Make: Fechters' Braking Regen Add-on for BLDC Motors

[fechter]

That would probably work.

It will need a heat sink and you should make sure it doesn't go over 35 amps.

 

[steveo]

That would probably work.

It will need a heat sink and you should make sure it doesn't go over 35 amps.

Hey Fechter

I was reading in an early post Gary did some hard testing runing his x5 at 30mph .. and seeing 50amp peak when hitting the resistive braking .. he was also using a 35amp 3 phase rectifier. Is it because some of the load is spread by the (coil) & rectifier? How would i go abouts testing for current running through the resistive braking anyways? Where how would i wire the watts up meter?

-steveo

 

[fechter]

You would place it between the rectifier and the resistor.
Those diodes can tolerate a fair amount of overcurrent for short times.

 

[sharkmobil]

Seems like a huge waste to convert all that energy into heat!
Going for a 90% efficient motor only to blow-off roughly half of that on braking?

How about devising a circuit to dump the breaking energy into a stack of super capacitors? Seems like a voltage converter is needed along with a controlling schematic to use the energy of the supercaps before batteries, if it is there. Or maybe set a current threshold and dip into the caps above the mark for boost?

Just throwing it out there... seems like converting current into heat ought to be less than a 9 page topic.

 

[Nimbuzz]

Where do you get "super capacitors?" as far as I know they are vaporous dream thingies that peeps talk and swoon about but are as often seen and touched as the Loch Ness Monster.

 

[TylerDurden]

Where do you get "super capacitors?" as far as I know they are vaporous dream thingies that peeps talk and swoon about but are as often seen and touched as the Loch Ness Monster.

http://www.maxwell.com/ultracapacitors/index.asp

build thread

 

[Nimbuzz]

That all looks very interesting.
1 -- Has anyone here or anywhere tested one?

2 -- Is there a spec sheet that includes size and weight?

3 -- If peeps here are seeing 50 A discharge in regen can these puppies take it in and how much can they put out?

4 -- Is there a price with real world ordering?

Thanks for any info,

 

[TylerDurden]

That all looks very interesting. 1 -- Has anyone here or anywhere tested one?

The forum has a search feature.
It seems to work.
It returns a topic titled: "Ultracap Bike".
You should try using it sometime.

The link to that topic was also in my last post.

 

[Comfortzone]

use circuit breakers - not fuses

 

[amberwolf]

Necro thread REVIVE! :lol:

I've been looking at ways to improve braking on the SB Cruiser trike, which with me on it and a bunch of cargo / groceries or a dog in the back weighs several hundred pounds; braking from up to 20MPH is a lot of inertia.

The leftside generic brushless controller with the MXUS4504 uses an active "EABS" braking where it actually forces teh wheel to a stop, and it's pretty good, but not as good as it could be...and it isn't proportional or analog, it's just on/off, without any ramping I can determine. That's pretty hard on the axles and dropouts.

The rightside "grinfineon" on the MXUS4503 (usually--presently an HSR3548 till i fix the axle of the MXUS) just uses regular regen, also just on/off. But it's not very powerful.

Even with one brake lever controlling both, braking on the generic is so much stronger than the grinfineon (regardless of which controller runs which motor) that it pulls hard to the side the generic is controlling. It's controllable, since I"m expecting it...but annoying. Plus, it just doesn't give enough braking on the grinfineon side.

Since for now getting two identical powerful controllers with proportional active-EABS-style braking (not just regen) is out of the budget (as is getting the rest of the parts for the two lebowski boards I have), I'm basically looking for something I might be able to implement with the stuff I already have laying around here, to improve the electric braking.

(I also am working on adding mechanical braking to the rear wheels, but that's a long project requiring redesigning and rebuilding the frame around them.)

The plug braking options aren't what I'm after, because I want control over the amount of braking (something none of the controllers I have now can do).


The most likely option I've found so far is the below; I could use the ebrake lever I'm using now to pull a cable controlling a throttle to control the level of braking, and the ebrake switch itself can be used to disable both motor controllers and enable the braking controller (via a relay).

The difference is that I would need to use at least a 12FET or more controller, because I need to brake *two* wheels at teh same time, and don't want to add two more controllers/etc to the trike to do it. So I'd run each phase wire to at least two of the FETs, with their drains separated out from each other for each phase wire, as noted below.

At the moment, I think the only working/useful brushed controller I have is an old Curtis golf cart controller (huge), but I think it's a 36v type. Been a while since I've laid eyes on it, though I know where it is in the shed. I may have a tiny cigarette-pack-sized brushed controller in a scooter, but that would need a complete powerstage add-on to be useful for this. (though...I could use a brushless controller's powerstage/heatsink, rewired, added to the brushed controller.


Thoughts?

Chances of smoke?

Update! Fechter, seemingly at my prodding, has been working on the regenerative braking system some more. Here's a summary:

The "new and improved" regen setup avoids using the bridge by using a brushed controller that has 6 FETs (it's what I had lying around). Something with 3 FETs would work too.

