Controller mounting inside vs outside

[Emoto]

Thinking of mounting my infineion 18 fet on the inside of the frame,Would like to hear from anyone who has done similar ive seen it some where before.

one main benefit would be by bolting the fet buss bar directly to the frame it would now be a huge heat-sink instead of the standard case.
But the big question is as its enclosed in the frame and no air flow will it be counter productive.
so what method would be better.
Externally mounted and vented vs internally mounted, fett buss bar bolted directly to frame as in pic and some form of induced air flow.

 

[nechaus]

seems like a great idea...
It will feel nice and warm in the winter

 

[dogman dan]

How it should be done in an ideal world I'd think. You might be able to retain the case, by using some kind of machined aluminum bar to make the good connection from the case at the heatsink screws to the frame. Even a so so connection should work well. Flat plate under the controller, and bolt it on normally? Then some small vents in the frame could let a tiny but dry breeze in.

I always object to smothering controllers, but that's when discussing just putting it into a padded unventilated bag.

 

[kdee122]

I use an "inside the box" controller and on my initial prototype I mounted it with double side foam tape. At 5kw the controller easy over heats after 2 to 3 minutes of full power use. I will soon switch to thermal bridge design with the heatsink bridged to the alum enclosure. So in short you will need to connect the heatsink if you are running high power.

 

[Emoto]

Yeah dogman it would be a real clean install, on mounting id mount as in photo heatsink bolted direct to frame, would lap controller heatsink and inside of frame for a perfect fit and thermal paste of course. but would make a some custom brackets instead of the case helping it breath posibly a strip of 1mm rubber for between batts if i get paranoid.
And pos later some clever induction tubing to blow air over controller . or a fan.

 

[Punx0r]

If you can get a good thermal connection to the frame then your indea of mounting inside sounds like a winner

 

[agniusm]

I will put my controller inside the box as well. My approach is to use water block on fets and 240mm computer radiator with fans and pump and 40C termoswitch on the mosfets to switch on pump and fans.

 

[snellemin]

I mount mine inside the frame. I use thermal conductive tape between the controller and frame. Heat transfers pretty good. I have a thermistor on the controller and I haven't seen 50C yet, but I haven't ridden it in the summer yet.

 

[Emoto]

I use an "inside the box" controller and on my initial prototype I mounted it with double side foam tape. At 5kw the controller easy over heats after 2 to 3 minutes of full power use. I will soon switch to thermal bridge design with the heatsink bridged to the alum enclosure. So in short you will need to connect the heatsink if you are running high power.

Thanks, OT love the frame concept almost transformer like, you got me thinking man for my next build.

If you can get a good thermal connection to the frame then your indea of mounting inside sounds like a winner

agreed thats not to hard, if we can get good air flow as well that just might seal the deal.

I will put my controller inside the box as well. My approach is to use water block on fets and 240mm computer radiator with fans and pump and 40C termoswitch on the mosfets to switch on pump and fans.

agniusm what do you mean by " water block on fets" sounds complex adding water.

I mount mine inside the frame. I use thermal conductive tape between the controller and frame. Heat transfers pretty good. I have a thermistor on the controller and I haven't seen 50C yet, but I haven't ridden it in the summer yet.

Snellemin Great! real data with the thermistor, i also will do the same, no point doing this if your not going to monitor progress.
Is that at 5kw+?
Did you measure temps before relocating
And i wonder what would be safe controller max temps 80c/100c ?

On the controler case i think ill make one out of this perforated ali good protection and will radiate heat nicely

 

[dogman dan]

I like the idea of something around the controller. Just in case of some damn loose metal object in there, some human boo boo with a loose wire, etc.

I think just heat conductive mounting of the bottom of a normal controller box will heatsink it fine. The box is a heatsink, so I don't think it must connect at the fets. Just needs to have the box cool by conduction if air can't blow over it.

Bolt it to a chunk of aluminum plate, to get a good contact on the whole bottom of the stock box.

 

[Emoto]

Yeah That would be tempting fate if not protected batt lead coming off or something nasty.
sorry i will re-edit original post i meant fett bus bar bolted directly to frame, eliminating original controller case/heatsink.
Biggest heatsink wins right.

