Two-Speed Regen

Kingfish

Kingfish

100 MW
Joined
Feb 3, 2010
Messages
4,064
Location
Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Mil
Greetings
For months I have be intrigued with the passive regen from my Infineon 6FET; without any special modification I would pick up a few tens of a volt on a long downhill; nothing to write home about. Finally this weekend I got around to shorting the BK to GND pad on the reverse side of the board to enable (pure) regen braking/ebrake via a momentary switch. The results from a diagnostic hilly ride went as follows:
  • From dead-roll at the top of my hill (>6% grade), regen builds to about 250W until the speed exceeds roughly 16 mph (boundary condition). After reaching the boundary the regen drops between 1/3 to 1/2 and the speed instantly accelerates as if letting off from a good braking effort though still with slight drag remaining.
  • On subsequent downhill tests at various speeds up to 34 mph light regen was noted until the vehicle was slowed below the boundary condition which led to an immediate doubling (or tripling) of the regen braking. This effect was consistent through my entire 25 mile hilly test.
System:
P0 configured as 10S3P. CA recorded trip params as...
Code: 
Ah      Start-V   End-V   W-hr     Wh-mi   Regen   R-Ah   Fwd-Ah	
13.31   40.9      32.1    498.49   19.4    3.70%   0.493  13.809	
Amin    Amax      Vmin    MaxS     AveS    Time
-14.3   33.4      29.6    34.7     21.3    1:10:09
(Start-V and End-V are my notations.)
Max Regen: slightly less than 500W @ 34 mph
Regen on flat, pedalling hard above boundary condition @ > 20 mph ~= 200W (to determine External Power)
Other factors -> ASV file:
Code: 
1:EB806
60
24.0
30.5
0.1
99
1:Cycle Mode
115
80
60
1.0
2.0
1:Only fake Indicate
1:Comm GND
2
15
Notes:
  • Last week I went for a ride with P0 configured as 10S9P. Small changes to AVS were Phase = 75, Rated = 30, Speed-1 = 110%. MaxW ~= 1356, MaxS was > 34.5, climbing steep hills at 26 mph, but range reduced to 70 miles.
  • Resolved to increase range by dropping Phase to 60 and Rated to 24. Starts from dead stop were slower though continued to build to MaxS; did not detect a difference between setting Speed-1 from 110 to 115, nor were any depreciating artifacts noticed.
  • With the previous weeks’ configuration there was a profound jump in max speed in WOT as the controller hit the 110% rating; this was hardly noticed with the lower Phase/Rated current settings likely due to the WOT Wattage being close to 860-990 on a flat.
Concerns:
Last week I noticed that when at WOT with Speed-1 at 110 and climbing a steep hill that if the grade became steeper and causing a loss of momentum such that the speed dropped below 25 mph the motor would drop-back and suddenly loose 4 mph, as if the forward-winding charge was lost.

Therefore, does setting the Speed-1 above 100 affect regen, creating a two-speed difference possibly in the same manner of how setting the parameter above 100 charges the forward windings?

Observed: Slowing below the boundary condition profoundly increases the regen effect, and yet when above the condition the forward charge is less capable, weakened, lacking full field strength.
Make sense? Does anyone else experience this phenomena?

For the record:
The only mods to this particular controller are power/phase trace-fattening (10AWG), Battery input wires to 10AWG, and replaced three 63V caps with 100V.
~KF
 
Though I won't say it as eloquently or as technically as you have, yes, my 12fet Infineon has a distinct border between two different levels of regen. I notice it at just under 30km/h (If I had to pin an exact number I would say 26-27) It's become an interesting challenge when coming up to a stop sign and trying to get the braking just right so I stop in the right spot (Not that I am going to overshoot, I can just squeeze the mechanical brakes a little harder, but I enjoy trying to get it just right)

Note that I have not programmed or altered this controller in anyway other than shorting BK to Gnd. It's a stock controller from Ebikes.ca. (The 35A) I am however quite curious about it, so I will ask Justin about it tomorrow when I see him. I'll then post his response.
 
