Game Changer Page 3 with video of 1700amps!!!!

Arlo1

1 TW
Joined
Apr 26, 2009
Messages
8,459
Location
Nanaimo
I have been holding this back only because I don't want any naysayers to call me out before I can prove its worth. I am doing this one open source as well, as a way to "give back" and because the world needs it. The fastest way we can get more EVs on the road is to stop the bull shit with IP. As Elon Musk said having a patent is like having a loto ticket to a lawsuit.
I have had this on the back burner for a good while. But as I wait for some parts for the Leaf inverter I will start a thread on this one.
This is the best way to lay fets out to current share. It solves MANY problems all at once. I will try to work though the start of this over the next couple weeks.

I will do my best to explain as I go.

But the basics are The top layer will be the positive buss which is an aluminum circle then a set of boards with high side drains on the top against the aluminum positive buss maybe with some silver grease under that so they conduct well. On the bottom of that is the source trace which is against Aluminum cut into 3 pie shaped pieces. Each of the 3 is 1 phase leg buss. Then under that is another set of boards in the same manner as the first set with the low side fets drains facing up on the top of them to a copper trace on the bottom with the negative buss aluminum circle.

The benefits are as follows.
Perfect (as perfect as possible) current sharing.
Matched, low inductance
Central cap locations which helps reduce inductance between all caps/fets.
Low profile so it will be compact.
Can fit in side a motor if needed.
And many more
All allowing for more power density with hi amperage.
 

Attachments

  • 347.jpg
    347.jpg
    76.8 KB · Views: 8,772
  • 356.jpg
    356.jpg
    68.7 KB · Views: 8,772
  • 229.jpg
    229.jpg
    57.4 KB · Views: 8,772
  • 280.jpg
    280.jpg
    45.6 KB · Views: 8,772
Just about done 1 phase for deast testing.

I will install all caps and rout all the power buss.

I am really liking this project.
 

Attachments

  • 101.jpg
    101.jpg
    51.3 KB · Views: 8,591
  • 102.jpg
    102.jpg
    52.3 KB · Views: 8,591
  • 103.jpg
    103.jpg
    53 KB · Views: 8,591
  • 105.jpg
    105.jpg
    55.3 KB · Views: 8,591
  • 106.jpg
    106.jpg
    47.4 KB · Views: 8,591
  • 109.jpg
    109.jpg
    52.7 KB · Views: 8,591
  • 108.jpg
    108.jpg
    43.3 KB · Views: 8,591
Top disc is Positive with the center 3/4" lug being connected to it.
The 3 outer 3/4" lugs are welded into the 3 pie shaped pieces and they are the phases respectively.
The bottom small disc is negative and has a set of 3 1/2" lugs attached to it.
The big bottom disc is a heat sink.
I have a new design that is about 1/2 the size of this with the same number of fets (48)
 

Attachments

  • 047.jpg
    047.jpg
    46.7 KB · Views: 8,639
  • 049.jpg
    049.jpg
    53.7 KB · Views: 8,639
  • 059.jpg
    059.jpg
    51 KB · Views: 8,639
This is the files I am starting with.
I have major revisions to come but this is what I am using for V1.0 for testing.
V2 is 1/2 the size with all the same power :)
 

Attachments

  • desat v2.5 split mosfet .zip
    89.7 KB · Views: 136
  • desat v2.5 split mosfet .pdf
    99.4 KB · Views: 203
  • game changer 11.JPG
    game changer 11.JPG
    71.4 KB · Views: 8,635
Ill note here I have a new design I will order after testing with this one.
It separates the driver off the actual board the fets solder to. That allows me to mount the driver boards outside the casing to keep them further out of the noise and put them on a perpendicular plane when compared to the discs of aluminum to make the diameter smaller.
 
I'm very excited for this controller Arlin! Way to go!
 
Lebowski said:
don't worry I had these boards ordered before you showed your low inductance power-stage design. And I just now realized how this is like having 8 of your low inductance power-stages made into a circle...
This is the best way :) Great minds think alike.... :)
 
Interesting, but you should change your thread title so people interested in controllers will want to look at the thread, and not just people interested in "random" thread titles. ;)
 
amberwolf said:
Interesting, but you should change your thread title so people interested in controllers will want to look at the thread, and not just people interested in "random" thread titles. ;)
Its the name of the controller. They will look at is as time comes no worries. ;)
This was more or less posted to make sure its "out there".
 
Check these caps out. http://www.sbelectronics.com/product-category/dc-link-capacitors-non-ul-recognized/
:) I also have them working on re-designing them slightly to work better with Game Changer :)
 
Arlo1 said:
Check these caps out. http://www.sbelectronics.com/product-category/dc-link-capacitors-non-ul-recognized/
:) I also have them working on re-designing them slightly to work better with Game Changer :)
I wondered where you put the capacitors, but these round caps look perfect.
It seems to be the lowest inductance structure ever and most even current density, looking forward to see how it performs.
 
Those caps are about as good as it gets right now for EVs.
 
is it the same as golden motor controller?
MagicPie4.jpg


where would you stick the controller? on the external case?
 
Very nice Arlo1. Those caps look good and the circle based layout looks really good for sharing current.

Are you going to start stacking layers to increase power density with more parallel devices?
 
cwah said:
is it the same as golden motor controller?


where would you stick the controller? on the external case?
Not even close
GM is good for cheep ebike kits and entry level stuff. But they put all the digital stuff inside the noise as well.
 
zombiess said:
Very nice Arlo1. Those caps look good and the circle based layout looks really good for sharing current.

Are you going to start stacking layers to increase power density with more parallel devices?
Thank you very much. I have had this on my bench for a long time and I decided I need to get it done.

I have some ideas on how to add more fets or IGBTs in parallel but I have not sorted out how to mount them when positioned in a different manner.

The trick is to keep 1 layer because the top of the board has a drain trace under the fets and a source trace on the bottom of the board to connect to the piece of aluminum (can use copper) below it. Making it laminated. As well keeping the top aluminum + buss circle and the bottom smaller - buss circle as close as you can keeps the positive and negative laminated to some extent.
But adding a few more fets and growing the circle is not to hard.

If you can current share properly we should be able to make this thing run at 160 amps rms per fet/igbt (leg limits for a 247/264 package) for a 10-30 second burst. With 8 150v rated fets we are talking
1280rms phase amps at 102v nominal would should be able to push 1200-1400 DC amps depending on field weakening that's 122400-142800 watts !
But when you up the voltage by simply swapping out the fets for IGBTs and re-tuning the driver to say 360 nominal we are looking at up to 504000 watts!
 
A round speed controller...sounds like a hard sell since it's not the typical "box" shape. While there are advantages to this design, mounting and placement on the EV will be more "interesting". When I first saw it, I thought maybe mounting it next to a hub motor would be good, but I don't know the more I think about it.

Do you have any idea how much you will be selling it for?
 
ElectricGod said:
A round speed controller...sounds like a hard sell since it's not the typical "box" shape. While there are advantages to this design, mounting and placement on the EV will be more "interesting". When I first saw it, I thought maybe mounting it next to a hub motor would be good, but I don't know the more I think about it.

Do you have any idea how much you will be selling it for?
First I prove it. Making money has never been a priority for me.
Round is best. Motors are round it can be mounted beside a motor on one end. Look at Tesla's inverter, the housing is round. They use a triangle type layout that is not as good as this.
BTW how many hub motors need 1200amps+ fed into them?
Priority number 1 is to make it work as best as it can and that means a round layout. I can make a housing square or rectangular if needed but that is stupid.
When finished and refined this will have the highest power to weight and the lowest volume of any controller layout. It will also last the longest because the fets will share current the best they can and the inductance is low from all fets back to the cap meaning the voltage spikes at turn off as the free wheeling diode starts to conduct will not be as high. This might take months or years to perfect... That will depend on cash flow. I don't like asking for handouts....
This is not a hard sell go get me a list of 1000 amp+ controllers that are relaible and get back to me ;)
 
At first I thought you where making e-bike controller, I now see that this will be motorsport controller first.
But I guess scaling down is hardly never a problem :lol:
 
macribs said:
At first I thought you where making e-bike controller, I now see that this will be motorsport controller first.
But I guess scaling down is hardly never a problem :lol:
I live for big power.

I have built 100s of <3kw controllers and that's not fun for me.
I can scale this down and in fact I have the boards for a 6 fet version.
But proving big power first is my style ;)
 
when I said "hard sell" it is difficult to determine physical size from pictures. It looks like this controller is about 10" in diameter. I may be waaaay off and that's totally cool, but that was my impression. What are the actual expected dimensions? Anyway, for most EV solutions on ES, a 10" diameter controller is probably biggish. Of course with the current handling you are estimating, this thing will run a car!
 
Back
Top