Smart Pie Regen

[Kris]

I've put a 48v Smartpie with in wheel controller as the rear wheel on a SideWalker Micro scooter. It works great but the brake lever operated regen is so strong it's unusable. The regen appears to be either on or off. Is it possible to get variable regen worked of a throttle type control?

[Bikemad]

Hi Kris andto the forum.

Unfortunately, I haven't been able to persuade GM to incorporate variable regen into their ebike controllers, but hopefully this will happen eventually.

I have seen posts over on Endless-Sphere where others have used an additional speed controller and throttle unit to enable the use of variable regen, but his might be difficult to implement with the Smart Pie's internal controller, but you will need to look into this further yourself.

If you have the programming cable you can try reducing the regen to a more acceptable level for your smaller wheel.

Another option would be to modify your wiring so that the brake switches just cut the throttle instead of applying regen, but the brakes would also need to be connected to the cruise control button to ensure power to the motor stops when you brake, even if you're using cruise mode.

The basic circuit is reasonably simple and you should only require two diodes and a 330 ohm resistor:



In your case, you would simply leave the regen wire disconnected and would not require the switch.
Even without the switch, it may not be quite so easy to install into the Smart Pie's wiring as you would have to cut into the nicely sealed main harness and locate which three of the eight wires you needed to use (Brake Switch/Regen, Throttle Signal and Cruise Signal)  in order to make the necessary changes to the wiring.

As you wouldn't be using a switch, it might be possible to install the components inside the controller housing, but I suspect that could be very fiddly connecting them directly to the controller.

Alan
 

[Kris]

Thanks for the reply Bikemad. Why don't the tech wizards at GM not put variable regen on the first 20% of the existing throttle leaving the other 20-80% for speed control? That way backing off the throttle from full power would give coasting when you got down to 20% then increasing levels of regen to full regen at 0% throttle? No hardware changes required just a software upgrade?

[Bikemad]

If it were feasible to do it that way, I think it would be virtually impossible to keep the throttle at the 20% position in order to ride on the level without motor assistance, or simply freewheel downhill without braking.

I can only imagine what might happen if you were riding along and had to release the throttle to indicate a turn etc. the last thing you'd want was 100% regen when you were approaching a junction on a busy main road whilst holding on with only one hand.

For this to work, the throttle would have to be spring biased to the 20% (or possibly mid-way position) instead of 0%, but I'm not convinced that this could be implemented purely with a software modification to the controller, as I think it may be a little bit more involved.

Although it would be a good solution for use with a twist throttle, I think it would be very awkward to use with a thumb throttle.
Perhaps a spring loaded button with a magnet and hall sensor arrangement could be used to provide proportional braking instead?

Alan
 

[Kris]

Given a suitable device BikeMad, a separate twist throttle or brake lever modified to have a variable resistance output is there any chance that the current controller regen input might one day accept continuously variable analog input rather than digital on/off?
I'm desperate to get variable regen going on this puppy. What would happen if I did PWM on the current on/off input do you think?

[Bikemad]

If GM want to stay at the top they will need to incorporate variable regen and introduce a decent digital display unit to compliment their motors and to allow some form of user control over the pedelec function.

I would also like to see the cruise function improved so that you could set the cruising speed, which could also be displayed on the screen, with the aid of an up/down button and a set/resume button like the car systems have.

A mode selector button could be incorporated to allow the cruise function to be linked with the variable regen to enable the option of a fixed constant speed on the level, uphill and downhill, as I previously suggested in this post back in August 2009.

PWM switching of the current is unlikely to do anything useful, as I've noticed that the electrodynamic braking effect still works at higher speed, even with the battery completely disconnected, and below 10mph there is no charge going to the battery during braking, so there would be no current to control via PWM anyway.

Alan
 

[Bikemad]

I have seen posts over on Endless-Sphere where others have used an additional speed controller and throttle unit to enable the use of variable regen, but his might be difficult to implement with the Smart Pie's internal controller, but you will need to look into this further yourself.

