
I think each controller should be evaluated on it’s own in terms of reliability and cooling… two very important things that we have little data on. From the evidence, it certainly seems like the Uboxs have some unique issues as they all seem to fail in the same way… a fiery death. It could be something as dumb as the positive wire shorting against a heatsink… or the high current pos/neg traces having too tight of a tolerance for reliable PCB manufacturing.
I recently had a great conversation with a design engineer who worked on the ESC design team for a big brushless power tool manufacturer. Turns out they also use DRV8323 chips… and also have quite a few problems with them. I brought a bunch of different VESC controllers with me and we were going over all sorts of pros and cons of the different designs. I brought up one specific failure mode that I’ve been racking my brain on how to solve and he just laughed and said yup… you and the entire R&D team at my work.