You are here

Custom 12S VESC seems to be power limiting or something

5 posts / 0 new
Last post
nickw1881
Offline
Last seen: 4 months 3 weeks ago
Joined: 2018-01-19 10:15
Posts: 17
Custom 12S VESC seems to be power limiting or something

My custom vesc has arrived! And it works pretty well and stays cool even when delivering 2.5kW!! But when I throttle up to 3kW or so, the VESC software seems to be limiting itself. It does it around 50V/95% throttle or a little lower throttle at higher voltages. Here is a video:

<iframe width="1068" height="601" src="https://www.youtube.com/embed/1_3IwvQKiv4" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>

https://www.youtube.com/watch?v=1_3IwvQKiv4

I started the test at 45 ish volts and ramped up the throttle to 95%, then turned up the power supply voltage to 50 and you can hear the controller throttling back and start surging, which is when I cut power. I’m sure its some setting in the software. I tried most of them, but I am a little lost as far as what they all do. I could use a little help.

  • BLDC mode
  • Current limits in hwconf 4.10 and board.h are set to 200A (sensor can only sense up to 165A)
  • max erpm is set to 200,000
  • max voltage is set to 61, sensor can only sense up to 61.8
  • no faults are reported
  • power supply current limit is set at 150A input (and its the kind where it backs off the voltage, dont see that)
  • haven’t messed with the timing advance, not sure what the process is for messing with it.

20180806_192716.jpg

20180821_003036.jpg

josh
josh's picture
Offline
Last seen: 1 year 3 months ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

Hmm, that seems odd. Could it be something with stator saturation? It seems like it loses the position of the motor and begins cogging. My guess is something with the back EMF and positioning. Maybe try and sample during the event through the VESC tool?

Cheers!

TechAUmNu
Offline
Last seen: 1 day 12 hours ago
VESC Free
Joined: 2017-09-22 01:27
Posts: 514

There are a lot of limits for power, erpm, duty cycle, currents, etc. It is likely one setting was missed. Otherwise it could be losing motor tracking. I had issues losing motor tracking at high speed high current and it blowing up a lot of hardware. Which turned out to be my current shunts being covered in solder so the current was actually 2-3 times what it was measuring. 

 

What happens when you slowly increase the duty cycle with the voltage already at the 50v it fails at?

 

TechAUmNu
Offline
Last seen: 1 day 12 hours ago
VESC Free
Joined: 2017-09-22 01:27
Posts: 514

Just watched the video. It is definitely either losing tracking or hitting ERPM limits. Are you running BLDC or FOC? What is your switching frequency set to?

What is your motor current when this happens? Is it over the limit for the motor? Have you tried using stator saturation compensation?

Other possibility is deadtime compensation is wrong for your power stage, I see you are using discreet gate drivers, what is your calculated deadtime and have you set it in the firmware?

You need to set both deadtime in firmware and, deadtime compensation in software if you are using FOC.

nickw1881
Offline
Last seen: 4 months 3 weeks ago
Joined: 2018-01-19 10:15
Posts: 17

Hey TechAUmNu. Your 200A ebike project is beautiful. I can't wait to see the latest version of it.

Think I figured it out the problem and the screeching. With this motor the measured peak phase currents are getting up to ~3x the average phase currents. I was hitting my absolute current limit and the VESC was cutting out like it's supposed to. This is reasonable behavior. With "filtered current" on, the VESC still seems a little quick to pull the plug, but I suppose it's a fine line between quick shutoff and blown up boards so I'm rather safe than sorry.

The screeching was throwing my head, but tell me if this sounds plausible: In the video I am sending throttle%-commands via serial port at 5Hz. When the motor cuts out from absolute-over-current it loses tracking, then it is commanded back to 90% throttle via the serial port at 5Hz, and the macro-effect is screeching. I have since switched the spam throttle%-commands to on-change current-commands and it seems to work much better. Throttle%-commands may be a poor way to control the VESC.