I have the older infinion based controller, but I do not notice this effect. My controller when turning on regen experiences the highest amps put back into the batteries if I am traveling at top speed (45mph), and hit the regen switch. The current gradually declines as the speed decreases, and there is no boundary speed where it kicks in harder.
 
OK, I ran more tests today. On a long downhill the Boundary Condition between full regen and part-regen is about 15.5 mph which is near 25 kmh as Demosthenes has discovered. The precise moment of shifting is difficult to ascertain without a data logger.

While flying downhill at a speed roughly equivalent to 30 mph on a constant grade I rolled the throttle forward and received a welcoming increase in regen. I reapplied the test with new interest:

  • Long downgrade at 30 mph/48 kmh, depressed momentary switch shorting BK to GND and no throttle, Regen = > 220W.
  • At ½ throttle Regen => 450W, and
  • at full-throttle Regen => 600W.
  • Slowing below the Boundary Condition is equal to full-throttle.
  • Retested from somewhat fast speed to complete dead stop with full-throttle the whole way and experienced full-regen with constant drag.
It would be very interesting to test this again with default Speed (2) set to 100% at power up to see if the Boundary exists, or with a Three-Way switch selector using different percentages.

Regardless, it makes for a more interesting and coordinated braking experience, if not more powerful and useful one :wink:
 
Not sure if this completely related, but.... With the help of MWKeefer I was able to set up regen on my E-bikekit controller. Really neat stuff. Additionally, he explained that while gripping my ebrake to engage regen, if I roll my throttle, it gives a variable braking effect. It did seem to increase the Amps/Watts coming out of the wheel. I have my controller programmed to only produce regen from the ebrake, not throttle(Only Fake Indicate).
 
I have never experienced any regen effects like this either, and I am using 6 and 12 FET 116-based controllers (from keywin). I do remember reading some similar description on Justin's web site though, one the page about the infinion controller. I found it bizarre at the time, but I guess that some controllers have a slightly different firmware version than others.

Kingfish - I find it odd that your controller looks to be programmed to not use throttle-based regen (1:Only fake Indicate), but it does seem to be using it anyways! Does it use the throttle to regen only when you jumper the BK pad to GND? Nothing done with the EBS- pad? And where/when did you get this controller BTW?

With my own 116-based controllers, shorting BK to GND enables fixed regen by using a momentary switch on the EBS- pad (but it cuts off throttle action while doing regen like this). What I *really* want to have is throttle based (analog) regen activated by a jumper... and your setup almost sounds like this unless I'm mistaken! Your throttle works as normal (you can coast) when you're not ebraking, is this right?

Pat
 
ZapPat,

I have the 846 chip on mine, not the 116. Could this be the reason for functional differences? I thought the 116 was more advanced... Also on mine, when I engage regen and then let totally off the brake, I get drag on the wheel until I use the throttle again.

I'm really surprised E-bikekit.com does not open the full functionality of their controllers to their customers. Or at least have an option to pay for such functionality if they have decided it would not be cost effective at their regular price. Maybe they are worried about the stress regen puts on the forks/frame? Dunno, but its awesome to have its full functionality available!
 
Gents, ZapPat, et al

  • Yes, I shorted BK to GND to enable Regen.
  • I am using the –EBS pad to GND connected to a momentary switch to enable regen/drag-effect. The pigtail coming out of the controller was setup for use with an ebrake and came that way from ebikes.ca. The connector is a 4-pin with only two contacts inserted, and that's what I am using to hook to the switch.
  • Slip Current Charge mode is set to the default “1: Only Fake Indicate” value since I wanted to freewheel when off-throttle. I have never tested it with the other param as it reminded me of how my tractor throttle functions: yer either power-on or braking.
  • When rolling the throttle forward while shorting –EBS to GND there is definite increase in wattage/voltage produced. I cannot say for certain that all of the regeneration is going into the battery; I need to test more.
Conclusion: I am definitely getting more regen by using the button than with passive regen (no button). Last night coming back into Redmond from Bellevue there is one big long downgrade and I used full-throttle-regen all the way down and picked up 0.4V which has never happened before. Passive regen would give me 0.1V. Button-only I think was 0.2, but more than passive. I just need to test full-throttle-regen more to see that the energy produced is really going into the battery and not to waste heat. The power spikes during regen (CA reports it so), and when letting off regen-throttle the power takes a second or two to settle down adding to the small challenge of determining the actual captured amount.

