VESC Overheating ?

Report your hardware problems here. Include as many details as possible with pictures to get the best possible help and feedback.
csomerv
Posts: 4
Joined: 01 Sep 2016, 16:15
Location: Cheltenham

VESC Overheating ?

Postby csomerv » 17 Mar 2017, 16:24

Hi I have an application with a small pump being run be an alien motor and a vesc.
The VESC / BLDC autodetect was used to get initial parameters and the motor itself runs smoothly.
Motor 130 kv Alien sensored motor
Supply 36 volts ( Currently via mains lab supply)
Integrator limit = 155
BEMF Coupling = 650

[img]
VESC%20Config.JPG
[/img]

At full pump speed we have typically 40 amps motor current as shown on the BLDC tool and 80% duty cycle, within a short time we hit 80 degrees on the FET temperature.
VESC Run .JPG
VESC Run .JPG (142.18 KiB) Viewed 1344 times


Considering the fact that the VESC is used for a few (quite a few) Skateboard designs where the user is on the board up and down hills something must be wrong if we can only achieve a few minutes run time before the VESC overheats.

I have run it in sensored, sensorless and hybrid modes. Obviously sensorless mode does not work too well at low speeds.
VESC Motor.JPG
VESC Motor.JPG (123.76 KiB) Viewed 1344 times



Any thoughts would be appreciated as this cannot be correct as the specs for the VESC seem way above what we are actually using ?

Brgds

Craig
Attachments
VESC Config.JPG
VESC Config.JPG (117.43 KiB) Viewed 1344 times

eTurbo
Posts: 4
Joined: 04 Mar 2017, 21:12
Location: Florida, USA

Re: VESC Overheating ?

Postby eTurbo » 18 Mar 2017, 19:41

Dumb question perhaps, but have you verified the 80 deg C reading? That is "burn you" hot... If the board is barely warm, then you may have other issues. Following the BOM and schematic of the original VESC 4.xx, mine reads 40C at an ambient of 20C.... (I'm not sure why - I haven't troubleshot that yet. Just assumed a wrong value or calculation was used somewhere)

pf26
Posts: 278
Joined: 28 Mar 2016, 14:37
Location: FR Valence

Re: VESC Overheating ?

Postby pf26 » 19 Mar 2017, 11:27

Be aware that 1200W is already quite a large amount of power for a longboard, or an e-bike. An average cyclist will only achieve 400W for short periods of time..
At sustained 30amps and above, I think you need to provide some kind of cooling to the VESC, because each FET, even with Rds in the low mOhm range, will have to dissipate a few watts. And this translates into 10s of degrees depending how you allow the heat to escape.

rew
Posts: 904
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: VESC Overheating ?

Postby rew » 19 Mar 2017, 12:04

The problem with the VESC-4 is that you could try to cool the side of the fets that sticks out. If you put a hefty 1K/W fin on that, the total thermal resistance will be about 16K/W as my best guess for the thermal resistance junction-to-top-of-the-case is 15K/W. Now, if we allow the junction to become 150 degrees C, and allow for a 30 degree ambient, you get a max power dissipation of 7.5W (per FET). And this is assuming a near-perfect cooling element.

Assuming zero switching losses that would mean the maximum current is about 80A. Contrary to what I expected, you might see some improvement from better cooling. Remember that it's the PCB that needs cooling, not the FETs. (the FETs pass on heat to the PCB, cooling them from the back is way less effective.

Just for completeness: The human on a pleasure-bike-trip will produce around 70W. A top level sporter in a race will do about 300-400W continuous. It's the non-top-sporters that do 400W for short periods.

csomerv
Posts: 4
Joined: 01 Sep 2016, 16:15
Location: Cheltenham

Re: VESC Overheating ?

Postby csomerv » 19 Mar 2017, 22:37

Hi thanks for the responses so far.
The vesc configuration is one from "scramboards" in the UK.. Where they have repackaged the VESC design and added a reasonable heatsink.
https://www.scramboards.com/vesc-vedder.html
Ive added a CPU style fan to the heatsink. I have checked the temperature of the heatsink and when the FETS are at 80 degrees the heatsink is considerably lower than this. That said with the heatsink bonded to the top of the FET and given the thermal resistance of the plastic portion of the case as opposed to the metal connection on the underside case I can see why the heatsink may be at a lower temperature than the FET's.
So perhaps the next step would be to reposition the fan and remove the heatshrink to allow better board cooling.
I agree 1200w is a lot , but it is within the spec of the controller..maybe water cooling :)

I had noticed than another deisner is using the directfet from IR and this may be a route to go along.

Thanks again for the input/help

Craig

artteth
Posts: 28
Joined: 22 Jul 2016, 00:28
Location: Belarus

Re: VESC Overheating ?

Postby artteth » 20 Mar 2017, 10:31

my heatsink version
very hard to solder (need powerful solder iron)
i set 300 phase amps and i can't overheat it (I'm 150 kg weight)
its notebook cooling system
solder it only on one side (where all fets are parallel connected )
if it will overheat add fan (but i can't overheat it (and can't unsolder it :D ))
Image
Image
Image
Image
Image
Image
Image

rew
Posts: 904
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: VESC Overheating ?

Postby rew » 21 Mar 2017, 07:47

Yeah! that sounds like a better cooling solution. Does not look professional, but could certainly work.

You soldered the heatsink to the top fets. It's the bottom fets that are now the trouble: the top fets are taken care of.

order-of-magnitude calculation: They have a 1K/W junction to bottom-of-case. Then 2K/W to the other FET, and another 1K/W to the heatsink. If that heatsink does 3K/W You're still way better off than a perfect cooler cooling the plastic on both sides. If that's a CPU heat-pipe, it needs to be better than 1K/W because a CPU dissipates around 80-100W.

artteth
Posts: 28
Joined: 22 Jul 2016, 00:28
Location: Belarus

Re: VESC Overheating ?

Postby artteth » 21 Mar 2017, 14:48

rew wrote:Yeah! that sounds like a better cooling solution. Does not look professional, but could certainly work.

You soldered the heatsink to the top fets. It's the bottom fets that are now the trouble: the top fets are taken care of.

order-of-magnitude calculation: They have a 1K/W junction to bottom-of-case. Then 2K/W to the other FET, and another 1K/W to the heatsink. If that heatsink does 3K/W You're still way better off than a perfect cooler cooling the plastic on both sides. If that's a CPU heat-pipe, it needs to be better than 1K/W because a CPU dissipates around 80-100W.


yes
its cpu pipe
cooling is too good so bottom fets didn't heat up too (previously i use two big radiators and thermal paste and try to cool thru fets plastic and it didn't work)
with this pipe ewen without fan its super cooling performance (those pipes in most cases with fan)

Vanarian
Posts: 9
Joined: 23 May 2016, 11:27
Location: Nissa

Re: VESC Overheating ?

Postby Vanarian » 02 May 2017, 17:28

Hi, looks steampunk! Do you have a link on the type of heatpipe you used? Is it simple copper bars/tubes soldered onto the FETs or is it gaz/water reaction based inside ?

How hot is the large part of the heatsink ? What's the silicon around the wires for, transmitting some heat through them too ?

Can't a thermalpaste do the trick ?

Great job !


Return to “Hardware Problems”

Who is online

Users browsing this forum: No registered users and 1 guest