The FETs are normally all in parallel in a brushed controller, so I separated the drains and kept the sources connected to ground. Other than that, it is the same as the bridge rectifier setup. The advantage is there is no bridge, which would not perform well under switching conditions, so the amount of power actually recovered to the batteries would be more. A bridge rectifier would need a large heat sink, which would be eliminated by going direclty to the FETs.

 

[fechter]

I blew up some stuff testing way back then, but I think it was from stupidly miswiring something and gave up on it.
Grafting the power stage from an old brushless controller to the PWM drive from a cheap brushed controller might work OK. The biggest danger is that thing coming on when the main controller is running, which will result in a plasma ball somewhere.

Also keep in mind the main controller has to stay connected to the phase wires when the regen unit is running as the body diodes in the controller FETs are being used to circulate the current back to the battery. This will cause a lot of heating in the main controller during braking. On a long downhill, things could get pretty hot.

 

[amberwolf]

Thankfully all my rides are essentially on the flats, with the braking mostly needed at the many complete stops I have to make at traffic controls during my commute. But whenever I'm on the road, I'm almost always riding in traffic, and sometimes other traffic does stupid things I couldn't predict (to just avoid them by slowing a little, etc) or read their "body language" for. Most of the time braking is what I need to do to avoid them, and it's usually very hard braking; what the trike can do now isn't sufficient (well, it's enough to have avoided everything so far, but some have been close and . Acceleration is the other thing I can do to avoid them, and right now that's sometimes safer than braking, but it's not that powerful or quick, either, at present.

I can improve heatsinking on controllers relatively easily if it's necessary. The EABS type generic controller already gets pretty warm just during a regular commute (so does the motor it's on), from the repeated braking.


I know the results of a failure to shutdown either set of FETs in time is probably a big fireball, possibly in both controllers. And then since the FETs will rpobably be shorted, the wheel would be "locked up" cuz the phases would all be shorted together, and I'd be stuck sitting there in traffic unable to get out of the way...very bad.

Unfortunately I cna't think of any protections I could install that would save FETs; they'd pop long before fuses would, though fuses might work to unlock shorted phases if current flows long enough to blow them.

In theory :lol: as long as my ebrake switch disables the main controller before it enables the braking controller, it should be "fine". (where's that puff-of-smoke-smiley? )

I'd guess the "ignition" line that disables the MCU and gate drivers would probably be better to use than the ebrake line (which I'd ahve to go into the controller and find the pads that enable regen, and disable taht, so ebraking would only cut off the motor, and not do any form of braking).

I have a "meh" generic basic brushless controller that I can use as a test guineapig, before I use the idea on the trike's actual controllers.

 

[Oith]

Hi! I know i'm a little late to this thread, but i'm kinda confused on fechter's diagram, about how the regen controller is able to get power to control the pwm if there isn't a complete path for the B+ to go to it's negative terminal(blocked by the rectifier).

 

[amberwolf]

Hi! I know i'm a little late to this thread, but i'm kinda confused on fechter's diagram, about how the regen controller is able to get power to control the pwm if there isn't a complete path for the B+ to go to it's negative terminal(blocked by the rectifier).

Do you mean this diagram:

and the diode in the B+ to M+ connection on the brushed controller, but no B+ connection from controller to battery?

Depending on the brushed controller design, the M+ and B+ can be internally effectively the same, with M- and B- being the ones "modulated" by the controller. Here's a random simple brushed controller schematic from a google image search to show this:



BTW, many cheap brushed controllers don't actually do current limiting (or not well enough to consider that they do it) and can just blow up on overcurrent. So I recommend making sure that the one you use really does do proper current limiting. (and that it limits to something the battery can handle).

 

[Oith]

Thanks for the reply! Sorry but I still have a hard time understanding how the controller itself is powered? I know that the B- and M- are the ones getting shorted by the mosfet similar to the brushed controller diagram you've sent. But i still don't know how the controller gets its power if we've only connected the B+ of the battery? Don't we need the B- of the battery in order to complete the circuit?

 

[harrisonpatm]

Just came across this thread. Any reason why this
https://www.aliexpress.us/item/2251800321247056.html?spm=a2g0o.productlist.0.0.72c1d0c3VELhNa&algo_pvid=9129d3e2-f71d-49c3-bd49-d3b8314b6f78&algo_exp_id=9129d3e2-f71d-49c3-bd49-d3b8314b6f78-5&pdp_ext_f=%7B%22sku_id%22%3A%229092260041%22%7D&pdp_npi=2%40dis%21USD%2120.0%2120.0%21%21%21%21%21%402101e9ce16647238142132750ead0c%219092260041%21sea&curPageLogUid=zhbGHl4ycElJ
and this
https://www.aliexpress.us/item/3256801257160815.html?spm=a2g0o.productlist.0.0.5af32022uNUbgq&algo_pvid=3fbf7f66-99c0-42c6-95c6-d0fe9079823e&algo_exp_id=3fbf7f66-99c0-42c6-95c6-d0fe9079823e-2&pdp_ext_f=%7B%22sku_id%22%3A%2212000026698674879%22%7D&pdp_npi=2%40dis%21USD%2110.32%2110.32%21%21%21%21%21%402101e9d216647238740686840e6aea%2112000026698674879%21sea&curPageLogUid=8Ur1ZnVK1bwq

won't work? Pretty high amp ratings, not very expensive.