=> The CA says the hub is making a lot of power... and I tend to trust the unit. 8)
 
Update:
The last few rides I have mixed it up a bit
  1. Regen without Throttle; Passive
  2. Regen without Throttle; Drag (Short –EBS to GND)
  3. Regen with Full-Throttle; Variable (Short –EBS to GND and roll Throttle forward to MAX)
Conclusions:
  1. Passive generates the least (we knew that).
  2. Drag: Regen w/o Throttle developed notably more than passive, but it’s not very impressive (don’t bank on it).
  3. Variable: Regen with Full-Throttle is substantial - providing the maximum effect, and it contributes a modest amount of braking energy, though I wish that the amount was greater.
Of the three, Regen w/o Throttle is great for a long steep downhill especially where momentum is increasing and adding a bit of Drag isn't a hinderance. On one such ride I was headlong downhill busy managing my station with traffic when I looked down and saw I was doing in excess of 35 mph and still picking up speed: Perfect for scavenging a bit of energy.

However Regen with WOT can be slightly more than a thrilling experience at times if the timings’ not right :shock:

More testing:
I would very much like to understand if there is a relationship between advanced timing, field strength, and regen. The next series will use various levels of Speed-x settings; however that will have to wait until I get P1 up and running since it has the 3-Way Switch. I already anticipate that 2WD regen would essentially double the effect all around 8)

Regardless, I am glad that rolling the throttle forward during regen does pump power back into the battery and does NOT convert it to waste heat (thankfully at least not all of it).

Hmmm, I might try building a static test stand: I have several controllers, hubs, and a clutched electric motor that would do nicely for spinning the hub. Dang! Now’s when I wished I had a data logger. Anyone have a lead on a DIY for the CA?
 
http://www.ebikes.ca/drainbrain/CA_Logomatic_Hookup.pdf
http://endless-sphere.com/forums/viewtopic.php?f=2&t=14397
http://www.sparkfun.com/commerce/product_info.php?products_id=8627
 
In all th stalk no one has yet mentioned the voltage of the pack they are using..and the re gen voltage set in the controller software.

That is going to have a great effect on regen . I have an 84 volt pack...and since the maximum I can set via the software is ...I think 75 volts, re gen , either throttle based or switch based does not work at all.

I am yet to do the R12 mod to allow regen to work
 
R12 mod for high voltage regen

At 84V, you will be hard-pressed to crank it up much higher without impinging upon the FETs lower-threshold.
For myself – I run 15S LiPo at 63.3V hot. LVC is set to about 49.5V although I rarely let it get there ever (actually I get a bit nervous at 54V because the pack just drops like a stone after that).

I bought a 2k Ohm resistor to parallel with R12.

If I program the new 12FET 100V Lyen controller with LVC = 31V and HVC to 55V, then with the resistor mod, LVC goes to 49.6V (good enough) and HVC goes to 88V.
That’s presuming my R12 = 1200. It is not; I measured it to be 1160 to 1170. That small change can have a dramatic effect depending upon the tolerance of the 2k resistor. The best practice is to make the mod, measure the new value, and reprogram accordingly.

Thus I presume within a degree of confidence that at 1170 + 2000 (paralleled) = 738 Ohms. With Hyena’s most-excellent spreadsheet, we can prefect the calculation wherein LVC = 31.5 and HVC = 60, lending a new profile to be LVC = 49.9V and HVC = 95.1V.

Very edgy. Above the 20% tolerance of the 100V 4110 FETs. I am taking a big risk. If I blow them, well, maybe I should upgrade to the 4115.