Here's a link to the posts that I was referring to on Endless-Sphere.
But after studying Fechter's diagram on the 10th post, I'm pretty sure that the "Cheap Brushed Motor Controller" would need to be of a variable regen type to allow the amount of regen to be controlled.

Alan

[Morgen 3Eman]

Hi Folks,

Just to look at this opportunity from a different light, I'm going to put my two cents worth of intellect and experience into the discussion.

I played around with regen on my Morgen,  and I came to dislike the feature immensely.  It was so damned annoying to have the variable effect of regen interfering with my efforts to brake smoothly.  The effect was also noticable on the Prius I rented on vacation a few years ago.  I will admit that I am incredibly fussy about brakes because of the cars I have owned.  I can't imagine how horrible braking would become with some magically automatic feature involving throttle position also influencing the  variability.   

The other side of regen is the attraction of increasing range per charge, but my Morgen experience showed such a small gain that it was a waste of time to keep fiddling with it to make the braking effect better.   Clearly there is a much larger opportunity with a car sized vehicle with much greater mass and speed.

I do however agree with Alan about the cruise control arrangement with a bult-in display.  It would seem to be an absolute no brainer to implement such a scheme.    There is already a USB data line, and 5VDC to support the idea.

TTFN,
Dennis

[Kris]

Sorry by 'current' I meant the 'existing' brake switch input, if I pulsed it rapidly with a solid state switch might I get some control over the regen? My ammeter is telling me regen on the smartpie is pushing current back into the battery over the full speed range.

[Morgen 3Eman]

Hi Kris

Electronically, your idea would work.  You could have (for example only) a left hand twist grip that controlled a  low frequency pulse generator to generate a variable pulse width to ground the regen line, and keep the hand brake signal line intact for hard braking situations.   It would be pretty easy to implement.  Might I suggest you have a bit of a dead band before the control becomes active?

I find it interesting that you see regen current over the speed range.  I remember an earlier thread where several of us commented that we saw a current pulse when initiating regen, but it didn't seem to continue after that brief pulse.  Maybe the SmartPie regen is set up differently from the MP3.

TTFN,
Dennis

[Bikemad]

My Smart Pie does not generate any useful current below at least 10mph. I can regularly ride for two or three miles with the dog and measure virtually nothing on my second watt meter which is wired specifically to measure regen.

I use the brakes a lot on the downhill stretches and feel the electrodynamic braking cut in almost as soon as I pull either of the brake levers, but the last time I checked the meter at the end of the ride it showed a maximum regen current of 0.43 Amp (13.5 watts), and indicated a total of 3mAh of battery capacity (0.003Ah) had been returned though regen, and that's with the regen set at 100% in the software setting.

Alan
 

[Kris]

OK, maybe I need to look into where the regen current goes more closely? I'd seen threads that said earlier (external controller) models simply shorted the back emf to ground (so called dynamic breaking) so I just put an ammeter in the main battery line, spun the motor up on the bench before it was even fitted to the scooter then checked the polarity when I grounded the regen line. Three milliamp hour is pathetic. Can you guys recommend a decent coulomb counting display for the smartpie?

[Kris]

I got a cycle analyst fitted & checked the regen current on the SmartPie scooter motor BikeMad, 10A going down hill, easily puts back 12% of our battery capacity except it's way to strong even with the software set to the minimum value. I got a quote from Phil Yao to do us a prototype with variable regen controlled from a pot but now he doesn't answer my emails any more so I guess he has changed his mind. Hell of a way to do business IMHO.

[Lollandster]

I believe the Kelly controllers can do variable regen breaking.

General Specifications:

• Support three modes of regenerative braking: brake switch regen, release throttle regen, 0-5V analog signal variable regen.
• Analog Brake and Throttle Input: 0-5 Volts. Can use 3-wire pot to produce 0-5V signal.

[Kris]

Yeah I been looking at them but their smallest hub motor is 800w (about double the smartpie) & wide (nearly 200mm across) way too big for sidewalk scooter forks.