Joe Perez

iTrader: (8)

Yeah, that's retarded.

The controller itself translates throttle input into requested motor current, which I find to be the most intuitive mode- it behaves like a car or a motorcycle. More throttle = more torque, regardless of speed. I'm surprised yours differs- most controller I have seen behave this way by default.

On this build, I'm going to be using a CycleAnalyst computer as well, which sits between the throttle and the motor (it senses throttle position, current, motor temperature, RPM, brake actuation, and has a couple of other general-purpose inputs. It's sort of like a piggyback ECU.) I am using it principally as a temperature-limiter and to enforce a soft-start on heavy throttle inputs, as gear-reduced motors don't like sudden throttle inputs. It's like dropping the clutch with sticky tires on a 1.6 Miata diff.

This device also supports the option to translate the throttle signal into a speed-request mode, which I find stupid and pointless, though there is a segment of the population who seem to prefer this mode. I assume that they all ride exclusively on flat ground.

I didn't know you had an e-bike. What setup are you running?

y8s

iTrader: (8)

ok ok then you need to build a second bike for crab-walk burnouts.

JasonC SBB

I have a cheapie Currie drive on a cheapie generic mountain bike which I bought cheap on craigslist. 24V lead acid battery. It's adequate for my hilly 3.5 mile commute, so I don't smell like a homeless person when I get to work. I haven't used it in a while, since I busted my ACL and had it repaired.

My controller seems to control PWM open loop to the motor, which is the same a requesting a voltage applied to the motor which is a speed request. I've been meaning to add some circuitry to it to change it to torque (motor current, not battery current) command but never got enough Round Tuits. I just live with it. The annoying thing is that the controller is built into the motor, which makes it a PITA to modify.

Joe Perez

iTrader: (8)

You're sure it's a Currie? All the ones I've seen used external controllers and motors which are external to the wheel hub. So far as I'm aware, only the Golden Motor "Magic Pie" hub uses an internal controller.


Very nearly every controller I'm aware of uses open-loop PWM to control the motor. (You can close the loop with a CycleAnalyst, but I'm taking about basic controllers here.) In such a configuration, I can assure you that the throttle does not function as a speed-setpoint controller, it merely increases the PWM duty cycle. This has the effect of varying torque with throttle, which you can easily see by putting an ammeter in series with the battery.

I realize that power control is your thing, but I'm not certain you're looking at this from the right angle.

JasonC SBB

It says Currie right on the (external, chain drive) motor.

A DC motor's speed is proportional to its "back EMF".
Current is proportional to (applied voltage - back EMF) / resistance
Torque is proportional to current.


Open loop PWM'ing the voltage of the battery produces (constant, regardless of load) voltage at the motor. i.e. the output of the battery + H-bridge equals some constant voltage (reduced from batt volts). Constant voltage on a brushless motor = constant speed characteristic.

If the actual speed is below said "constant speed" it delivers lots of torque at zero speed then diminishes as speed goes up. This characteristic only has slight "rubberiness" from the DC resistance.

In order to get a constant torque command, one would have to have a motor current sensing feedback loop, which modulates the PWM % to hit target (commanded) motor current. (Note, not battery current). One could also design the loop so it's not a perfect torque command, but a combo of torque and speed. Gas engines btw are a power command (throttle is roughly a constant airflow device). That's why at low RPM, 10% throttle may deliver max torque.

In a former life I designed AC and DC motor drives.

Joe Perez

iTrader: (8)

Ah, the ole side-mounted chain drive. Yup, that's definitely a Currie. I've just never seen one with an internal controller- every Currie bike I've come across has had the controller either inside the battery housing or as an external box.

Moving on...

Typically, when the e-bike community speaks of current draw, they are speaking of current as measured at the battery. When a controller is specified as having a "30A" limit, this means 30A RMS at the input as measured across a shunt on the line between the battery and the FETs..

For a typical controller, battery current increases proportionately to throttle position. You can see this for yourself by placing an ammeter in series with the battery cable, as I had on the first bike. No two ways about it, current increases proportionate to throttle.


When you say that throttle position is proportional to speed, this is a misnomer. Throttle position is proportionate to power, which does not translate to speed once you account for variance in terrain. There do exist some controllers which allow you to map throttle directly to wheel speed, but these are horrible to ride- you have no control over torque at all, and the thing wants to run at maximum current until it achieves the requested-speed setpoint. Imagine driving a car with nothing but the cruise-control buttons.

That said, I don't want this to devolve into an argument over semantics.

rleete

iTrader: (21)

Why are you so anti-semantic?

blaen99

iTrader: (6)

A better question: Why is Jason so pro-semantic?!?

JasonC SBB

Bike's been packed away since my knee surgery last year so you may be right, controller may have been inside the battery housing as opposed to the motor housing. There's no external box, IIRC.

This can be true for a design with a feedback loop, in particular if the loop is closed around the battery current.

It will be true for open loop PWM, which my Currie seems to be.

This would be true if the feedback loop is closed around battery current. It is not true for my Currie which is simple PWM open loop. When the wheel is in the air, the throttle gives me a speed command. I can regulate speed with the throttle with the wheel in the air.

A battery current feedback loop, or constant power, would be an improvement over open loop PWM. It will still be more sensitive at lower speeds, unlike a true motor current command.

Joe Perez

iTrader: (8)

Yeah, that makes sense.

The Golden Motor "magic pie" series is a direct-drive hubmotor where the controller is actually inside the hub itself. A great design if you are trying to put the controller in the worst possible place insofar as vibration and heat dissipation.


That's because when the wheel is in the air, load is a constant.

But consider the following experiment:

1: Put the bike on a flat, smooth road and set the throttle to 50%. The bike will accelerate at a certain rate, and settle in at a steady-state speed of, say, 15 mph.

2: Now, put the bike on a road with a 10% uphill grade. Set the throttle to 50% again. Now, the bike will accelerate much more slowly, and achieve a final speed of 5 mph. If you then increase the throttle to 100%, the bike will accelerate again, and achieve a final speed of 10 mph.

That's not a closed-loop system, just a stable one. It's like the throttle in your car. 50% throttle calls for 50% PWM duty cycle, but has no absolute relationship to either wheel speed or battery current. You can increase the load on the wheel (by dragging the brakes, for example) and the wheel will slow down, but the controller will not react- it will continue to operate at 50% DC, and the only change in RMS current through the system will be due to the slight drop in back-EMF from the motor.


If we speak of a "speed request" system, then the controller must do whatever is possible to achieve (or attempt to achieve) a fixed speed. Thus, a throttle opening of 50% would call for the controller to operate at whatever duty cycle is necessary to achieve 15 mph (assuming that 15 mph = 1/2 of maximum speed.) This is like the cruise control on your car. Closed-loop, and reactive.

JasonC SBB

Agreed. What you are describing is a constant speed system that droops with load. That droop is due to the motor resistance in the equation in my earlier post.

The resulting characteristic is quite different than if it had a closed feedback loop around the battery current, which would have been better, because at say 10% throttle at zero speed, the latter wouldn't result in max current and thus max acceleration.

P.S. It's not RMS current from the battery that matters wrt power delivery, it's average current (Iavg*Vavg). What the RMS current is good for is for calculating resistive losses in the wiring i.e. Irms^2 *R.

Joe Perez

iTrader: (8)

First, Iavg * Vavg gives you average power, not average current.

Second, if RMS power out of the battery is not within 1% or so of average power delivered to the motor, then you have a serious problem somewhere. At 30-35 amps, any I2R losses in the wiring are going to make themselves know pretty damn quickly. (eg: fire.)


Aaaaaanyway, a large box arrived today from Shanghai! My living room is being over-run with wheels.



The middle rear wheel (with the brake disc mounted on it) is the new one. To the right is the old motor, to the left is the stock rear wheel from this frame.

It's amazing how much smaller and lighter the new motor is than the old one. I haven't weighed them yet, but it's pretty significant.

MazDilla

iTrader: (7)

Subscribed.

Looking forward to the low down the cellman/MAC kit.

y8s

iTrader: (8)

Joe, do you expect the new setup will travel 40 miles on a charge? If not, how far (assuming most hills wouldn't kill you if you had no motor)?

JasonC SBB

Correct. What I meant in my earlier statement was, average power from the battery is NOT Irms*Vavg.

"RMS power" is not a very meaningful number in this context. Average power is. RMS current is useful for calculating I^2R losses. Average current out of a DC source is useful for calculating power delivered.

Consider a 12V battery driving a 1A lamp, that is turned on for 1 sec, then off for 1 sec, repeating (50% duty cycle):

The power while the switch is on, is 12W.
While the switch is off, 0W.
Because 1 sec on and 1 sec off, average power is 6W.
Because 1 sec on and 1 sec off, average current is 0.5A
12V * 0.5A = 6W ... this is Vavg * Iavg at the battery
Battery will last as long having 6W drawn continuously.
The number is meaningful.

RMS current is 0.707A .... (This is 1 A * sqrt(duty cycle)), this is equation for RMS of a unipolar square wave).
Irms*V = 0.707*12 = 8.5W is NOT EQUAL to power drawn from battery

Let's say the wire resistance is 1 milliohm
Every time the switch is on, it dissipates
1^2*R = 1 mW
Because 1 sec on and 1 sec off, average wire dissipation is 0.5 mW.
Let's try the I^2*R equation:
Irms = 0.707 A
Irms^2 * R = 0.5 mW ... correct
If you try to use average current in the wire, the result will be wrong:
Iavg^2*R = 0.25 mW .... wrong

Joe Perez

iTrader: (8)

Ok, Jason is officially banned from this thread for excessive semantics.



So far, it looks pretty good.

The low-current wiring uses the same crappy white connectors as all the other kits, so those will be replaced with Molex Micro-Fit connections. On the plus side, the motor wiring is extremely stout (each phase has two 2mm2 wires, which equals roughly 11AWG, and they are terminated at Anderson PowerPole connectors. The phase wiring on the controller is also quite hefty, although it looks like someone got confused at some point, as the blue wire has yellow heatshrink and a yellow Anderson on it, and vise-versa. The area where the wiring exits the axle also seems a tad more robust than on the Amped kit.

In terms of mechanical quality of the motor, it seems to be a lot better than the Amped unit. The threads on both the freewheel mount and the axle are much cleaner, the freewheel that it came with (DNP 11T-7) seems to be of much higher quality, and the overall fit-and-finish seems better in general. I haven't miced it yet, but the axle itself also seems thicker in the major OD (obviously they're all cut down to 10mm to fit standard dropouts.) I can detect no axle play by hand, and it has a solid, hefty feel to it. I was expecting the internal clutch to ratchet like a freewheel, but it's actually very smooth.

I do note that the spokes used are of a narrower gauge than on the Amped kit, and the wheel, while of seemingly comparable quality (and still double-walled) isn't quite as deeply V-eed as the Amped wheel. it's possible that this wheel might be less robust if you are really beating the snot out of it (eg: aggressive off-roading, riding down concrete steps, etc). Cell_Man does offer two different rim styles in the 26" size (he uses Alex rims) and I deliberately chose the narrower of the two (19mm) as it better matches my slick tires. The wider rim (24mm) is specifically called out as being suitable for downhill trials.

The accessories (throttle, brake levers, etc) are all of typical Chinese quality, just like every other e-bike kit out there. Kind of "meh," but that's hardly surprising.

Tonight I will test-fit the wheel into the frame. I'm a little concerned about what it's going to take to properly align the disc with the caliper- I've never owned a disc bike before, so it'll be interesting.


The battery hasn't shipped yet (it's being fabricated) so we're not quite ready to rock.



I honestly have no idea. The computer I'm using on this build does allow me to monitor cumulative discharge (in amp-hours) versus distance traveled, so I'll be able to give a decent estimate once it's up and running.

Also, note that there are four different winding styles available for the motor, which determine the RPM vs. torque characteristics of the motor. The one I have is biased towards lower-speed, higher-torque applications, so for a flatlander, this motor would not represent the best choice for long-range travel.

40 miles is probably unrealistic on the particular battery I'll be using (48v @ 11.5 Ah.) A lot of folks who are aiming for either extreme speed or extreme range opt to use LiPo batteries as opposed to LiFe. The power density of these batteries is about twice what I've got, meaning you get double the Wh for the same weight. The downside is that there are no pre-fabbed, off-the-shelf LiPo solutions as of yet, so you have to cobble together your own pack along with a charging / BMS solution, and then monitor it like a hawk to ensure that it doesn't burn your house down while charging. If you needed to build a Long-Ranger, I'd probably fabricate a stout steel battery box, pack it full of LiPo, and make sure that there is a bottom-facing vent to direct the flames away from anything flammable when the bike is parked in the nominal charging location.

On the plus side, inexpensive and lightweight LiFe chargers are readily available, so you could always carry one with you and charge the bike while at work, assuming that your 40 mile range represents 20 out and 20 back.

JasonC SBB

y8s

iTrader: (8)

Indeed it is round trip. And given 8 hours of charge time, I'm sure that's feasible moreso than having a Nissan Leaf and a 200 foot extension cord.

Unfortunately though electric bikes are somewhat illegal on the awesome bike trails here. Which is sad because the bike trail is actually slightly shorter than the car route and is almost exactly a 13 mile straight line from my house to my work starting about 3 blocks from home and ending across the street from the office. I'd only share about 1/4 mile with cars and that's on tiny backstreets.

y8s

iTrader: (8)

PS did you edit out the details of the two companies you were debating between for buying e-bike parts? I can't find that anywhere.

Joe Perez

iTrader: (8)

From the folks at ES, I've come to surmise that e-bikes are, to some extent or another, illegal pretty much everywhere on earth.

In practice, it seems that most e-bikers tend not to get hassled provided that they ride in a safe and reasonable manner. IOW, doing 35 MPH down a dirt trail filled with other cyclists and pedestrians will probably get you noticed, though to what degree your local LEO's patrol the bike trails is probably questionable. Out here in north SD county, we have a pretty comprehensive "Rail Trail" system (bike paths built alongside the right-of-way for the light rail trains) and yet I don't think I have ever seen a single bike cop.

Even in Manhattan I see e-bikers all the time (mostly couriers / delivery workers), despite the fact that e-bikes are specifically and completely prohibited in the entire state of New York.

To date, the only folks who have really noticed that I'm not riding a conventional bicycle are the bio-cyclists who I've passed going up a hill, along with the occasional motorcyclist who I've pulled up next to at a stoplight.

I honestly can't imagine that you'd have a problem. Just keep the speeds reasonable on the bike path and, as Ernie the Giant Chicken says, don't be a dick.