 

[agniusm]

In the photo you can see copper bar that I milled chanel inside to circulate water and cool fets. I hope it will give some benefit. Mine is xheap greentime controller and i will squeeze more than it can give, just curious to see how much more power it can handle with active water cooling. I know a person who uses same approach like yours. He is building 4motus bikes. It is a lot of area to shed heat away. Good stuff

 

[Samd]

I bolt my aluminium FET heatsink bars straight to the cases.
http://endless-sphere.com/forums/viewtopic.php?f=6&t=51661&start=50#p801235

Take a good look at the back of the controller. See all those PCB tracks with power across them - DO NOT LET THEM MAKE CONTACT WITH BARE METAL!

I use some thin neoprene fastened to mine. 2mm thick poly chopping board might be useful too.

I know I am stating the obvious but can't see if anyone mentioned it above.

Kepler also mounted his direct to the bike.
http://www.electricbike.com/stealth-hot-rod/

 

[Punx0r]

I think it was Jeremy Harris who posted some calculations suggesting that watercooling offered little or no benefit compared to a good passive heatsink. The main restriction seems to be the thermal resistances between the FET junction and the heatsink, with the cooling of the heatsink only an issue if you are forced to use one that is too small or has insufficient airflow.

I hope that you find some benefits from your watercooling but I think you must take care to monitor the FET case temperatures because at high peak current your heatsink might be cold but the FETs dangerously hot.

Are the FETs insulated from the heatsink in the Greentime controllers? If so you will probably find you can find a lower thermal resistance material

 

[Emoto]

In the photo you can see copper bar that I milled chanel inside to circulate water and cool fets. I hope it will give some benefit. Mine is xheap greentime controller and i will squeeze more than it can give, just curious to see how much more power it can handle with active water cooling. I know a person who uses same approach like yours. He is building 4motus bikes. It is a lot of area to shed heat away. Good stuff

Nice to see your work it will be interesting for sure.

Hay Samd thanks for that, agniusm has reminded me Thats where i saw the bus bar bolted to frame on a nice custom emoto,
Dont worry about the obvious many minds are a better collective. on the neoprene another good idea really light and another good use for old wetsuits.
Might have some questions after ive had some time to look at your build.

PUNX0R -- real interesting / encouraging if by any chance you come across jeremys data link away.

 

[Punx0r]

There may be more, but here are a few posts:

http://endless-sphere.com/forums/viewtopic.php?f=2&t=44475&p=654874&hilit=water+cooling#p654874

http://endless-sphere.com/forums/viewtopic.php?f=3&t=40549&p=592794&hilit=water+cooling#p592794

 

[John in CR]

The frame looks like steel, and if so you'll need a much larger than normal aluminum heat spreader bar to connect to the steel bike frame. Then include a fan to get some air flow over all of the components. I'd include a bit of insulating material between the controller space and the battery, and then a fan should make fresh air entry unnecessary.

 

[Punx0r]

Ah. I assumed it was aluminium...

 

[Emoto]

Good guess punxor i should have mentioned that and one of the main reasons this method should benefit me over other frame materials, but a good point john for others considering this mod. of course ill insulate it, but im thinking ill use the perforated ali as a case and get insulation + radiation and pos a fan but i like to keep things simple if i can.
Got another idea if you use long enough mounting bolts you could also mount a large external finned heatsink [ on the outside ] as well this could be a turbo sink .
Something like this

 

[Samd]

Good to know the info was useful. Holding it as in the very first photo of this thread may even result in the caps giving a light crackle as they discharge to the frame.

 

[agniusm]

I did heatsink on my first project. Also internal controller:
FWD to 02:02, reveals the heatsink.
[youtube]ZmK1yWoZNtw[/youtube]

 

[Emoto]

Good to know the info was useful. Holding it as in the very first photo of this thread may even result in the caps giving a light crackle as they discharge to the frame.

Yes sam and if sitting on the frame waiting for the lights to change could be err
The link to Kepler's http://www.electricbike.com/stealth-hot-rod/ is good to read.
Before i start drilling holes in my frame, I hope some one else beats me to it to and comes back with the before and after data.[external vs internal busbar bolted to frame] as i still have my reservations.
Dam agniusm i thought i was the first to think of this method.

 

[agniusm]

Not really