Interest in a higher voltage ODrive for larger CNC conversions/builds?

I am building a PMDC powered 4th axis for my machining center. I bought an old stepper motor powered 4th axis and retrofit an odrive brushless motor and optical encoder.

I’m thinking about whether it would make sense to design an Odrive hardware variant that makes integration with machines like this easier. My VMC (like many designed for big fat brushed servo’s) has a 90V DC bus that gets stepped down from mains to power all the drive servo’s. Most of the brushed machines I’ve worked on are similar (70-120v). It uses differential signal encoders for all the axes. In the case of this machine, the motion controller outputs a digital velocity command. Most machines use a +/-10v analog signal for velocity control.

The current Odrive-HV could be used for this but putting in an additional 48v power supply is a bit inelegant especially since the machine already has a beefy DC bus (wiring the differential encoders and control signal up is inconvenient but doable). A variant like this would make retrofitting full scale CNC machines with brushless motors a lot easier since you could reuse the existing power system and connect easily to the original control (or convert to LinuxCNC). Although overkill it could even be used for brushed motors. I wondered if anyone else was considering similar projects.

One question is what max voltage would be most popular. Seems like 120V would cover most older (up to ~mid 2000s) industrial CNC machines that originally came with brushed DC servos. Newer CNCs that use AC servos tend to have even higher DC bus voltages (some over 300V). Supporting that is likely less practical, and maybe unnecessary anyway since the newer machines are high enough performance that converting them is usually a major downgrade.

1 Like

Yes I too would love a 350V variant for industrial servos. But the problem is that ODrive uses a monolithic gate drive / SMPS / current sense amplifier from TI that only tolerates up to 60V Max. There is another chip in the range that goes to 100V but that’d be your lot unless you completely re-designed the power stage.

1 Like

Somehow I had missed the existence of the STMBL project, which seems pretty suitable for this application. It goes up to about 400v and is mostly designed to drive common industrial AC servo’s. The project has sadly stalled over the last few years due to the IGBT they designed around going EOL. Worth taking a look at for larger CNC projects though. Not sure how I missed this in past internet trawling.

Yeah the DRV8301 provides gate drive, LDO step down and two channels of current sense. The DRV8353 is a similar device that goes to 102V. For my particular application that is enough but that isn’t a huge improvement over the existing design, but from a schematic standpoint is an “easy” change. Bump FETs and caps up to 100v capable parts… re-layout the PCB.

2 Likes

@crinq is around here somewhere if you want to talk STMBL

Would love to hear if folks have comparative experiences with STMBL vs Odrive. Mostly different communities I think but similar goals! I am not using LinuxCNC on this project although I’ve done a few conversion builds with it in the past.

It looks like the heavily anticipated Odrive V4 will use the TI DRV8353, despite only being rated at 58v. I presume to take advantage of the three channels of current sensing.

Not strictly related to odrive but I designed a 600V servo drive for micromod controllers, one of which is ä the same stm32 that odrive has.

I was thinking about trying odrive firmware on it eventually. Just never got that far with the component situation being what it is.