EV Charge Controller

I've looked around on the Internet at some of the home-brew electric vehicle conversions out there. One thing that bothers me a little about them is that so far as I can tell none of them actually implement the entire J1772 specification. In particular, so far as I am aware, none of them actually attempt to detect the incoming pilot duty cycle and insure that their current draw is under what's allowed.

In fact, most of the true home-brew conversions I've seen use an 882 ohm resistor and a diode on the pilot pin. This will work in most cases, but there's a great deal of difference between what merely works and what is compliant with the specification.

In particular, simply jumping straight to state C will fail on a Hydra in sequential mode. After all, sequential mode's intent is to insure that the cars take their turn. The Hydra will not even offer a pilot until it's prepared to actually give power, and the specification requires a car to stay in state B when the pilot is not offering at least as much current as the minimum the car is prepared to draw.

In addition, there are a variety of EVSEs out there with various levels of available charge current. That's the entire purpose of the pilot signal to begin with - if a car draws more than is allowed, the best case scenario is a popped circuit breaker.

As if that weren't enough, most (though not all - the Modular EV Power EV controller gets this right) ignore the proximity signal. The intent of the proximity signal is to demand an immediate shutdown of the charger to insure that the current draw is zero when the plug is pulled from the inlet. This prevents arcing. The spec also demands that the car return to state B when the charge is completed.

With all of that in mind, I have a design for an EV charge controller for home-brew EV conversions. It connects to the proximity and pilot pins of a J1772 inlet; and on the car, it connects to +12v, ground and two charger interface signals - "charge request" and "charge enable."

Charge request is a signal from the car that indicates that if AC power were available, the car would want to charge the batteries. The intent is that when the charge cycle is completed, this signal will drop, and the charge controller will transition to state B. This is a very good thing, particularly for public charge stations, which can text you that your charge is done, and indicate to other nearby users that your plug is free for their use.

Charge Enable is a signal from the controller to the charger, and represents an enable for charging. The charger must not draw power unless this signal is asserted, and must stop drawing power when it is dropped. The controller will drop the charge enable signal a short time before transitioning the inlet from state C to state B. This will insure that no current is being drawn when the EVSE's contactor is opened, which prevents arcing there.

In the absolute worst case, the charge enable signal can be used to open and close an AC power relay to supply power to a dumb charger. That relay may potentially arc, but that's better than the alternatives.

The charge controller has two DIP switches and a rotary configuration switch. One of the DIP switches is "vent required." EVs that use lead-acid batteries should turn that switch on so that state D is used in place of state C. The other DIP switch inverts the sense of the charge request signal to be a charge inhibit signal. The rotary switch selects from a set of 8 choices the maximum current draw of the charger. If the J1772 pilot ever drops below the configured minimum, the charging cycle will be interrupted.

The charge controller is powered by an ATTiny84 and an LM393 comparator. It is wired to see the configuration options, the charge request signal and the state of the proximity and pilot pins. It can switch between state B and C/D and control the charge enable output.

I haven't gone ahead and actually built one of these, however. I don't have a use for one, and I don't know anyone who does. But I think it could be made and sold for around $20, and would greatly improve the compatibility and safety of home-brew EV conversions that have built-in chargers.