Taking the Infineon 12 Mosfet controller to the Extreme

TMaster

1 kW
Joined
Dec 1, 2009
Messages
396
Location
TX
Infineon XCKJ8B116A 12 Mosfet Brushless Controller, stock for 45 Amps peak with 72v.
Model: EC-124110-LYEN EDITION

Stock controller review: http://endless-sphere.com/forums/viewtopic.php?f=4&t=17679

After testing this controller out of a few days I decided it had a lot more potential; and being a power hungry freak, I decided to beef this thing up to the max!

First thing to do was to replaced the power side resistors with 870 ohm 3W total of resistance, which actually came out to 856 tested. This is for running 22S A123's 72.6V nominal. I replaced them with Vishay brand resistors, they are a tad smaller in size compared to the stock ones, but are the same Wattage.

Next thing I did was replaced the 12AWG wire with 10AWG. As I found out after de-soldering the 12AWG wire, the holes in the board are too small for the 10AWG. So it was either drill the holes larger, or modifiy the wire ends. I decided to mod the wire ends. Also the grommit holes at the end of the controller for the wires were too small, I had to drill those out larger and carefully. I used colored heat shrink and 4mm bullet connectors. The 4mm bullets are good for around 75 to 100amps continuious, and higher peaks.

Next was to Pull up the stock trace solder and lay down 2 layers of 12AWG solid copper wire. By my calculations, the 2 stacked layers of 12AWG and the solder, the trace are is now equal to 8 to 9 AWG. Plenty of beef for going to 100 amps or more. I spent a long time on that area.

Shunts, What to do? I decided to mod the controller physically, rather then rely on the programming to do the job. I went with 4 shunts total. Thanks to Lyen for hooking me up with a couple spare shunts. Being there is not much room on the top side of the board, I decided on putting the other 2 underneith. Some of you may ask, why not just solder up the shunts instead of adding more shunts? Well, I thought to myself physically having more mass and conductive area for the current to go through is better. 2 shunts are almost equall to 11AWG, and 4 shunts are almost equal to 8AWG.

After finishing these mods, the controller is pulling 100 Amps peak without any settings changed in the programming. The phase current and rated current were still set for 80 and 30. So adding the 2 extra shunts gave me 40 more amps! :D The programming settings are able to go up to 157 phase and 50 something rated current. I'm guessing the controller can do 120 Amps easy, but 95 is plenty for me. I am also using Arctic silver 5 thermal paste on the mosfet block.

Inside the controller was a tad warm. The controller wasn't getting any good air flow though from the outside. I haven't moved it to the new location for this setup. I measured 55c on the mosfet inside after hammering the throttle for 3 or 4 min. The temps outside were only 65F. Thats pretty cool for here in Texas. I thought to myself the controller may get up to 65C when the temps were 95F this summer. The small wires inside the controller are only rated for 80c. This was on the edge for me, and I had to do something else, to put me in the safe zone in my mind.

I had to figure out a way to cool this thing under extreme conditions. My new setup is to mount the controller in the center of the bike. The controller sides will be open to the air with my new setup, so that will help out. But I still wanted to make sure I wouldn't ever have problems. I drilled 4 intake holes in the front end plate and cut out a large exhaust hole in the rear plate of the controller. I ordered a 40mm 6.5CFM case fan for the rear. This really turned out well. The suction of air was excellent. The fan runs of 4S off my pack. Which is 13.2v nominal, and just over 12V under throttle load. I have an external switch so I can turn the fan off and on when I want. I have measured around 41c now with using the fan 25% of the time durring the ride. Going easy you don't need the fan on, but it hasn't been over 80F here yet in Texas, and I haven't been hammering the throttle to the extreme yet either for a long period of time.
 

