2WD (Two-Wheel Drive) FAQ

Kingfish

100 MW
Joined
Feb 3, 2010
Messages
4,064
Location
Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Mil
Greetings

This topic comes up occasionally and deserves special attention to prevent us from repeating ourselves. Much on this subject has been learned since the oldest known post was made on this forum back in April 2008. There is a lot of good useful experience out there, and some disinformation. The purpose of this FAQ is to demystify the concepts, point to relevant topics, and provide concise answers.

Disclaimer: This thread does not address All-Wheel Drives or dual-motor mid-drives, although these are certainly interesting topics that rightfully deserve their own FAQ. :)

What is 2WD?
Two wheels driven by two separate motors and providing forward motion. Used within the context of "bikes", these wheels may be attached to a bicycle or to a tricycle (recumbent or 3-Wheel bicycle with 2 rear wheels or bicycle with sidecar). Motor-driven cycles (other than motorcycles using ICE) may also use 2WD. Technically, the primary employment of 2WD is through the use of Direct Drive (DD) hub or Geared-Drive motors, although not exclusively: The two types may be used in combination. Brushed motors may also be used although they are less frequently discussed on this forum. In brief: The concept is simply about using two driven wheels to provide electric assist towards forward motion, whatever the technology.

What is required to create 2WD?
Two motors affixed to two sturdy wheels securely, two controllers – each attached to one motor, and at least one throttle. It is rare to find a single 2WD controller specified for "bike" use, therefore creating a 2WD bike is nearly always a custom-job requiring basic wiring & soldering skills. Each motor requires its own torque arm to secure the axle from spinning.

Do I need two throttles?
No; modifications can be made so as to drive both controllers with one throttle. However care must be given on how the mod is applied. Not all controllers are equal, and neither are throttles – therefore a bit of experimentation will be required to find the correct match for the application. Some folks enjoy having two independent throttles which works well for Front & Rear configurations.

Do I need two CAs?
No – although given with trepidation & reluctance: Having one attached to each controller will help to monitor activity and diagnose issues, especially if one unit malfunctions.

Will I be able to use Regen braking?
Yes. Regen can be applied to one, or with both controllers using careful modification. Note: 2WD-regen can be jarring and hard on the frame. Adjust the strength according to need.

What other features can be used with 2WD?
Any feature may be used. If both controllers are paired equally, then it is possible to use one dashboard device to affect both wheels. Typically, Cruise-Control and 3-Way Current Control may be unified with good results.

Do both motors require synchronization?
No, especially motors aligned Front to Rear. The motor attached to the steering wheel will turn at a different rate and therefore it is preferred to have both motors run independent of each other, hence employing two controllers.

Is 2WD more efficient than one-wheel drive?
Yes and No. A 2WD bike outfitted identically with a single-wheel drive bike (sans extra motor/controller) will outperform the single-wheel in acceleration off-the-line, in hill climbing, when facing cross or headwind, or otherwise pulling a heavy load. In these conditions, 2WD is more efficient at applying power to the ground by splitting the load between two controllers and two motors; neither will get as hot as a single controller/motor at the same power rating. 2WD bikes therefore run cooler and better under load farther and longer than a single-motor. In essence, a 2WD bike will have approximately twice the available torque and traction in all conditions.

However, 2WD bikes are not more energy-efficient. On flat level ground without any headwind (discounting wind resistance) a 2WD bike will consume more power in cruise mode because of the energy consumed to manage the second system, including overcoming resistance to freewheel. Likewise coasting downhill will cost more due to drag created by the extra driven wheel. This loss can be partly offset by employing dual-regen braking. Two motors will use up to twice the energy twice as fast over that of a single motor! Be prepared to double the battery capacity of your pack, or cut the estimated distance by at least 1/3 when using 2WD. The easiest way to visualize the difference between 2WD and single-wheel drive is to imagine the difference between a 2WD truck and the same model & engine as a 4WD: Which will have more power (should be the same), which will climb hills better (4x4), and which will get better gas mileage (2x4). There’s no free lunch and it’s the same physics with electric 2WD bikes. :)

Will 2WD experience contention?
Contention means two forces competing for control, and the answer is Yes, they will. Contention is most noticeable on flat level ground. There are however several ways to reduce contention to nearly imperceptible levels.
  • Program the Front & Rear controllers to output slightly different power levels.
  • Use different tire treads or types or sizes.
  • Use different rim sizes.
When using Cruise Control I often found that either the Front or the Rear motor would pull much harder than the opposite, and this strange artifact appeared to be completely arbitrary, though on the plus side – there was no contention.

