Arlo's power stage Leaf controller runs and drives page 103

Arlo1

1 TW
Joined
Apr 26, 2009
Messages
8,460
Location
Nanaimo
As you know I am working on a BLDC controller Here OK So Its time to design my power stage. From this I will build 3 6 fet prototypes to test irfp4568 x6 irfp4668 x6 and irfp4468 x6 to their destruction Probably on my X5 but we will see it it can handle the power.... Once done I will look at the make power numbers and max current to see what I was to build for running colossus.

The Plan its to build the best layout possible. What I am thinking is using aluminum to prototype with then maybe built the finished product with Copper if it shows any advantage.

My goals are in this order. Best location of parts, Best cooling, Best size, best weight, best price.

What I am thinking is 3 layers all horizontal to each other with the fets laid on their backs the drain from the metal back on the body and the leg both attached to the Positive rail on the Hi side fets and on the phase rail on the low side fets. To help cool as good as possible and to keep the resistance as low as possible!!!

SO can anyone point me to something that's good to use to electrical isolate three layers of copper or aluminum but conduct heat though so I can cool the bottom?

and or a non conductive liquid??? This it my schematic so far....
 
Simular to this.
 

Attachments

  • HBC40063_4bd04a27da474 (800 x 558).jpg
    HBC40063_4bd04a27da474 (800 x 558).jpg
    68 KB · Views: 17,006
Something like this.... http://www.lairdtech.com/Products/Thermal-Management-Solutions/Thermal-Interface-Materials/Electrically-Insulating-TIMs/. They have electricaly conductive and insulating thermal stuff!
 
So far dave linked me this stuff. http://search.digikey.com/ca/en/cat/fans-thermal-management/thermal-pads-sheets/1179751?k=PGS
It conductrs heat at 2-5 times better then copper! But its not cheep the thin .025mm stuff adds about $1 per fet its under!
I have found some pastes that conduct electricity with silver or aluminum or copper mixed into them. I think cooling the fets needs to be Tied for #1 with low resistance as priority. Here is a data sheet.
 

Attachments

  • BEST-Panasonic-PGSPyrolyticGraphiteSheet.pdf
    130.5 KB · Views: 269
This is what I am thinking for my layout sorry I have not mastered google scetchup or anything. But I need to save time for other things right now.
 

Attachments

  • 003.jpg
    003.jpg
    54 KB · Views: 16,833
Ok so today I bought 65$ worth of 68oz 12gauge copper :mrgreen: I cut up what I need for a 6fet controller and here is what it will look like. Once I test this I will cut the rest to make a bigger 24 or 36 fet controller with TO-247 package fets of what voltage I find best to work with what winding works best as a package for colossus. I have seen the 100v fet on paper will produce ~3 times less heat so....
 

Attachments

  • 002_renamed_8246.jpg
    002_renamed_8246.jpg
    48.3 KB · Views: 16,782
  • 003_renamed_12896.jpg
    003_renamed_12896.jpg
    47.1 KB · Views: 16,782
  • 004.jpg
    004.jpg
    39.8 KB · Views: 16,782
  • 005.jpg
    005.jpg
    44.8 KB · Views: 16,782
  • 006.jpg
    006.jpg
    45.8 KB · Views: 16,782
  • 007.jpg
    007.jpg
    46.8 KB · Views: 16,782
  • 011.jpg
    011.jpg
    60.2 KB · Views: 16,782
  • 012.jpg
    012.jpg
    67.2 KB · Views: 16,782
  • 013.jpg
    013.jpg
    56.7 KB · Views: 16,782
Yeh I have been recomended to use this stuff from Dave, http://search.digikey.com/ca/en/products/EYG-S091203/P13690-ND/1630955 It has 2-5 times less thermal resistance than copper!!! The only thing better is diamond. I will use it under the fets and the thin stuff conducts electricitiy 2 times as good as iron. I will also solder the drain leg to the plate under the fet to make sure the electrical resistance is uber low!
 
I thought graphite only transmitted heat well in two dimensions (i.e. parallel to the plane of the material). Isn't it a pretty good insulator in the other direction (perpendicular to the plane of the material)?
 
mlim said:
I thought graphite only transmitted heat well in two dimensions (i.e. parallel to the plane of the material). Isn't it a pretty good insulator in the other direction (perpendicular to the plane of the material)?
IM not sure it says A-B plane... But it doesnt list anything for the other direction
 
This video show some examples so I think I will get some. http://bcove.me/tua8w67g
 
parabellum said:
Arlo1 said:
This video show some examples so I think I will get some. http://bcove.me/tua8w67g
Is it electrically conductive as well?
You can get it plain as electricaly conductive and electricaly insulative as well.
 
Arlo1 said:
You can get it plain as electricaly conductive and electricaly insulative as well.
Using my best Mr. Burns voice: "Excellent!" :twisted:

...and exciting! Very good info; I can see many useful applications :)

Keenly observing from afar, KF
 
Arlo1 said:
2-5 times less thermal resistance than copper

Good find!

Imagine if there were PCBs made with this stuff as core. Then you could make high power electronics with surface mount transistors instead of the big through hole ones. That way you could put more transistor in a small space, to get better current sharing. With a big heatsink on the bottom side of the PCB, cooling of the SMD components would be as good as, or better than, a TO220 on a heatsink.
 
This might take a while lol but it will be worth it and should not be to bad for replacing fets.
 

Attachments

  • 007 (800 x 600).jpg
    007 (800 x 600).jpg
    58 KB · Views: 15,011
Alan B said:
Good conductor. Hard to solder without overheating the FETs.
I have a plan for that ;)
 
bearing said:
Arlo1 said:
2-5 times less thermal resistance than copper

Good find!

Imagine if there were PCBs made with this stuff as core. Then you could make high power electronics with surface mount transistors instead of the big through hole ones. That way you could put more transistor in a small space, to get better current sharing. With a big heatsink on the bottom side of the PCB, cooling of the SMD components would be as good as, or better than, a TO220 on a heatsink.
It doesnt conduct electricity as cood as copper...
 
Tomorow I will drill and tap the heat sync and solder it all together and build the driver stage!!!
 

Attachments

  • 010_renamed_21910.jpg
    010_renamed_21910.jpg
    58.9 KB · Views: 14,973
  • 008_renamed_28008.jpg
    008_renamed_28008.jpg
    68.5 KB · Views: 14,973
Arlo1 said:
It doesnt conduct electricity as cood as copper...
I imagine that non conductive type just has a tin layer of other non graphite stuff inside or outside. I wonder how heat conductive it is and how many volts it can handle before spark jumps, also how resistant to cuts of sharp transistor edges.
 
parabellum said:
Arlo1 said:
It doesnt conduct electricity as cood as copper...
I imagine that non conductive type just has a tin layer of other non graphite stuff inside or outside. I wonder how heat conductive it is and how many volts it can handle before spark jumps, also how resistant to cuts of sharp transistor edges.
The backs of these IRFP4668 fets are metal in the center with plastic on the edges. Its all listed on the data sheet for heat transfer rates. And the way I made my layout is for the back of the fets TO conduct as well as the drain leg, I will test it with just metal on metal and Krypton paper in between the plates for now.
 
I just found this stuff its price is cheeper and the specs are better. http://www.digikey.ca/ca/en/ph/LairdTechnologies/Tgon800series.html
I also ordered some t gard for the plates.
 
I really like your layout. Nearly the same as I've got here in my notes, will post it up later :)

As far as soldering problems go: I would try to get some 43Sn43Pb14Bi, it melts at 163 degrees C, and have good mechanical properties. (similar to 63Sn37Pb, but 20 degC lower melt temperature). Bismut allyos aren't THAT expencive, but Indium totally is xD
 
Back
Top