So for you at 84V, maybe the (presuming) 4110 FET is not for you if maximum regen is in your gameplan :)

Best, KF
 
Just after some regen -braking really...I have 20S Nano Tech LiPo, and as I discovered today, 70v is my best bet for LVC level...I only charge to 83 volt (4.15) so looking at that as the regen voltage.

Lyen has the controller at the moment,as I cooked a couple of the FETs I have the 18 FET 4110 controller.

What are the limits for the 4115's? higher voltage and higher current?
 
Well it is all a bit beyond my understanding...but looking at the 115...it is only 105A to the 4110's 180A.

So what is the answer for me? With a Lyen 18 FET controller with 4110's ? can I do re gen at 83 volts or not? and if so what is the required value of R12 and what should I set the software to ? 83 volt regen and 70 volt LVC?
 
Great info, thanks for this- did not know that you can use regen in passive mode without excessive drag. Need to test this and check the effect to average speed on my ordinary route.
 
NeilP said:
Well it is all a bit beyond my understanding...but looking at the 115...it is only 105A to the 4110's 180A.

So what is the answer for me? With a Lyen 18 FET controller with 4110's ? can I do re gen at 83 volts or not? and if so what is the required value of R12 and what should I set the software to ? 83 volt regen and 70 volt LVC?
Click to expand...
Neil, I am not privy to your setup. Can you tell us briefly what you have:
Motor?
Controller = 18FET 100V 4110, check.
Batts = 84V, got that bit.
Ah?
Limits: What are to wanting to do? Go fast, go far, climber?

Just to put a spin on it, I run in two modes: Commuter, and Cross-Country long distance. Commuter I just want to go fast as I’ve got plenty of capacity to go far enough – up to 50 miles at WOT (> 30 mph). Regen is important to me for cross-country where 2-3% change is a big deal when yer going 150+ miles. At 63V hot, I want the maximum regen and there’s enough headroom that I can get what I want. My phase current doesn’t go past 45A, so the 4110 is overkill, particularly a 12FET controller. Although last summer when I did my long-distance trek the 6FET controller and motor both got very warm on the hill climbs lugging my heavily provisioned bike. This year it’s 2 motors and 2 controllers with twice the FETs so I don’t anticipate the same issues.

Therefore we spec out the controller to meet the needs.

BTW – Alan B is absolutely correct; going to a higher voltage FET gives as much as it takes. I only suggest 4115 for the sake of discussion and technical merit. Also consider that although the amount of current is lower, the voltage is higher; the two go together to provide the total power available. You could always go to 24S/101V and use 20% less current :)

Best, KF
 
OK, I'll tell you what I have. But I am not after changing anything, just want to know what R 12 value and what regen setting I need to get some regen braking. Brake saving rather than all out max recharge


5304
Lyen 4110 18 FET
Nano Tech LiPo 20 Ah made of eight 10S packs 4p20s

Speed...35 to 40 is fine, our speed limit on the island is 40 mph anyway .... And officially 15 mph electric assist
.range...well what I am getting is great at present....between 30 to 35 miles.
I only want the regen braking as a braking effect...anything it puts back to the pack will be a nice bonus


I was getting 35 to 36 Watt/hr per mile, in a Trek fram.but when I cooked the Lyen controller, a I bought a 48Amp Crystalyte controller....and changed the bike frame too...power usage has now gone up to 42 to 44 W/h per mile. Not sure if that is a controller or frame change issue though
 
Interestingly, I have now verified this two-level regen issue exists with the Ecrazyman EB812 controller I've got on CrazyBike2 as well. AFAIK it's an 846-based unit, but I can't read the chip number even with the best light I have; have to find the magnifying glass wherever I put it. :roll:

I found this thread again (saw it before and forgot about it), because I thought something was odd at first, after I hooked regen up finally a few days back. Last couple of work commutes I verified that it gets less amps when braking from above ~15MPH than after speed drops below that, and I'm pretty sure I can feel that the braking becomes significantly greater below that speed than above.

