High Power VESC

Discuss hardware related to the VESC such as the NRF nunchuk.
User avatar
TheFallen
Posts: 84
Joined: 26 Oct 2016, 10:49
Location: UK
Contact:

Re: High Power VESC

Postby TheFallen » 21 Jan 2017, 18:07

rew wrote:Where did you get that funny idea that the current will only be going one direction? You need a bidirectional current sensor even when it's in the low-side-fet-return-path.


I thought that as it was essentially connected to ground you'd only have current flowing one way? My modifications to the maths should be safe for bidirectional current so this is more of a learning point for me.

Anyway, I've been struggling with a massive problem for the last week. Phase C is weird.I would connect up the phase A drivers, phase B drivers,phase C drivers, phase A current sensor and phase C current sensor, power up he ESC and I'd hear a buzzing noise. Playing 'remove wires until it goes away' (and powering down when removing wires) I found it was entirely due to the phase C gate driver circuit. I could put any phase gate wires into the driver circuit and it'd would buzz. Probing the gate driver supply found it would not provide a stable +12V supply.

Image

So as you can see what should be a +12V rail drops to +5V which is pretty bad. The weird thing is if I disconnected the the MOSFETs from the phase C driver then the rail would stabilise at +12V. The +12V switching power supply is rated for 500mA and the 3 gate driver chips theoretically only pull 2.5mA even when charging the boost capacitors so it shouldn't be an overload condition.

I continued on thinking it had to be a short. Of course without the MOSFETs attached I couldn't find a short between any of the pins or between the pins and ground or power. This confused me somewhat so I decided to measure the voltages on all the chips while they were powered but inactive. I found:
Image

  • VDD is the +12V rail, so is fine.
  • HB has a difference of one diode drop so is also fine.
  • HO is where things go different, phase C is ~4.00V higher than phase A & B.
  • HS needs to be the same as HO to ensure the high side is off, but is ~4.00V higher than phase A & B.

De-soldered the offending UCC27211A a douple of times without any change, also tried changing the boost capacitor between HB & HS but that had no effect either. I could see any shorts that could have caused the problems before or after the soldering.

In short, I'm completely stumped.

arvidb
Posts: 234
Joined: 26 Dec 2015, 14:38
Location: Sweden, Stockholm

Re: High Power VESC

Postby arvidb » 21 Jan 2017, 19:29

Without really knowing what I'm talking about:

  • What's the voltage HB-HS supposed to be? Isn't ~6 V (as in phase A & B) a bit too low to drive the gate of a normal MOSFET?
  • I'm guessing your measurements are without a motor connected (since you've got ~4 V between the C phase and the others)?
  • Can the voltage difference between the HO's drive any current? Edit: Sorry, between the HS's, I mean.
  • With all MOSFETS turned off, the HS-HO-HB part should be floating? Edit: With respect to Vdd, that is. It should be referenced to the motor rail positive side? Still weird though that you get a consistent difference on the C phase...

Edit again: From the datasheet 8.3.4: "With the VHB capacitor connected to HB and the HS pins, the VHB capacitor charge is refreshed every switching cycle when HS transitions to ground." So with no motor rail connected, the voltage at HS (and HO) is simply Vdd - boot diode drop - VHB (voltage of boot capacitor). Maybe you've been chasing a red herring?

arvidb
Posts: 234
Joined: 26 Dec 2015, 14:38
Location: Sweden, Stockholm

Re: High Power VESC

Postby arvidb » 21 Jan 2017, 20:04

Another idea: pull HS to ground (through a resistor) to manually charge the VHB cap. What happens?

User avatar
TheFallen
Posts: 84
Joined: 26 Oct 2016, 10:49
Location: UK
Contact:

Re: High Power VESC

Postby TheFallen » 21 Jan 2017, 20:20

arvidb wrote:Without really knowing what I'm talking about:

  • What's the voltage HB-HS supposed to be? Isn't ~6 V (as in phase A & B) a bit too low to drive the gate of a normal MOSFET?
  • I'm guessing your measurements are without a motor connected (since you've got ~4 V between the C phase and the others)?
  • Can the voltage difference between the HO's drive any current? Edit: Sorry, between the HS's, I mean.
  • With all MOSFETS turned off, the HS-HO-HB part should be floating? Edit: With respect to Vdd, that is. It should be referenced to the motor rail positive side? Still weird though that you get a consistent difference on the C phase...

Edit again: From the datasheet 8.3.4: "With the VHB capacitor connected to HB and the HS pins, the VHB capacitor charge is refreshed every switching cycle when HS transitions to ground." So with no motor rail connected, the voltage at HS (and HO) is simply Vdd - boot diode drop - VHB (voltage of boot capacitor). Maybe you've been chasing a red herring?


