case for VESC

General topics and discussions about the VESC and its development.
scienceworks
Posts: 29
Joined: 27 Mar 2016, 16:27
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

case for VESC

Postby scienceworks » 27 Mar 2016, 16:49

hi,

after i almost fried my brand new VESC while prototyping with my settup, i decided to 3d print a case until i can design a real case to be milled from aluminum.

has anybody else designed a case for an individual VESC? is it ok to encase the pcb from a thermal point of view? the aluminum case will be clamped on the mosfets and therefore be used as a heatsink, but there wont be much air circulation left in the case itself to cool other components.

might that be a problem for the DRV or MCU?

thanks!

rew
Posts: 943
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby rew » 27 Mar 2016, 20:02

It helps to know what voltage and current you are expecting. But even at the maximums, I would think that thermally connecting the fets to the case will be better than nothing.

scienceworks
Posts: 29
Joined: 27 Mar 2016, 16:27
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby scienceworks » 27 Mar 2016, 21:27

does the voltage and current even influence the heat that is produced by the MCU and DRV?

i will be running up to 12s and 2-3kW.

rew
Posts: 943
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby rew » 28 Mar 2016, 20:51

12S is about 50V.
3kW, 50V is 60A.

The current determines the heat-losses in the FETs. IRFS7530 at 1.15mOhm gives 4W per motor phase, or about 12W total. That will be in the FETs. Oh! there are also 1mOhm shunts. Two of them. So add another 7.2W from the shunts.

The FETs also incur switching losses. A "worst case" estimation is: I * Uin * t-switch * F-switch. This would come to substantial numbers, but it seems that practice is a whole lot better. There are much better approximations, but I don't know them by heart. Anyway, the voltage does play a role here (Uin).

The DRV generates the gate-current with a linear regulator. So (Uin - 12V) * Igate is dissipated in the DRV chip. Igate equals F-witch * Qtot. Qtot you get from the datasheet, and VESC software switches at 25 kHz max IIRC.

So... Current influences the heat a lot, as it mostly counts SQUARED. Voltage also counts, but mostly linear.

Benjamin says: About 50A continuous. IF you're aiming for 60A, that's more than what Benjamin deems safe. So you'll have to be better at cooling the whole thing than Benjamin. (As far as I'm concerned, the aluminium case just might do the trick). The thickness of the case does matter. Too thin and you're just adding a barrier that doesn't allow air to flow. And it doesn't help (enough) to get the heat away from where it is produced. Too thick and there will be a too big heat gradient across the case. Somewhere in between it will be best... I don't know the best thickness.....

joe90
Posts: 49
Joined: 15 Dec 2015, 16:52
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby joe90 » 29 Mar 2016, 06:49

I found that forced ventilation help a lot.
for 60A continous. you will need a lot
Also I am not able to go more that 80A phase without falling into faults
100A fall into fault easily
Many accelerations at 90A phase current bring my board to 50C
This is with forced ventilation, but not so well designed.
I think you can do it if well though out.
Michel Cote
Twin v4.7 + SK3 6374 149KV
Bicycle
2500W peak/unit on 12S

chrisbaron
Posts: 3
Joined: 23 Dec 2015, 22:35
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby chrisbaron » 30 Mar 2016, 00:42

It is best to work backwards from heat for MOSFETs.

These MOSFETs have an Rja of about 63C per Watt. Rule of thumb is about 1.5W to 2.0W per MOSFET with no heatsink. So a rough estimate:

2.0mOhm per MOSFET gives about 31A continuous I^2 * R = W thus sqrt(2.0W / .002Ohm)

Each MOSFET is on for 2/3 of each cycle so 31/.6667 = 53A

That does not include switching losses so I think 50A with no cooling is a good estimate. :ugeek:

joe90
Posts: 49
Joined: 15 Dec 2015, 16:52
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby joe90 » 30 Mar 2016, 01:05

Just want to share my findings:

Without cooling
In the bench test I have done 50A was shooting to overheat.
I stopped test at 70 Celcius but temperature was climbing steadily.
Michel Cote
Twin v4.7 + SK3 6374 149KV
Bicycle
2500W peak/unit on 12S

rew
Posts: 943
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby rew » 30 Mar 2016, 12:18

We did a similar test here. At a certain point we noticed the temperature about 75C. Then we set "max temp" configuration to 60-80, and we noticed that the board would automatically start slowing down when the temperature hit something like 70 degrees. That works! So then we were confident enough to set the temperature limits back to 80-100 and that worked fine at a slightly higher current and temperature.

(but the board we did that on has quite a good thermal connection between the FETs and the temperature sensor.)

scienceworks
Posts: 29
Joined: 27 Mar 2016, 16:27
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby scienceworks » 02 Apr 2016, 02:46

thanks for your imput.

i finished a quick'n'drity aluminum case, and it looks quite good (from a thermal point of view, the case itself looks horrible :D)

the first try was a 3d print, just to keep it from shorting out in my messy working environment :roll:
Image

i milled a test in wood first to make sure i dont have any collisions with components on the board:
Image

the same version in aluminium
Image

the two sides of the case dont match on the outer conturs, i used leftover stock and didnt set it up carefully enough, but the sides are indexed by the screws so it doesnt really matter.
Image

the VESC is basically clamped between the two case halfes and only contacts the mosfets.
Image


i put some thermal compound on the fets to make sure they do contact the case. should be allright once the case is screwed together.
Image


a first test showed that the case gets warm pretty quick, i might do some further testig, but i think the thermal conductivity through the fets should be good enough to keep the board cool.

one thing i didnt thought of before is that i now have the case on ground level, but i dont see why that should be a problem, its the same with cars, so who cares :D


Image

joe90
Posts: 49
Joined: 15 Dec 2015, 16:52
[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Re: case for VESC

Postby joe90 » 03 Apr 2016, 19:23

Nice work !

I noticed that the edges of the board are clamped by the case.
This is a good thing I beleive because the FET do heat the board and hence you can extract heat from this too.
I think that your case should improve quite a lot the heat removal.

Keep us posted about max continous A you are able to get with this.
Michel Cote
Twin v4.7 + SK3 6374 149KV
Bicycle
2500W peak/unit on 12S


[phpBB Debug] PHP Warning: in file [ROOT]/vendor/twig/twig/lib/Twig/Extension/Core.php on line 1236: count(): Parameter must be an array or an object that implements Countable

Return to “General”

Who is online

Users browsing this forum: Bing [Bot] and 1 guest