I am only using the regen via EBS short to ground via a Stop-Run-Stop switch on the ex-scooter controls, with the regen pad shorted to ground inside the controller to permanently enable it. AFAICT I don't have the option (without programming it, if possible) of variable regen via throttle (although it could be useful if I did, if regen braking was hard enough to warrant it (it's not, at the moment)). I tried a few variants of operating the throttle vs the brake lever, and none of them do anything beyond just activating the regen at the same level every time, with the brake lever. Either I'm not doing it right or it's simply not working (or not activated).

I am replacing my Stop-Run-Stop brake switch with an actual brake lever to make it physically easier to brake and throttle at the same time, so my next work commute will show me if that is all the problem was, or if it really doesn't have variable regen on this unit.


Battery is a Vpower/CammyCC 48V (59V fully charged) LiFePO4 pack that does support regen thru the BMS (charging is done thru the same port as discharge).


Anyway, the real problem I have is that this 2-speed regen thing, with lesser braking at higher speed is a problem...I wonder if there is a way to defeat this?
 
Hmmmm
 
IN the last 3 weeks i have been testing for the first time the Regen at HIGH POWER.

You know me and i was always worried about regen because with the wrong torque arm setup it could turn out into DANGEROUS dropout or axel problem by making the nut to loose and finally break.

NOw with a total of 1/2" of steel on each side of the motor for the torque arm, i can use regen at the power i want WITHOUT any problems.

Guess what?... I dont use any lock nut and the nut is alone and on both ebike i activated the regen, they are still as thight as when i thightened them last time.

I have done the R 12 mode bu replacing the 2200ohms resistor by a precision 910 ohm.

That alow me to set the LVC to a conservative 86 volts and the regen HVC to 99.5V ( i use 100V max on both ebikes)

On both 18 fets controllers the BK to Ground is linked via an external jumper out of the controllers.

I use the EBS- to ground connected to an ebrake grip with the built-in switch and i adjusted the cable tension so that i can activate the regen without having he disk brake having any friction applied. that allo me to choose first the regen and than if i want to have stronger braking force, than the brake cable become thightened bt the grip and the normal friction brake work in second place.

I used regen and saw on live reading a power up to 1900watts back to the battery. But my C-A recorded -39A reverse current.. so it might be also 3800watts or so...

I used the regen at really high speed ( 72kmh down a hill with my mongoose with the 5305 on a 22" motocross wheel)

I still never had problem with regen at voltage set up to 99.5V and power up to 1900watts.

if you wat to increase the regen power you can modify the shunt of the controller. That will have a direct and proportional effect on the regen power.. but also on the overall power in our out of the controller..

To give you an idea, my ebike and me weight 290 pounds together.

When i go at 70km/h and use the regen to stop me, it take less than 100m long

This is a great feature.. but on my average ride style, it represent only 2 to 3% energy back to the battery

Doc
 
Well, the brake lever doesnt' change the problem. I thought it did at first, but it doens't now, so I'm not sure. When engaging ebrake and then throttling up all teh way it doesn't feel any different, though at first I saw a peak regen amps of almost 16, vs the 8 or so I got before. Must've been a glitch or coincidence, as no manner of operation or combination of brake and throttle changes anything. :(

Oh, well. :roll:
 
Noticed that you had used 115% as maximum setting. When I try to set anything above 100% all the values in the configuration file are not anymore updated and are ignored from user interface.

Does the values over 100% really work regardless of the above ?
 
You must log in or register to reply here.

Similar threads

R
Battery current vs. phase current --- is it always MC x 1.73?
Replies
0
Views
629
richj8990
R
E
90's Model Suzuki GSX1100 Shaft Drive
Replies
4
Views
382
E-GSX
E
R
    • Like
First Electric Motorcycle Conversion Advice (1974 Honda CB360)
2
Replies
38
Views
1,357
harrisonpatm
H
J
New Bafang SWX02 48V500W with 30A KT Controller and LCD03 Max Speed Issue.
Replies
12
Views
675
AGS
AGS
A
  • Question
golden motor edge regen braking voltage setting for cycle analyst
Replies
4
Views
760
E-HP
E-HP
Back
Top