Attachments

  • IMG_6179.JPG
    IMG_6179.JPG
    56.4 KB · Views: 9,262
  • IMG_6195.JPG
    IMG_6195.JPG
    67.7 KB · Views: 9,745
  • IMG_6198.JPG
    IMG_6198.JPG
    57.2 KB · Views: 9,784
  • IMG_6199.JPG
    IMG_6199.JPG
    43.6 KB · Views: 9,744
  • IMG_6201.JPG
    IMG_6201.JPG
    42.5 KB · Views: 9,744
  • IMG_6188.JPG
    IMG_6188.JPG
    47 KB · Views: 9,784
  • IMG_6220.JPG
    IMG_6220.JPG
    65.5 KB · Views: 9,261
  • IMG_6223.JPG
    IMG_6223.JPG
    46.1 KB · Views: 9,261
  • IMG_6206.JPG
    IMG_6206.JPG
    67.9 KB · Views: 9,744
  • IMG_6203.JPG
    IMG_6203.JPG
    70.3 KB · Views: 9,261
  • IMG_6204.JPG
    IMG_6204.JPG
    69.8 KB · Views: 9,745
  • IMG_6209-5.jpg
    IMG_6209-5.jpg
    86 KB · Views: 9,262
  • IMG_6226-5.jpg
    IMG_6226-5.jpg
    50.6 KB · Views: 9,262
  • IMG_6259.JPG
    IMG_6259.JPG
    49.1 KB · Views: 9,261
  • IMG_6253.JPG
    IMG_6253.JPG
    40.4 KB · Views: 9,261
  • IMG_6258.JPG
    IMG_6258.JPG
    32.3 KB · Views: 9,261
  • IMG_6263-5.jpg
    IMG_6263-5.jpg
    86.5 KB · Views: 9,262
  • IMG_6276-5.jpg
    IMG_6276-5.jpg
    36.7 KB · Views: 9,261
  • IMG_6261-5.jpg
    IMG_6261-5.jpg
    54.4 KB · Views: 9,236
Nice job!!! You should see the one I did for KiM (that he doesn't want anymore :( )
 
Thanks!
You should see the one I did for KiM
Send link.

I donno, But I love my setup now. Everything in the middle! Leaves for plenty of room on the back rack. And the enclosure is only 3.5" wide. The bike doesn't wheelie anymore either, and the weight is closer to the ground.
 
That's neat work. I'm part way through beefing up a 12 fet also.
What value did the shunt end up as?
 
Oh i have no idea, the two new shunts that came with the controller are 230 or 250, and the old shunts under the other side are 250 or 280. I have no idea how to calculate things. Having 3 shunts would have been ok, then I could have used the programming to bump the current up. But it seemed everytime i modded something it was always not enough for my crazyness, so I just put 4 shunts in like the 18 fet controller. I figured it would be good like that!! WEE!!!!! :mrgreen:
 
Yes, very nice job.

A few questions:

- Are you worried about water ingress through the fan? I don't think so, given the battery box.

- Did you insulate under the PCB? It's very close to the casing. I duct taped my 18 FET Infineon.


And you wouldn't have been able to hit those currents with programming. The 18FET model tops out at 57A I think.
 
I just went for a ride today and i was hitting 100 amps. So i will modifiy the first post LOL...

No insulation, I was thinking about how close it was also, but you know the regulator on the underside of the board sticks out already, and the pad of it is the output side. So with that the beefed traces don't stick out any further then it, so it will be fine.

I don't ride in the rain, so I wont have to worry. :mrgreen:

The 18fet model should be able to hit 100 amps no problem, geeze, the gurus were doing 170 amps on them. :D
 
No, I meant the programming tops out at 57amps.

If you want to go higher you have to mod the shunts or circuit to trick it.
 
There is ample room to enlarge the holes for fitting 10awg wires too
i drilled mine out to fit 10awg. I like the job on the traces i did very similar also, i messed up the shunt
though not knowing what i was doing.

Would you be open to modifying another? I still need one myself and have been generously forwarded the $$ to replace
the first controller i purchased and sent to Luke, since i tried beefing up the Infineon myself and it ended up not so good
I would gladly pay for parts and yours services to do another? pm if interested.

liveforphysics said:
Nice job!!! You should see the one I did for KiM (that he doesn't want anymore :( )

..yes well feel free to show us all Luke be nice to see the controller i bought & heard so much about but never sighted.
After waiting 5 weeks i gave up...the rest is history.

KiM
 
Nice soldering job, TMaster - it's pretty hard to add lots of solder while avoiding having it pool together and make potential shorts between the battery and the outputs.

I do have some suggestions though, and believe me some of these details make very big performance differences, and/or they will make your controller much more reliable over time.

Additionnal modifications suggestions:
- Change that POS silicone pad for thin kapton (the glueless kind, and use just a little thermal paste on both sides to have it stick instead).
- Those big stock caps are fairly crappy (high ESR), so they may heat up and destroy themselves over time.
- Watch out that the controller PCB doesn't arc too much from beefing up your copper battery traces (the copper shrinks as it cools). I use smaller lengths of wire with stress relief bends in each section and this helps a lot. I don't know if your PCB is bent much or not though (no photo of that angle).
- Phases "A" and "B" on these controllers (EB212) have a design weakness: The low side FETs (connected to battery negative/shunt) of these two phases do not all have a good connection to the negative/shunt trace. One FET on each phase uses an indirect trace on top of the PCB instead of the more direct traces the other FETs use. What I do is add two flatened 12/14AWG jumper wires from the negative trace over to the FET's source pin - this gives a much better path for the current instead of going through the top PCB trace.
- I add extra ceramic chip capacitors very close to each phase's + and - FET connections. One small MLCC like this on each phase (even only 1uf) can have an ESR even lower than a huge electrolytic capacitor for higher frequency bypassing. This is most important if you are operating your controller very close to it's maximum values (specially when close to the FET's voltage rating).

This is my idea of a really extreme 12FET controller build! :twisted:

Pat
 
ZapPat,

Did you post any photos of your mods elsewhere on ES? Pictures would help confirm my understanding of your suggestions.

Dave
 
dbaker said:
ZapPat,

Did you post any photos of your mods elsewhere on ES? Pictures would help confirm my understanding of your suggestions.

Dave
Sorry Dave, I have no pics of my own mods right now. But I will be opening up two of my older the controllers from my 2WD dirt bike and update them a bit (install 1mil kapton MT + AAVID ultrastick thermal compound, current override circuit to be disabled), so I'll take some pics then (about 1 week). If you have any specific questions in the meanwhile, ask away.
 
Nice soldering job! Would it help to use square copper bars?
What iron did you use? Wattage, Tipsize?
Did I miss something or did you leave the original Fets in? No IRFB4110 neccessary?
I have a similar looking Controller, that has 'X8M06C' lasered on the controller-chip and my Fets are 'ST P75NF75' stamped.
Can you confirm, this is the same than yours?
Mine came with kapton tape under the Fets, but is only goop'd on one side. Is it a big difference to use arctic silver?
Where else can I get the shunts from?

Sorry for too many questions
Olaf
 
Zapat, Thanks for the pointers!!!

- Change that POS silicone pad for thin kapton (the glueless kind, and use just a little thermal paste on both sides to have it stick instead).
Yeah I noticed it has a .5 or 1mm thick soft pad between the mosfets and heatsink block. My old controller has the thin fim of plastic.
Those big stock caps are fairly crappy (high ESR), so they may heat up and destroy themselves over time.
Eh, im not too worried right now. I have the fan blowing through, should help out. I can panasonic them someday!
Watch out that the controller PCB doesn't arc too much from beefing up your copper battery traces (the copper shrinks as it cools).
Oh, I spent hours on the trace laying. I made sure the copper lengths were flat. There is no hendering forces. I use a 80W iron and make sure the entire area's are flowing before they cool.
I add extra ceramic chip capacitors very close to each phase's + and - FET connections
There is one small .1uf cap next to each phase. But you would have to show some pics for me too see what you did.
Phases "A" and "B" on these controllers (EB212) have a design weakness: The low side FETs (connected to battery negative/shunt) of these two phases do not all have a good connection to the negative/shunt trace. One FET on each phase uses an indirect trace on top of the PCB instead of the more direct traces the other FETs use. What I do is add two flatened 12/14AWG jumper wires from the negative trace over to the FET's source pin - this gives a much better path for the current instead of going through the top PCB trace
Some pics would be nice to see of this.
 
olaf-lampe, Thanks!

Square wouldnt matter, really the round is fine, it gives nice area for solder flow underneith the round.
80W weller Iron with the 3/8" tip
I had lyen install 4110's from the start
That may be a older controller you are looking at, best to show pics. Also the 75NF75 Fets are 75V, 80amp. The more capable the fet, The better overall it will be. The 4110's are so much better in every way.
The arctic silver 5 may help a little. Thats all I had laying around here, so thats what i used.
Shunts are actually resistance wire, you can buy rolls of it from places. But if you want a cheap solution, ask an es member or put up a wanted ad :D Someone with a broken controller may have some even.
 
Actually I have two controllers, one is mounted and the other is where I took the data from.
The green PCB doesn't seem to have a program connector and I'm also missing the regen input. :(
Tomorrow I'll swap them and see what I can do with the yellow one.
Olaf
 

Attachments

  • 36V700W 001.JPG
    480.6 KB · Views: 6,928
  • 48V_Inf_12FET_front.jpg
    48V_Inf_12FET_front.jpg
    82.3 KB · Views: 6,876
Back
Top