  • The IRFP4468 only requires a VGS of +2.0V to +4.0V to turn on. But as VS = VG that means VGS is 0V so the high side MOSFET is most definitely off.
  • I didn't have a motor connected at this point, I did later and found it permanently pulled the +12V supply down to +1.2V, very much unwanted.
  • Not too sure what you mean here?
  • HO will 'float' with respect to HS. To turn on the high side MOSFET HO needs to be >4.0V higher than HS, to turn off it needs to be <4.0V higher than HS. This means that if the MOSFET is on and HS is then VCC (minus MOSFET losses) then HO needs to be >VCC+4.0V. This is where the boost capacitor comes in. Ground is still ground so I can use a multi-meter to give me sensible measurements.

This may be a red herring, but 1 of 3 identical MOSFET gate drivers doesn't work, and that gate driver has funny voltages on its pins. I'm not sure else where to look. I shall have a poke about grounding that capacitor tho.

arvidb
Posts: 234
Joined: 26 Dec 2015, 14:38
Location: Sweden, Stockholm

Re: High Power VESC

Postby arvidb » 21 Jan 2017, 21:19

TheFallen wrote:
  • I didn't have a motor connected at this point, I did later and found it permanently pulled the +12V supply down to +1.2V, very much unwanted.
  • Not too sure what you mean here?


What I meant was, that with everything off HS (i.e. the bridge outputs) should not be able to drive/supply any sustained current. You pretty much answered this with the previous point - somehow, your bridge supplies current even in the off state - current that it takes from Vdd! The question is, how does it do that :?: Hmm... Which phase is supplying the current to the motor, and which phase is sinking it?

arvidb
Posts: 234
Joined: 26 Dec 2015, 14:38
Location: Sweden, Stockholm

Re: High Power VESC

Postby arvidb » 21 Jan 2017, 21:40

What voltage was Vcc at when you did the measurements in your first post about this problem? I'm wondering if there could be a problem with the C driver that makes it source Vdd current from HS, and maybe, if Vcc was not connected but you have something else on your board that loads Vcc, that something could sink that same current? Edit: nah, strike that, then the motor would not have to be connected for Vdd to go down to 1.2 V...

lizardmech
Posts: 171
Joined: 19 Jan 2016, 10:54
Location: Australia

Re: High Power VESC

Postby lizardmech » 22 Jan 2017, 06:51

I just finished my ucc27211 based controller. It works, HB and HO remain at 1.8V when it's not running.

rew
Posts: 943
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: High Power VESC

Postby rew » 22 Jan 2017, 09:49

For static measurements, I would recommend putting a 1k resistor on each phase to ground. This makes it possible for the high side drivers to charge their boost capacitors. The voltage at "HS" would otherwise be undefined. This allows it to float to any value. Even a small leakage current could cause the voltage to drift up or down.

Double check your mosfets. When they are connected there are 12 measurements that reasonably quickly give a good indication of that they are still good: put the multimeter in beep/diode position and measure from each phase to each of the power supply leads. Now turn the multimeter around (swap red and black) and do it again. You should NEVER get a beep, and "open circuit" in half the measurements and a diode drop in the other half. My mosfets measure about 420-440mV. (I don't remember exactly, just that it was in the 400's That's good enough...)

A slighly more advanced measurement will also check that the gate is free to do as it pleases. Open circuit to the source in both directions.

I suspect your C-low mosfet might be shorted. Possibly not now, but in the past and it might have blown up your driver.

User avatar
TheFallen
Posts: 84
Joined: 26 Oct 2016, 10:49
Location: UK
Contact:

Re: High Power VESC

Postby TheFallen » 22 Jan 2017, 18:25

rew wrote:I suspect your C-low mosfet might be shorted. Possibly not now, but in the past and it might have blown up your driver.


Thing is I'm on the 3rd UCC27211A and I have been running most of the single phase tests on the phase B MOSFETs. I'll definitely have a go with the diode tester.

User avatar
TheFallen
Posts: 84
Joined: 26 Oct 2016, 10:49
Location: UK
Contact:

Re: High Power VESC

Postby TheFallen » 25 Jan 2017, 21:38

rew wrote:Double check your mosfets. When they are connected there are 12 measurements that reasonably quickly give a good indication of that they are still good: put the multimeter in beep/diode position and measure from each phase to each of the power supply leads. Now turn the multimeter around (swap red and black) and do it again. You should NEVER get a beep, and "open circuit" in half the measurements and a diode drop in the other half. My mosfets measure about 420-440mV. (I don't remember exactly, just that it was in the 400's That's good enough...)


OL and ~430mV on all 12 MOSFETs. Goodstuff!

Went round testing the PSUs but seem to have broken the +5V in the process!

VBat = +20.64V
+12V_PSU = +12.02V
+5V_PSU = 0V
+3V3_PSU = +3.310V

Also tried shorting out +12V_PSU with a 22Ohm resistor, in an attempt to replicate the oscilloscope trace. I noticed that the voltage dropped to +5V similar to the oscilloscope waveform so it does look like something around the phase C driver was shorting it to ground.

I couldn't do anything more as the ACSxxx current sensors run off of the +5V_PSU and with that not working I'm a bit stuck.


Return to “Related Hardware”

Who is online

Users browsing this forum: No registered users and 1 guest