cboy said:
I certainly want to use the regen feature of my Kelly controller. Are you saying that to use regen I would need to wire things up as you propose? If I set the regen in the Kelly software to engage via the throttle instead of the ebrake would that eliminate the need to wire as you propose.
Well - I've never used a Kelly and can't run the configuration app (no controller so it won't run). However, My understanding from the manual is that the Kelly three regen modes - the two useful modes for regen have fixed or variable regen but both require that you engage the ebrake for them to kick in.
cboy said:
Or is there an alternative to just wire the ebrake to the controller?
Whenever you trigger the controller ebrakes directly, you should also run the ebrake signal to the CA so it doesn't keep applying or increasing the throttle to try to counteract the strange battery current reading that it gets and doesn't understand. Again, my understanding is that the Kelly will not apply regen until the throttle signal goes to ZERO, so if the CA keeps applying throttle, the controller will never turn on regen.
As far as I know, there is no way to program the Kelly to simultaneously use the throttle to control both regen and throttle. You can make a little external circuit with a few parts to do that, but I believe that normally the Kelly requires a separate variable control for the regen.
cboy said:
BTW, I'm actually running 2WD and thus two matching controllers. I didn't think that info was necessary for my original question but when we get into regen and ebrake connections, the dual controllers might start coming into play.
Whoa - important item on the list!
You will want to be careful about grounds with two controllers. This is one of those cases where the difference between a schematic diagram showing
what is connected is very different from a wiring diagram that shows
where those things are connected. The wire runs can become important as currents are trying to scoot around between the controllers and the CA by whatever means they can find.
There are different schools of thought on this (and frankly some are arguably Bad Ideas) but for a 30,000ft view I recommend:
- tie the controller and CA-DP grounds together using as short connection as possible and run to the battery GND from there.
- The CA3 solves a lot of problems by providing a 5V throttle supply that is independent of both controllers.
- You can join the controller throttle connections and drive them by the CA, but don't use the controller throttle ground connections.
- Similarly, you can join the controller ebrake connections, but don't use the controller ebrake ground connections.
- The best plan is to use a wheel pickup for the CA speed sensing instead of a controller hall. This ensures the speed distance stuff will always works regardless of which controller is powered up.
The general reasoning behind this stuff is mentioned in
this post. There's a link there to a very good overall 2WD wiring diagram by teslanv.
The only issue with the wiring in those posts vs your situation is the annoying Kelly 12V brake signal.
cboy said:
Hmmmm. In reading further into section 5.5 it would appear the idea of running cruise control from the controller rather than the CA3 is not a good one.
Very bad plan. Battery monitoring/limiting and centralized control for autocruise or PAS are the sticky parts of 2WD design. Here is where the CA3 works out nicely. It can do all that stuff for two motors as easily as one so you only have one set of controls and everything plays together nicely.
cboy said:
For example, I don't know what a "divider" is or what it might look like. I also don't know resistors...
As best I can decode it:
a) I would attach my ebrake switch wires to the CA3 ebrake connector....
Nope - but your next post is better:
cboy said:
I think I'm beginning to see your solution. Basically, the 12v ebrake wiring to the controller is hooked up normally...with one exception. The input wire to the controller has a "branch" which goes to a resistor (to convert to 5V) and then onto the CA3 ebrake input (would this be the blue wire on the brake harness?). Thus, when the brake is engaged a 12V signal is sent to the controller and a 5V signal is sent to the CA3.
If that is correct, I don't get the part about grounding the 5V wire at the controller. I can see sending that 5V signal back to the CA3 and THEN grounding it within the CA3 to complete the circuit. So there is obviously something I am still missing about that part of it.
You have fallen prey to a common misunderstanding that suggests that resistors alone can lower/change voltage. It's really the current through the resistor that determines the voltage across the resistor (how much it drops the voltage), so you need to look at the entire circuit. Since some of the CA ebrake circuit is unknown to you and inside the CA, we use an external resistor divider to set the current through the resistors which allows us to say what the voltage drop will be without any foreknowledge of what's going on circuit-wise inside the CA. We just want to be sure we are using enough current so that any little extra bit the CA might use won't upset our calculations by more than a bit.
So - we might redraw the above diagram like so to help clarify the presentation:
KellyCA_ebrake.png
Here we see that this is called a divider because the 12V is divided proportionately across the two resistances. We can pick any resistances we want to get 5V as long as the resistor ratio is the same. So - we pick some values that allow high enough current that we think we'll overpower any little bit the CA is going to use.
Here I would recommend putting the R2 resistor on the CA PCB or directly across the CA ebrake connector pins since Bad Things will happen if the power is ON and the R2 GND connection is broken (12V will shoot into the CA EBK pin and bye-bye CA).
cboy said:
Also, another possible wrinkle/solution. I have hydraulic brakes (fronts hand operated, rears foot pedal operated) so I would have to rig up the ebrake NO switches anyhow. So could I accomplish the same end by simply putting TWO switches on each brake lever, one hooked to the CA3 and the other hooked to the controller? That set up would send a 12V signal to the controller and a 5V signal to the CA3. Is that what we are shooting for? Simultaneous signals to the CA3 and controller each at their own appropriate voltage? And does that allow the regen to work without creating any safety issues?
That's a bit of a mechanical PITA, but would work fine.
Any solution that gets braking signals to both controller and CA (as you suggest) will allow your regen to work fine - for fixed regen. To get variable regen you will also need to do a little fudging (extra thumb throttle), but that is exactly in keeping with the Kelly design concept.
The foot-operated rear brake makes this sound like a moped/motorcycle. If so and you already have secondary 12V power running a brake light, you could cheat and just hook a small 12V DPST DIP relay across the brake light to signal the controller and CA. In spite of being electro-mechanical, I think it's arguably a better solution than fiddling with multiple switch adjustments, etc. You still need to address the front brake issue, but one step at a time...
2WD regen can be problematic, but you might set your front motor for light fixed regen and set the rear for variable regen with a left thumb throttle. Many folks have wished for brake lever variable level regen control, but there doesn't seem to be a good commercial solution yet.
cboy said:
It seems for many controllers the CA3 is far from plug and play.
Quite the opposite. The problem is that your Kelly has a 12V brake circuit, etc and really shows an automotive or motorcycle heritage. Most ebike controllers and some Kellys use 5V brake logic and everything is much easier and only needs wire to hook up.