I'm also wondering, if you connect this as suggested, your main BLDC motor controller is essentially paralleled with the rectifier. Does enabling regen in this fashion risk backfeeding into the BLDC motor controller and damaging it? Presumably not, seeing as these are designed to push current back and forth at high amps, they've got reverse current protection up the wazoo.

 

[amberwolf]

According to this post from Fechter a couple posts above your first one in this thread, the main (BLDC) controller does this:

Also keep in mind the main controller has to stay connected to the phase wires when the regen unit is running as the body diodes in the controller FETs are being used to circulate the current back to the battery. This will cause a lot of heating in the main controller during braking.

I suppose you could use a second three-phase-rectifier set to output back to the battery instead.

However, at high enough speeds for the motor's generated voltages to be higher than battery voltage, at the two-lead output of the second rectifier, braking would automatically occur, and so it has to be able to be disconnected from the circuit if you don't want that to happen.

So it would be switched onto (connected to) the motor phase wires as the main controller is switched off (disconnected from) the phase wires, but the switching mechanism (presumably a 3P2T relay, and associated relay driver engaged by the brake switch) would add some complication (and extra points of potential failure), as well as a little bit of extra resistance in the controller-to-motor phase current path. The relay would need contacts that can handle the full normal phase current load easily, as well as whatever regen currents might be expected.

Alternately, use a 2P1T (DPST) (or even 1T1T (SPST) relay or switch (on the brake switch) that just disconnects the battery side of the rectifier; this needs to handle only the regen battery current, and never handles any "normal" motor current--if a problem develops with it then it only affects braking and not driving the motor (unlike the one in the phase wires).


Note that I haven't tested any of the above, just coming up with ideas in a few minutes as I work on my once-a-week morning coffee.

 

[amberwolf]

Just came across this thread. Any reason why this
https://www.aliexpress.us/item/2251800321247056.html


https://www.aliexpress.us/item/3256801257160815.html
won't work? Pretty high amp ratings, not very expensive.

I'm sure the rectifier would work (anything that will take the generated voltages and currents will; you'll still need a heatsink for any rectifier, capable of dissipating the accumulated heat from braking).

Whether the brushed controller would work depends on its' actual capabilities vs the claimed ones, and the voltage you'll be generating out of the motor. Since this one is only for up to 14.8v, it won't work for any typical hubmotor system. Note also that the high current rating is only for burst / instantaneous startup current, and not for continuous. The 160A is only a 35A, 320=60, 480=80.

I'm also wondering, if you connect this as suggested, your main BLDC motor controller is essentially paralleled with the rectifier. Does enabling regen in this fashion risk backfeeding into the BLDC motor controller and damaging it? Presumably not, seeing as these are designed to push current back and forth at high amps, they've got reverse current protection up the wazoo.

As Fechter notes in the bit I quoted in my post above, it specifically uses the BLDC controller's FETs (as diodes) to backfeed the current to the battery.

It's not drawn in the schematic, but it is effectively the same as a second three-phase rectifier wired to the phases on it's input (just like the first), and it's output connected to the battery + and -.

And no, there is no reverse current protection in any of these controllers. (if there were, regen wouldn't work, on those controllers that support it natively, which is most of them these days, even if it is only on/off and not variable).

There might be some controllers that detect reverse current and shutdown operation, but it doesn't stop the current from flowing (because that's thru the body diodes of the FETs that there is no control over). Unless the controller has control over a disconnect mechanism for the battery itself, like a contactor, etc., it can't do anything about that. (if it did, it would effectively be a self-kill-switch, since the controller would then have no power to it to turn itself back on; would require user intervention or an external wakeup circuit still powered by the battery). Additionally, shutting power off this way disconnects the load the battery puts on the generated voltage spikes from the motor, so they could go high enough in voltage to blow up the controller FETs or caps or it's LVPS (since it's still connected to the motor).

No reverse voltage protection either, so if you hook up a battery backwards, you're shorting across the battery directly thru the controller's battery-shunt thru the FET bridge body diode pairs (two diodes in series, as many in parallel as half the number of total FETs--which is why a bigger controller is more likely to survive a very short reverse-connection event than a smaller one, at least for the FETs. The shunt may still desolder itself, capacitors may still explode, and the LVPS may still be destroyed).

But the lack of reverse voltage protection is also what allows this whole "trick" to work, without other external parts (see my other post above).

 

[harrisonpatm]

Alright, alright, alright... more thinking required. Hmmmm....

 

[amberwolf]

FWIW, I still have this on my "someday" list of things to try...but that list is pretty long, and I already have all the parts to do more than a few of the other things I still haven't gotten around to, while I'd have to get parts to do this one. :lol:

Since more and more controllers are getting variable regen of some type, there's less and less reason to do it; the Lebowski/honda-IMA hybrid controllers I still need to finish and test would obviate the need for this for me.

 

[fechter]

These days you can buy a controller that has regen for pretty cheap.
I found that using a lot of regen really heats up the motor. If you’re riding on steep hills it can become the limiting factor.