VESC cooling on longboard

[RogerD]

Looking at getting some air around my VESC on my eboard but don't want to just stick it exposed under my board. I'm thinking of a project box with large holes where the FETs are to allow airflow, and provide light "bash and splash" resistance.

Anyone got examples of what they've done?

 

[RogerD]

Cheers. I have a 3D printer and can design in 3D so I'll make a case to hold one VESC, with either a 30 or 40mm fan mounted in/in it. Watch this space!

 

[Trbt555]

You may not even need to do this.
It all depends on your voltage. Higher voltages are more efficient.
I was running 12S on one VESC during my test ride yesterday and I maxed out at roughly 24A. Barely any heating.

 

[RogerD]

Was your vesc opened air or sealed in a box?

 

[Trbt555]

Sealed in the box, check my build thread for details.

 

[RogerD]

Hmm - just tested mine. Air flow made a big difference. Still managed to overheat it.

With the esc mounted outside I could go up and down my long drive with no slow downs or cut outs, which I was getting before when it was boxed. I had to ride holding a laptop whihc was not ideal!

Temps climbed into the 70's plus. I pulled around 65amps max.

When i forced it to work, by doing slow starts up a speed bump, it soon triggerd the temp protection and the max speed was severely capped until it cooled down.

I need to work out how to save logs of temperatures rather than just seeing them real time.

Bear in mind I'm driving one large motor, and I'm around 90kg.

 

[Trbt555]

What's your battery setup ?

 

[RogerD]

10s 4000Mah 25x x 2 in parallel:

so 10s 8Ah 50c equivalent.


http://www.hobbyking.co.uk/hobbyking/store/__21367__ZIPPY_Compact_4000mAh_10S_25C_Lipo_Pack.html

 

[chaka]

I have found that the lower fets are what overheat first but it really depends how your VESC is placed. I mounted a bottom plate to a VESC with thermal paste connecting the plate to the fets and it stopped the overheating issues I was having. I then went a step further and added a heatsink to the upper fets.

 

[saul]

I have found that the lower fets are what overheat first but it really depends how your VESC is placed. I mounted a bottom plate to a VESC with thermal paste connecting the plate to the fets and it stopped the overheating issues I was having. I then went a step further and added a heatsink to the upper fets.

Where did you find that heatsink? perfect fit. 8)

 

[RogerD]

They have thermal cut off. It's configurable in the VESC sofware.

When I was testing a board with the wrong gearing it kicked in a lot, slowing the VESC down.

 

[Vanarian]

Up guys, I was actually wondering something a bit on the extreme side!

VESC FETS are rated for 240A continuous at 25 degrees and 200A continuous at 100 degrees. So if we can keep the temp under these 100 degrees for the FETS and under 75 degrees over rest of the PCB, we can safely assume that continuous power should be raised up to 8KW - 10KW (10-12S)!

Apart from looking like a useless BIG figure, it means you could look for motors rated to produce multiple KW continuously and feed them as much as you'd like without losing acceleration. Sprint climbing abilities for off-road bashing? Or something sized like a go-kart or buggy?

What do you guys think about some non conductive synthetic oil as cooling system to put the VESC inside a thermal chamber? Could be a compact and lightweight solution to drastically drop temp thus raise continuous power.

 

[WeeChumlee]

Hi Vanarian

One must take these data sheet values seen from a technically theoretical point.
They are of course correct, but to achieve the values is not quite as straightforward as it may seem.
Your idea of a thermal chamber filled with oil is not a bad one at all, indeed this is done in some applications.
The problem we have though is size and weight. If the size is to be kept relatively small then a system must be made where the oil would circulate, taking the heat from the chip to a point where it can cool. Lets assume a small sump with cooling fins.because of low volume the flow would have to be reasonably fast, this means a pump of some form. Heat pipes like one knows from a PC microprocessor would not be fast enough. The junction within a MOSFET where the heat occurs is very small, this then spreads to the case, which is not very large either. At large currents the heat is very substantial and must be carried away immediately.
That all being said, I realise you do not wish to achieve the 240A continious current and a small enclosure, with cooling fins and open to ambient air flow would definitely increase the power one could draw.
Not sure if this is practical on a skateboard but for other applications, such as you mention, definitely a possibility.
Personally though I think, taking space, cost and manufacture into consideration, a heat sink with a small fan would be an easier solution to achieve similar results.
(But hey, that would be the easy way - go ahead, push boundaries, make one)

 

[bobfandango]

Up guys, I was actually wondering something a bit on the extreme side!

VESC FETS are rated for 240A continuous at 25 degrees and 200A continuous at 100 degrees. So if we can keep the temp under these 100 degrees for the FETS and under 75 degrees over rest of the PCB, we can safely assume that continuous power should be raised up to 8KW - 10KW (10-12S)!

Apart from looking like a useless BIG figure, it means you could look for motors rated to produce multiple KW continuously and feed them as much as you'd like without losing acceleration. Sprint climbing abilities for off-road bashing? Or something sized like a go-kart or buggy?

What do you guys think about some non conductive synthetic oil as cooling system to put the VESC inside a thermal chamber? Could be a compact and lightweight solution to drastically drop temp thus raise continuous power.

The oil filled box won't help much because the heat still needs somewhere to go. The oil has a higher specific heat capacity than air, but only by maybe a factor of 1.5-2. So, it could maybe absorb twice as much heat as the air, but that doesn't buy you all that much in an enclosed space.... i.e. you would still need to be dumping the heat into the atmosphere somehow.

Per the datasheet for these FETs, the on resistance of the FET at 100A is in excess of 2 milliohms (with junction temp of 100C). But let's just assume 2mohm at 200A. That means the FET would have to dissipate 200^2*.002=80 Watts! That's roughly the same as the thermal design power of an intel i-4790k cpu! A fanless CPU cooler for that processor is enormous... Even with a fan, they can be pretty darn big. Moreover, the CPU heat is being ejected from a much larger case than a FET.

So the point is, running 200A through these FETs won't happen in a skateboard form factor, but I suppose you may be able to make it work in a buggy or something where space is not an object.

 

[randyc1]

..Ever see the Big Heatsink on the Evolve skateboards !!...that would.probably solve any heat problems.

 

[whitepony]

I ordered VGA aluminum heat spreaders with double adhesive thermal tape quite a while ago for some other things I was fiddling with - 16 pieces for 3€ or so. just slapped them onto the trampa vescs and it will probably help a little bit with potential overheating.

 

[randyc1]

Whitepony you have an Evolve correct ?, is it actually a Heatsink at the back of the enclosure for the ECS's ?

 

[whitepony]

yea, it is! its never anything but cold though!

 

[WeeChumlee]

I did exactly the same with the VESCs on my dual, the small heats sinks and a bit of thermal glue is so cheap there really is no reason not to.
I will change my single 6374 build to a VESC as well, if I can I will do a run without the heat sinks and then a run with the heat sinks - just to get a comparison.
Have to strap the laptop to the board again and run the BLDC tool. I wish there was a function that one could log things to a USB stick or the like.

 

[Vanarian]

True the size would be too big for a board, but you're right guys I was wondering how it would act inside something like a go-kart And well that's a huge amount of temp to cool!

On a more usable topic a vapor-chamber based on water could also do the trick yet remain compact as long as heatsink gets direct aeration! That's how some graphic cards are able to run mad power without melting. I'm tempted to try some experiments there, first I'll finally get some VESCs and then try after ! :lol: