View Single Post
Old 10-22-2014, 12:14 AM   #126 (permalink)
Cyruscosmo
EcoModding Lurker
 
Join Date: Jun 2009
Location: Seattle, Wa
Posts: 71
Thanks: 7
Thanked 31 Times in 26 Posts
Quote:
Originally Posted by e*clipse View Post
First qualification: I haven't read the whole thread.

Issues that bother me:
1) Isolation. What is the rated voltage? Yes, there is one. Remember, we're talking about 650V here. Many isolators, etc have voltage limitations far below that.
2) GND is local - the cell's local ground, I assume. This is ok, BUT data must be completely isolated from GND. (see below)
3) Dout is near the cell's voltage. That is fine when we're talking about cell 1. how about cell XXX - the one near 650V??
4) Any safety for the balance resistor's control? If it locks on, any sensor or override? Lee Hart has some excellent suggestions about this.

In one of my basic "how to convert an EV" books, it makes the very wise suggestion that the ENTIRE drive train's HV system is completely isolated from the frame. This book was written for relatively simple 144V DC conversions. Any detection of a frame to battery/drivetrain connection shuts down the main contactor. This is kind of like ground fault detection, but helps prevent a "hot" chassis. It's imperative that the BMS doesn't form this connection.

I'll pass. I've gone through that learning curve.

I'm sorry I don't have any better suggestions or alternatives. I do like that it prevents the spaghetti wiring, which (believe it or not) is all over my Leaf battery pack.

- E*clipse
Isolating the main pack from vehicle ground is first on my list. I have been bit by a case grounded winding on a 480 motor and that is as close to flying as I ever wanna be again. Lucky for me it was from one hand to the same arm.

Just a curiosity question, why not link the individual BMS boards to a controller with a light fiber? The plastic ones are cheep and "Non" conductive spaghetti sounds much safer. The batteries already supply the power for the board and you can use the same fiber for communication both ways right?

Yes I know all the ends will have to go to one control box but the total bundle would not be much bigger than my finger. Or you could make a few smaller control boxes that handle say ten cells each.

Just thinking out loud...

I got an idea for your "One unit per wheel" setup.

Instead of paying for a solid spool to be made, you have access to a lath and mill right? Make a shaft with a drive line flange on it like what the unit has already.

Working in from the flange you make an area on the new shaft for the oil seal then an area that is a close fit to the inside diameter of the carrier bore at the outer bearing. Next you have a hole perpendicular through the shaft that is the same diameter as the spider shaft then another area that fits tight inside the opposite carrier bearing and maybe a bolt on cap that rides against the other seal.

So what you do is remove the drive line flanges and dismantle the case. Next you remove the ring gear and pull the spider gear shaft. Remove spider gears and bushings and the side gears and bushings.

Now you install the new shaft with single flange to whichever side you want to drive and slip the spider gear shaft back in place. The "differential" part is now removed for single drive axle connection to one wheel. And you don't have to mess with splines.

Cyruscosmo
  Reply With Quote