In a Front-to-Rear alignment, which motor should have more power?
Traditional physics suggests dividing up to 70% of the total system power to the Rear motor. If programed evenly the Front motor will occasionally slip when accelerating off-the-line, especially on loose or slick surfaces. This can actually become a liability when cornering due to the loss of traction. Conversely, Regen should be set highest on the Front motor for the same reason why motorcycles have dual disc brakes forward and not aft.

On a 2WD bike, where is the best place to put the batteries?
In the Triangle; the center region of the bike. I have found that saddle bags over the triangle increases the weight distribution forward of center and helps prevent the Front wheel from spinning in rainy weather.

What is the best bike frame for 2WD?
Depends on the use. If money can afford, a full-suspension Downhill frame with rear horizontal dropouts is quite possibly the best.

How many torque arms do I need for 2WD?
At least one per wheel, although one on each side of each wheel is strongly preferred. It really depends upon how much power is being applied. Best to err on the side of caution and use two per driven wheel.

How fast can a 2WD bike go?
How fast do you want to go? Imagine that you have a single wheel drive and you push it as hard as it can go when the batteries are freshly topped off; write down that speed. Then run the freewheel test. Note the gap between the actual speed and the freewheel speed. With 2WD it is feasible to cut that gap in half using the same voltage ~ and possibly more with current adjustments. The motor winding has a large effect on how well the top-end is improved with lower wind motors benefitting the most.

What happens if one motor fails?
The beauty of having 2WD means you are not stranded! 2WD = Redundancy; two separate & independent systems. In most cases a single-wheel drive can get you home safely.

What are the most common problems with 2WD?
  • One motor periodically cuts out under load:
    • Check to ensure that the Battery Voltage does not sag below LVC when asking for WOT. It could be that the Battery cabling is too small, C-rating insufficient, Capacity (Ah) is too low, or pack is nearly depleted.
    • Check to ensure the Throttle is provided enough voltage to drive both controllers. Good cabling should solve this.
  • Motor makes horrible grinding noise and appears to lock-up:
    • Phase and/or Signal wires are damaged. R&R.
  • CA gets real flakey, powers-up, dies, resets, reports battery voltage falsely against the opposing well-functioning CA; may appear in addition with no throttle:
    • Check the power connection to the controller; it is likely dirty or damaged. If it can be repaired, the CA should appear normal and in parity with its opposite. If the throttle is still not functional, cycle the power OFF, then ON again to reset the controller.

Is 2WD more fun and worth the trouble than a single-wheel?
You have to ask? Yes, many believe so; it’s a kick in the butt! 8)

Related threads sorted newest to oldest:
This is by no means the complete list of topics on 2WD, though easy to find with a quick search.

Opinions expressed are my own based upon direct experience. I thank the Community for providing knowledge which helped advance my own.
Go for it! KF
 
Now for some correct info on 2wd bikes. :)

AFAIK, nobody have been able to get a single throttle bike to share current effectively, it always glitches to one or the other handling a larger share of the load. (Has anyone solved this without doing the double-hall throttle mod or running 2 throttles?)

If your COG (center of gravity) is above the surface of the ground (as it is for our bikes obviously), then the more you accelerate, the more weight transfers to the rear wheel, right up to 100% of weight on the rear wheel, the point at which a wheelie starts. If your bike can wheelie, then you have NOTHING to gain from powering the front wheel, and lots to lose (like having a front motor, which cripples a bike's ability to do anything fun off road IMHO).

Lugging the parasitic losses of 2 motors will drastically reduce light throttle efficiency, and moderately reduce medium throttle and high throttle efficiency, while adding nothing to the bike that a single rear motor couldn't do alone if it wasn't all ready at it's thermal limit.

Which brings up the single reason 2WD can ever have a purpose. If you can't fit a properly sized rear motor for your continuous thermal needs, you could add a front hub, pay a stiff weight and handling penalty, add a bunch of complexity and double your potential failure points and increase system cost, and add a front motor. :)

If a rear motor can meet your thermal needs, then it's always going to be the best option. Single exception being something like snow or extremely slippery mud to the extent that you can't lift the front tire of your bike. All wet pavement, and 99% of dirt riding would only be traction penalized by having the front wheel powered rather than simply running an appropriate rear tire for the application.
 
As always Luke hits the nail firmly on the head...well said mate :wink:

KiM
 
Kingfish, good stuff. I believe you should qualify the statements about efficiency, because I believe that's only true for systems running at lower power with light loads, so they are already running near peak efficiency. In that case there would be virtually no gain on flat ground, but the only way there's a loss is if the 2nd motor pushes them past peak efficiency, which would be pretty difficult. Play around with Justin's simulator using a 74V pack and a 40A controller, and you'll quickly see that sharing the load with a 2nd motor will pay huge dividends even on the flats. That's because you can run at significantly higher speed with several points higher overall system efficiency.

Guys are looking for 2wd do it for more power with less stress on the system, and whether it's climbing hills or during takeoff, it's the same thing, and 2 is far better than one.

While I don't doubt that much more battery may be required for 2wd, it's not because of lower overall efficiency. Higher consumption is the price of greater performance up hills and during acceleration. There is no fixed component of our motors' operation. It's all variable, yet your wording almost sounds like there's some fixed power consumption to overcome just to run, which isn't the case. Motor cogging is something different and that's only when coasting, not under power. Also, much of the no-load power required to run a hubmotor wheel off the ground is related to wind resistance of the spokes and tire, which are incurred whether the wheel has a motor in it or not. That means even changing the load in the simulator will understate the real benefits. It would actually be fairly difficult for a given level of performance to have lower energy consumption with one hubmotor than with 2wd, and the difference would be quite small. OTOH with a high power system it would be fairly easy to pick up 5-10kph using the same or less total power even on flat roads. For take offs and hill climbs it wouldn't be even close, and while coasting down hills the difference may be significantly in favor of 1 motor, the power consumption is insignificant and not really worth considering (a big portion of a little ain't much).

It's excellent that you pointed out that it's undesirable to run a closely matched front and rear 2wd. Do you think running just quite different current limits is enough, eg 50A front and 100A rear, or does having virtually identical wheels cause problems as controllers go in and out of sync with the pulses they draw from the battery?

Luke's points are out at the other end of the spectrum from Kingfish, but I'm somewhere in the middle. I want both wheels firmly on the ground, so I have no problem at all trading a 7-10kg weight gain for 3-5kw more power and 50% more torque, as well as better performance through the curves. All hubmotors bump into their thermal limits if you have sufficient controller, and I never discuss off-roading, since I don't do it yet. Regarding the parasitic losses, I see them as immaterial. Turning 2 extra bearings and powering the extra controller board are almost nothing AFAIC.

John
 
Thanks for the info! definitely thinking of 2wd over here..
 
As far as I have been able to tell, everyone with a 2WD ebike has doubled their e-grin. None have indicated it wasn't a lot more acceleration, power and fun. That doesn't mean it is optimal, but it is a useful option in some cases. Also there are reports that dual motored setups have slightly higher efficiency overall because both motors are running at a higher efficiency point on their operating curve. This is not a large difference, but it underscores that the battery capacity does not need to double, just the peak current capability and then only if the double torque is actually going to be utilized.

Anytime you are traction limited, thermal limited or torque limited then adding a second motor may be useful. Of course if you are racing on pavement with lots of traction available and your motor never overheats, then one motor rear wheel drive is superior. But if you are climbing Pike's peak and overheating your rear hubmotor a second motor to share the heat dissipation and drop the torque load per motor in half could be the winning ticket. Anytime one motor is bogging down and overheating, but would not overheat producing half the torque, a second motor would help.

But Luke is right on - it is more things to break, and carry the weight of, mount on the bike, pay for, and maintain.

It does provide a second motor to limp home on (instead of pedaling) after you cook the rear motor. Providing the battery didn't flame out and if the rear motor still turns freely.

One technique for power control that has been discussed but apparently not tested yet - using the Cycle Analyst current control mode to distribute the power between the two motors. In this setup a CA is programmed to current control mode (see the operating manual). Two CAs are used, one for each motor. One throttle supplies a signal to both CAs, essentially providing a 0-100% throttle. Each CA converts this to current control, 0-N amps as is set into the CA parameters. Since both motor controllers are fed from one battery they see the same voltage. So the current control becomes power control. Thus you can set the rear motor for 60% of total power, and the front motor for 40%, or whatever you want. You can also set the maximum speed on each CA separately, so you could set the front motor to drop out at some speed if you want the extra traction at low speed, but you don't want the front motor handling effects at high speed.

If the front motor is a geared motor with a freewheel it will have essentially no drag when not pulling, and lots of torque at low speed to help get the bike accelerated out of the low speed inefficient zone to keep from overheating the rear motor, which might be a higher power direct drive type. This is great for starting and climbing those really steep hills especially in traction limited dirt (such as our Marin ride up the hill off-pavement).

Update 12/2016

Having done this now (see the Bonanza 2WD link in my sig) using a rear DD motor and a front gearmotor, I can say - Wow. It really makes a difference to double the torque. With the rear 9C DD alone steep dirt gradients were a challenge that nearly melted my motor one day. It was running close to the traction limit so pedaling tended to cause slippage and loss of control. Adding a torquey front gearmotor changed everything. I can run up those same gradients, on dirt, at excellent speed with almost no heating.

I used a pair of PhaseRunner Sinewave FOC Controllers with Torque Throttle. This automagically balances the motor speeds and makes the control smooth and easy.

However changing a flat in either tire is no fun. Before it was only the rear tire that was hard to fix. Perhaps this is not really much of a change since flats are almost always in the rear tire...
 
I forgot the biggest benefit I'll get, which is not having to worry about blowing controllers anymore. The front hub will relieve enough load from the rear motor that I won't have to sweat it every time traffic slows me down going up a long hill without turning the rear controller down to unacceptably low levels, and the front hub won't be running at stressful levels for it's controller even without the rear carrying most of the load.

I like Alan's point too. Performance for that small weight gain in the front wheel will be like I dropped over 100 pounds over night. Who knows, maybe I'll even be able to safely turn my settings up enough that number is more like 150lbs or more, and a sub 100lb person on my current bikes would be :shock:
 
This has been a good discussion. About six years ago before getting into ES I got a WE BD36, (Wilderness Energy 36V Brushed Direct drive hub motor), with a 35amp controller. Even with hard pedaling I had to stop and cool off both the motor and myself on the 1/4 mile 15% grade leading to my house. My solution was to get another kit and build a two wheel trailer using both hub motors. Thats when my ebike grin started growing. Then, just for kicks I got a third kit and put a motor back on the front of that old $50 Wal-Mart MB. If I remember correctly it topped out over 25 MPH and only took about 4 seconds to hit 20 MPH with no pedaling. Since then I have probably put together a half dozen two and three hub motor rigs on a delta trike, recumbent bike, and tadpole trike. I always felt the multiple motors exceeded expectations in giving higher top speeds and much better climbing ability.

Alas, all the extra weight and trailers was a pain, particularly when hauling them out of town to ride with others. Also, I preferred to have a "legal" ebike. So now with the mid-drive I am content to go up the 15% grade at half the speed, but with no strain on anything. For over a year I have just used a 48V 10AH Ping battery since most of the time I pull less than 10 amps.
 
liveforphysics said:
AFAIK, nobody have been able to get a single throttle bike to share current effectively, it always glitches to one or the other handling a larger share of the load. (Has anyone solved this without doing the double-hall throttle mod or running 2 throttles?)
<I raise my hand> Me :)

liveforphysics said:
If your COG (center of gravity) is above the surface of the ground (as it is for our bikes obviously), then the more you accelerate, the more weight transfers to the rear wheel, right up to 100% of weight on the rear wheel, the point at which a wheelie starts. If your bike can wheelie, then you have NOTHING to gain from powering the front wheel, and lots to lose (like having a front motor, which cripples a bike's ability to do anything fun off road IMHO).
Agreed - but then I do not do wheelies. This thread was spawned to support those that WANT to build a 2WD and have questions.

liveforphysics said:
Lugging the parasitic losses of 2 motors will drastically reduce light throttle efficiency, and moderately reduce medium throttle and high throttle efficiency, while adding nothing to the bike that a single rear motor couldn't do alone if it wasn't all ready at it's thermal limit.
Rubbish: Throttle effectiveness is all on how it is programmed. For light, medium, and heavy power I used the 3-Way Current Control programmed as 52%, 85%, and 120%. The first setting is so fine and delicate - yet limited the top speed to 20 mph. On the Road, I perferred the medium setting which allowed up to 34 mph. Once I used the heavy setting - and that was to qualify for the 40 mph Club :p The only part of your statement that specifically applied to 2WD is that the parasitic losses are greater with 2 motors, but then I already explained that. :lol:

liveforphysics said:
Which brings up the single reason 2WD can ever have a purpose. If you can't fit a properly sized rear motor for your continuous thermal needs, you could add a front hub, pay a stiff weight and handling penalty, add a bunch of complexity and double your potential failure points and increase system cost, and add a front motor. :)
Well, obviously 4x4 and AWD have their utility or people wouldn't buy them. I own a 4x4 and swear by it ~ except for the fuel economy. 2WD appeals to me for the same reasons (and not for the fuel ecomony); it's the ability of having that instant power and traction available at a moment. A single-wheel drive will not be able to do the same in all conditions. Two wheels driving will always be safer than one. Sure it's more complex; that's what your paying for. It's a fun challenge, though I would agree that it's not right for everybody, and I'm not selling the idea that way. It costs money and it costs time, and it takes a resourceful person to pull it off.

liveforphysics said:
If a rear motor can meet your thermal needs, then it's always going to be the best option. Single exception being something like snow or extremely slippery mud to the extent that you can't lift the front tire of your bike. All wet pavement, and 99% of dirt riding would only be traction penalized by having the front wheel powered rather than simply running an appropriate rear tire for the application.
I run cross-metro and cross-country with 2WD; works great in all conditions of rain or... well rain if we're talking Seattle. In sunny California, worked as well as could be expected. I can't wait for it to snow! :D

Speaking of which: My Schwalbe Ice Spiker tyre 60-559 (26 x 2.35) arrived today from Germany! That is going on the Front wheel. The rear will have the Panaracer Fire FR 24" x 2.4 :twisted:

~KF
 
All you guys are failing badly at permanent magnet motor basics. lol Some of what you've said is true if you were using induction motors, but we're not.

I just spent 10mins with Biff (professional motor engineer) to make sure that I've got this all peer reviewed with all technical content perfect.

To make a PM motor turn, even if the wires aren't hooked up to ANYTHING, you have a substantial energy penalty. >90% of your no-load power is just going to make the motor itself turn (core loss), which happens anytime the motor is spinning, no matter if it's powered or not. The other <10% of no-load power is copper loss, which doesn't happen if you're not powered, but core loss is the same powered or not. Core loss is mostly the hysteresis losses in the lams as the permanent magnets field swap there field lines back and forth about 50times with each revolution, the eddy currents of the rotor magnets and lams, and associated recirculating currents in any parallel paths (even just in the individual bundle for a phase, even when connected in WYE)from BEMF. The last hubmotor we did had something like 2.5amps of no-load at 96v (240w, which may be a more extreme example for folks who normally ride at low speeds and may only see half of that), at no-load currents, >90% of that is not copper loss, but the rather the above parasitics.

So, if you had a 1kw-hr pack, and you added a front hub, not even connected to anything at all, you're going to be giving up ~210-220w of your drive energy just to make that front wheel turn (because to drive a 210watt mechanical load, you're going to be using at least 250w going into your hub). Which is going to mean over the course of an hour, you gave up about 1/4 of your pack just to have a front motor in your wheel turning (and it has nothing to do with bearing drag or whatever).

Now, when you go to power it, those losses do NOT disappear at all, they just slightly increase as field intensity picks up and they develop eddys on magnet surfaces and a few other things they didn't have before.

Then you also introduce copper loss, which brings up an excellent thing to understand for folks. If you power two motors lightly, vs a single motor at higher current, it's extremely unlikely for the two motors to more efficiently drive the bike, right up until the single motor begins to flux saturate the teeth in the stator, or until the copper enters thermal runaway and the resistance goes up a large amount (neither of these things should happen for normal applications).

The point at which your break even between a single vs dual motor setup is when 2x the core loss is equal to 1/2 the copper losses of a single motor. At this point, it is exactly equal to run a single motor vs a pair. Above this power, it's better to have 2 motors, below this power it's worse to have 2 motors. For all the time you're commuting around, it's simply going to be at a penalty.
Unless you guys are just shreding through motors in 30 seconds like I do, it's extremely likely you will never experience a point where having a pair of motors would operate more efficiently for any period of time.


If you can do it with a single motor, it's ALWAYS going to be better, and better in a huge way for the intangibles if it means not having the ability to hop your front end up robbed from you, along with nearly doubling system complexity.
 
Not more efficient in all cases. Possibly in some. Let's see if we can find one.

I think the configuration of a geared front hub with a freewheel and a more powerful rear direct drive hub is the best. Then when the front hub is not needed it has no significant drag. The value of the front hub is primarily in climbing steep stuff that overheats the rear motor or exhausts the available torque.

Let me do a little modelling and see if I can find a case where efficiency is higher.

Take a look at ebikes.ca's model.

Starting with the default values and then changing a few.

Ok, on level ground it is difficult to find a more efficient configuration. Probably not going to happen. Luke's right. Now try a good hill. Yes. Here we go.

Clyte HT3525
Ping 16S 20AH
16% grade
one motor
speed 4.9mph
203 wh/mi
overheat in 8.6 minutes

if we had 2 of those motors it would be similar to running an 8% grade.
speed 15.3 mph
65.1 wh/mi
no overheat

So maybe I've got this all wrong, but it looks like lifting the motor out of the inefficient region saves more power than the losses of a second motor, but only if we are bogging way down.

If we set the front geared hubmotor to help up until 16 mph or so it will lift us out of the inefficient region but not use power above that speed. The freewheel won't consume significant power. We'll get strong takeoffs and great hill climbing.

Thanks, Luke. Good points to clarify thinking and drive an example. This is EXACTLY where I was when climbing the hill in Marin, since I didn't have an AGNI and my motor was nearly stalled. It got so hot it nearly melted, and at least one other 9C did melt.
 
Luke,

While I agree with your points individually, I assure you that you're looking at it from a point of view that isn't applicable.

You're leaving out that all of the losses are variable not fixed. Otherwise running at higher power or having bigger motors would result in tremendous efficiency gains, but they don't. There's simply not a fixed loss to overcome, and it's refuted simply by the fact that overall efficiency of our motors increases with rpm up to peak efficiency. What we're able to take advantage of in terms of efficiency is the increase in efficiency above peak power. The single motor has nothing more to give, but adding a second motor easily results in both more power and greater overall efficiency.

Take a typical hubmotor and put it under enough load that it's running at peak power. It will be running at an overall efficiency in the mid 70's. Now if you take that same motor at the same voltage and cut the load in half it will run at a higher rpm and closer to peak efficiency. 2 motors each with half the power output (half load) will run at much closer to peak efficiency than the single motor running at peak power. Guys running these 26" wheels with bicycle hubmotors at higher voltage are running their bikes at much closer to peak power than peak efficiency.

Here's a specific example using the Ebikes.ca simulator. Assumptions input, which took some trial and error to get to 50% power output for the dual motor case. I tried with larger wheel, but even slight grades put it below peak power, and the negative grade results to cut load in half simply don't seem correct, because it showed power required to go down 8-10% grades. I had to take loading the motor down with a hill to get to something I could also simulate half power without changing the throttle lower too much:
2807 9C hubmotor
80V .05ohm battery
40A controller
20" wheel
100kg gross MTB

For the single motor I used a 13% grade, which loaded the motor to operate just above peak power for that voltage. To cut required power in half to simulate one of dual motors, I made the grade 0% and still had to reduce throttle to 92%.

Results:
Single motor 2309W motor output, 76.7% efficiency so 3010W power input, speed 44kph.
half load 1163W motor output, 84.4% efficiency so for 2 motors 2756W power input for 2326W dual motor total output, and a speed of 50.6kph. 2 motors for the same combined power output as one results in higher overall efficiency.

Here screen shots of the simulations (sorry, I couldn't make shrink to display work):
2807 80V 40A at full load.JPG

 
Alan B said:
Not more efficient in all cases.

Not in all, but in most sure, especially since we always have to accelerate to speed. On the flats there would be little to no benefit for those running at lower voltages. Guys running higher voltages, especially with large wheels, are running closer to max power than max efficiency on the flats. Just simulate a 700C wheel at 80V and 40A to see how adding a second motor will put the load toward max efficiency and therefore greater overall efficiency for the same power out.
 
Clearly Luke is correct about the various loss sources involved with having a second motor, but this begs the question - how much is the actual loss in different driving conditions. Above we looked at the bogged down overloaded motor case and found some benefit from two motors. Here we look at the level ground full throttle case.

What is the inefficiency of a dual motor setup at normal speed on level ground?

How can we model this with ebikes.ca's recently upgraded simulator (thanks Justin)?

Starting with the default setup (9C 36V 20A level ground 26" mtn bike 100% throttle 100kg weight) we get:

equilibrium speed 21.3 mph
load 392W
efficiency 22.9 watt hours per mile
battery power 488 watts
efficiency 81.1%

So we need a condition where we are going 21.3 mph at half the load power, or 392/2 = 196 watts

We can do this by adjusting the grade in the model, making it slightly downhill for the load power of 196 watts at 21.3 mph.

The model only allows grade adjustment at 0.5%, the closest is a -2% grade where the load is 207W, just slighly higher than our half load goal.

Now we adjust the throttle to find the point at which 21.3 mph is the equilibrium speed.

We find 90% is about as close as we can get, producing 21.2 mph, 203 watt load and 205 watt motor power. The Data for one motor is then:

equilibrium speed 21.2 mph
load 203W
efficiency 11.9 watt hours per mile
battery power 252 watts
efficiency 81.5%

We need to double this to compare with the one motor data above, so 11.9 wh/m doubled becomes 23.8 versus 22.9 watt hours per mile for the single motor. This is a 4% penalty but note that we are producing 205 watts times two or 410 watts instead of 392 watts or about 5% more output power, so this might indicate a 1% gain in efficiency.

Note that motor efficiency has increased from 81.1% to 81.5% so this is not enough gain to explain an increase in efficiency of 1%. Note also that controller and battery impedance losses have been reduced, and while the controller gain is correct the battery and wiring losses are only half of what they would be with two motors.

The accuracy of the model and this method of interpreting the data cannot be expected to produce meaningful predictions at the 0.5% level, but it does show that the loss of having a second motor may indeed be very small if the ebikes.ca model is accurate with respect to motor losses. We could easily see a 200 watt power loss and we do not see that. Perhaps Kingfish should report on his 2500 mile experiences with power consumption as I think he would easily be able to see a 200 watt loss on such a long trip.

In summary, if the model is correct about motor loss the loss from a second motor may be quite small compared to normal losses.
 
The only way you could ever simulate an efficiency loss adding the second motor would be if the load for 1 motor is so light that the single motor is running very close to peak efficiency and the lowering the load pushes them past peak efficiency and to the downside of efficiency toward no-load rpm. I submit that this is next to impossible to achieve.

For all intents and purposes, 2 motors will always be more efficient for a given load. With a low powered bike the gain would be minimal and not worth it for same speed at cruise. Even with modestly powered rigs, however, stop and go riding would pay significant dividends in terms of efficiency if you control yourself and accelerate the same.

In my case, my cruising speeds will remain unchanged, because I don't ride at top speed or even close with one motor. I will however, use the extra power to accelerate much harder, but the reduced load each motor sees during acceleration means each will be more efficient, and the greater power will be for a shorter duration. That means the increased performance during acceleration will be free in terms of overall consumption, other than the slight increase in average speed since I'll reach cruise faster, and the slight increase in energy required to accelerate the extra mass of the second motor. Other than the effects of wind resistance, the energy required to accelerate to a given speed requires a fixed amount of energy, so ignoring the wind, 2 motors enables you to launch at greater efficiency as long as you stay below double that of one motor.
 
John and I overlapped in our analyses postings. Quite Interesting. In any case the power cost of the second motor is not a major loss increase, according to this simulator. In some cases it appears to be a slight gain. That was reported on the Magic Pie forum as well where a number of folks have deployed multiple Pies - no significant loss, possibly a slight gain.

Anyway this is Kingfish's FAQ on running two motors, so we don't want to fill it with disagreements, but with useful information and experience. Perhaps those who have done dual motors it can report on their watt hours per mile consumption. Perhaps we can compare Kingfish's results with two motors this summer to his previous trip, and to Justin's cross Canada trip with one motor a few years ago.

I plan to do two motors. Maybe soon. Thanks for the thread, Kingfish!
 
Luke has some great, well researched points.

That being said...
I ran my MAC motor on 4kW the other day and enjoyed the feeling of being shot out of a cannon with all that damn torque.
Then my gears and clutch died 4 days later. HM.... coincidence? :lol:

The only way forward would then be to run dual motors at 2.5kW each = 5kW total.

I'll take a hit in efficiency, yes.. but do i have the fabrication skills, space, tools to do a non-hub build? hell no.
For some of us, this is the only option.
 
Luke is looking at it wrong, there is no efficiency hit. Lower load means greater efficiency all the way up to peak efficiency.
 
John in CR said:
The only way you could ever simulate an efficiency loss adding the second motor would be if the load for 1 motor is so light that the single motor is running very close to peak efficiency and the lowering the load pushes them past peak efficiency and to the downside of efficiency toward no-load rpm. I submit that this is next to impossible to achieve.

umm, you mean like when the front wheel is slipping or completely up in the air... :wink:


liveforphysics said:
To make a PM motor turn, even if the wires aren't hooked up to ANYTHING, you have a substantial energy penalty. >90% of your no-load power is just going to make the motor itself turn (core loss), which happens anytime the motor is spinning, no matter if it's powered or not.

3 ways around that i know of fer sur, possibly more.
1. don't spin the motor by utilizing a freewheel as has been discussed.
2. a mechanism to physically pull the core out of the PM field.
AF is the simplest to implement & commercial versions have been around for at least a decade.
3. flux cancellation.
something like a 'back-iron', except a caliper mounted front-iron to short out the magnetic field.
however a more elegant solution would utilize 5-phase windings to generate an anti-phase flux waveform.
we'll soon know how well that works out, hopefully by the beginning of the new year.
 
Toorbough ULL-Zeveigh said:
John in CR said:
The only way you could ever simulate an efficiency loss adding the second motor would be if the load for 1 motor is so light that the single motor is running very close to peak efficiency and the lowering the load pushes them past peak efficiency and to the downside of efficiency toward no-load rpm. I submit that this is next to impossible to achieve.

umm, you mean like when the front wheel is slipping or completely up in the air... :wink:

I definitely concede Luke's point about the front motor wasted when there's so little traction. I was getting at comparing 1 motor to 2, and if the single motor rig is optimized and running at or near peak efficiency, then reducing the load via a 2nd motor could conceivably push operation to the downside of the efficiency curve.

Toorbough ULL-Zeveigh said:
liveforphysics said:
To make a PM motor turn, even if the wires aren't hooked up to ANYTHING, you have a substantial energy penalty. >90% of your no-load power is just going to make the motor itself turn (core loss), which happens anytime the motor is spinning, no matter if it's powered or not.

3 ways around that i know of fer sur, possibly more.
1. don't spin the motor by utilizing a freewheel as has been discussed.
2. a mechanism to physically pull the core out of the PM field.
AF is the simplest to implement & commercial versions have been around for at least a decade.
3. flux cancellation.
something like a 'back-iron', except a caliper mounted front-iron to short out the magnetic field.
however a more elegant solution would utilize 5-phase windings to generate an anti-phase flux waveform.
we'll soon know how well that works out, hopefully by the beginning of the new year.

I really didn't understand this one. I used throttle activated regen, ie I let off the throttle and regen comes on, so I never coast, and I ride the throttle long enough that I always have some mechanical braking at the end. I never coasted to any significant extent before I had regen, and I never coasted when I rode a motorcycle. AFAIC cogging resistance is just low level braking, not a loss. There is no cogging during operation (except maybe the 1 unpowered phase), because the coils are active, and if I'm wrong about that it's part of each motor's efficiency curve, and doesn't change the fact that efficiency increases with decreasing load all the way up to peak efficiency. That means 2 motors has to be more efficient at all power levels less than double up to peak efficiency, which is generally pretty low power.
 
John in CR said:
Luke is looking at it wrong, there is no efficiency hit. Lower load means greater efficiency all the way up to peak efficiency.


I'm afraid that's not the case my friend.

Efficiency for any electric motor is 0% at both ends of the scale. It's an arched curve with 0% efficiency on both sides of the curve, and the peak efficiency always occurring when copper loss matches core loss, somewhere in the middle.

When you unload a motor to the point where core loss is more than copper loss, you're on the decreasing efficiency side of reducing load. This is pretty much the 95% operating mode for folks commuting with a large hubmotor (or a large type of any PM motor).

In return, while climbing really steep things, and/or when accelerating very hard, the larger motor hopefully has a bunch more copper in it, so they can minimize copper loss under those high load situations.

By all means, if you were going to be operating mostly in very high load situations (playing the waiting game to see if the hill will end before your motor smokes, or trying to maintain extreme high speeds or something), then yes! A larger motor will be more efficient for you, and if you can't find a large enough single motor to do what you need, then you could add a second, and be more efficient during those times of very high load.

I know everybody has a different area they ride, speed they ride, hills they ride, bike weight, rider weight, etc. But I envision 95% of folks spend 95% of there time riding in a situation where the single motor is going to be most efficient.
 
John in CR said:
AFAIC cogging resistance is just low level braking, not a loss. There is no cogging during operation (except maybe the 1 unpowered phase), because the coils are active, and if I'm wrong about that it's part of each motor's efficiency curve, and doesn't change the fact that efficiency increases with decreasing load all the way up to peak efficiency. That means 2 motors has to be more efficient at all power levels less than double up to peak efficiency, which is generally pretty low power.


I used to believe this same thing my friend, but after putting a pair of motors on the dyno and driving one of them unloaded (even with the phase wires not connected to anything at all), it totally changed how I see things. Your cogging torque ripple, and ripple ONLY sums to zero, but you have a substantial net loss as the rotor turns in the motor. This core loss is just a reality of having your magnets turning inside the motor. Your core loss is >90% of your no-load power (which I think was a few hundred watts for your motors if I'm remembering right?) This is just a penalty you have to eat for making the magnets change the field of the iron around as they rotate around inside there. Nothing can avoid it without the suggestions that Toorbough had listed (pulling the rotor out when you don't need that motor, making the rotor freewheel, etc).
 
If anyone wants to see the losses/inefficiency of a second (electrically disconnected) motor, it's easy enough even without putting it on a single bike and riding it.

Take any two ebikes, and flip them upside down.
Put them lined up so that the tire of one motor will be able to friction-drive the other, but not yet touching.
Disconnect the phase and hall wires from the driven motor.
Hook up your wattmeter on the driving bike.
Run the motor at your desired speeds, noting down the (no load) power usages.
Push the bikes together so the disconnected motor will be driven from the other.
Run the motor at your desired speeds, noting down the (now slightly loaded) power usages.

That's approximately how much power will be "wasted" just driving the second motor on the bike, even without it being connected to anything, and even without a power load on the first motor.

If you like, redo the test with the second motor still connected to it's controller and battery, and powered on, but at zero throttle. Power usage on the first motor will probably increase noticeably (maybe not significantly).

Redo it again, this time with the second motor also being monitored for power usage, and set it's throttle to match the first one's speed just enough so that there is no "drag" on the first motor by the second one. Total up the power usage between the two motor systems, and it will still probably come out a net loss. (this test may not work as expected because there is no real load on the first motor).

Now, for the sheer fun of having the second motor, this doesn't matter. :) Nor does it matter if part of your reason is to have a backup motor in case of system failure on the first.

But if you're trying to make it more efficient, I'd venture that in most cases, for our typical purposes, it probably wont' be.
 
Has anyone mentioned the added weight factor and the extra
Power that is needed because of it? Surely The weight of the
Extra frock and controller would would result in more
power required to move the bike along correct?

KiM
 
Back
Top