Published on 08/15/2007

This is the first in a series of technical articles covering some of electric bikes complexities in quite full detail so that all aspects of the subject can be appreciated. You are then free to participate in our forum to further discuss the topic.

Which is better, a hub motor or drive through the gears?

This is the first in a series of technical articles from our resident expert Flecc, covering some of e-bikes complexities in quite full detail so that all aspects of the subject can be appreciated.

We frequently see posts proclaiming the superiority of motor drive through gear systems on e-bikes, slighting wheel hub motors as inefficient, but in every case to date, the poster has fallen into the trap of using conventional motor vehicle gear theory. This does not apply to e-bikes.

They fall into that trap because they fail to realise that virtually all e-bikes drive through the gears to a considerable extent, because they are true hybrid vehicles with two power sources, one of which always drives through any available gears. Motor vehicles either drive through gears or they don't, that's an absolute, but e-bikes are infinitely variable in what they do.

Riding on the flat or on very slight downhills at above a hub motor cut off speed, a commonplace situation, all drive is through the gears since it's from the rider. In moderate conditions as when riding at the bike's normal cruising speed, the part of the power supplied by the rider is still through the gears, and the part supplied direct drive by a hub motor is at maximum efficiency anyway, so there can be no possible gain in it driving through the gears then. In fact there's a loss of efficiency due to the gears if they are used for the motor as well, and this loss can be substantial, making motor drive through the gears less efficient most of the time on an e-bike.

Only when conditions force the speed down and the hub motor drops well below it's peak torque point on the power curve does the inefficiency assume higher proportions that might give justification for using gear drive for the motor.

I believe that justification is wholly inadequate (that's my opinion) for the following reasons (these are facts):

1) Wheel hub Hall effect motors have advanced and can do the whole job well (though not all do) and quite efficiently. Due to their simplicity, they are also more reliable than complex drive through gear systems, often lasting for many years and frequently outliving the bikes they came in, this clear to see in second hand sales of these ex-bike motor units.

2) Where a small efficiency difference exists, it's offset by all the complexity costs, in design, manufacturing, reliability etc of drive through gear systems.

3) The European and UK legal restrictions mean the usage band is from 5 mph to the legal power phase down point of 13 mph, only 8 mph in bandwidth, so the Hall effect motor torque curve with a peak at 8 mph covers that speed band quite well.

3) Cycle gears are not strong enough to take both rider power and the higher motor powers customers now want, being designed only for rider power. The Panasonic system is extremely low powered by todays standards which enables gears to survive reasonably well, despite which there have been quite a few failures of the hub gears on these bikes. The latest Panasonic unit has had it's power reduced further, possibly for this reason.

4) Gears have inefficiency, a hub gear typically being about 84% efficient, so there's a 16% loss of the power passing through one.

5) Multiple gears are not needed to cover an e-bike's effective 8 mph speed range, and result in far too much idle time when drive is not applied. The road tested figures for the Lafree Twist bikes prove this. Under a given set of conditions with the standard 6.5 Ah battery, the Twist with the 3 speed Nexus hub gear had a range of 20 miles, with the 5 speed SRAM hub 17 miles, even though the SRAM is more efficient than the Nexus, and I've seen even lower figures for the 7 speed ST version. Clearly an excess of gears is harming the efficiency due to drive down time and loss of momentum.

6) It is irrational to pass the motor drive through the rider gears, since it means that at rider gear changes, motor drive and momentum is lost as well. Then that momentum must be made up again, that's very inefficient. A hub motor maintains the drive through the period of a gear change, so maintains the bikes momentum. This is especially important when hill climbing, this being when forward momentum lost during a gear change is at it's greatest. There's also the fact that cyclists in general are notoriously bad at selecting optimum gears for themselves, so the possibility that they might choose the optimum compromise gear for themselves and the motor is extremely unlikely.

There is a probability that two gears for the motor only and separate from the rider's gears, with maximum power/torque at a choice of either 6 mph or 10 mph would provide better for steep hill climbing, but that is qualified by the following:

a) Less power is needed to climb hills at 6 mph than at 8 to 9 mph on the maximum torque/power point.

b) A powerful hub motor/average rider combination rarely drops to 6 mph when hill climbing anyway, so the theoretical inefficiency off the peak of the torque/power curve hardly ever exists in practice. It's largely illusory and this shows how relying only on theory instead of tempering that with practical knowledge of what actually happens is always mistaken.

c) These are hybrid vehicles, human and electric. The rider's power through their own gears does not have a power curve for our purpose. The rider's 150 watts or whatever is constant and just as available at 6 mph as at 8 to 10 mph. That alleviates the supposed loss of power of the system as the motor moves from it's maximum on the power curve, since the riders contribution is rising in real terms as speed reduces, due to lower hill climb speeds requiring less power. Therefore the rider power proportionally increases as the motor power reduces, and the loss is not just self cancelling, it actually becomes a gain.

For example, to illustrate this with power/climb equations, take a typical 25 kilo legal e-bike with a moderately powerful Hall effect hub motor of 500 watts peak, together with a 75 kilo rider, and so as not to unduly benefit my case, assume the rider is not very fit and can only output 100 watts. That combined 600 watts is available at the peak torque point of 8 mph on a legal bike, and that can climb a 17% hill at that speed. At the reduced speed of 6 mph, only 390 watts is available from that typical Hall motor, plus the riders 100 watts, so 490 watts, down 110 watts. But at 6 mph that reduced power climbs a steeper 18.5% hill. At a still further reduced speed of 5 mph, the motor gives just 335 watts, plus the rider's 100, so 435 in all. But at 5 mph that even further reduced power climbs a 20% hill.

So you can see that on e-bikes, the Hall effect hub motor bike's supposed performance loss when off the torque curve peak is a myth, these e-bikes actually climb better when off the curve peak at lower speeds, simply because less power is needed for climbing at lower speeds, while the rider's power contribution doesn't reduce at lower speeds. In fact for average riders who are capable of 150 watts or more, the gains with reducing speed are even greater. That's just one of the many reasons why standard gear theory for motor vehicles does not apply to e-bikes.

The power consumption efficiency loss that does still remain after all those considerations is substantially offset by the increased performance of the powerful hub motor in most riding conditions, higher average speeds and much higher hill climb speeds. I've proved this with my Q bike, but since you may think I'm biassed, let's look at how someone arguing that drive through gears are more efficient unwittingly proved completely the opposite in two of his posts entered at the same time.

In one post he argued that hub gears like those on the Twist had to be more efficient due to the fact that a hub motor covering the range had to be off it's maximum efficiency point part of the time. In his other post he stated that his Torq was giving an average of 37 miles range with efficient use of the throttle. On his 360 watt/hour battery that means a consumption of 9.7 watt/hours per mile. The typical 5 speed Twist drive through gear bike gives a 17 mile range on it's 156 watt/hour battery, that's a consumption of 9.2 watt/hours per mile.

So the Twist uses 5.2% less power than his Torq. But the the Twist averages 12.5 mph at best across the ground, viz tests by A to B, myself and others, while a restricted Torq like his typically averages at least 15 mph across the ground, which is 20% faster. Many riders do much better. Since the performance is better by at least 20% for a cost of only 5.2% of extra power, his Torq hybrid combination is the more efficient, and thus he defeated his own argument.

Of course the Torq or any other hub motor bike isn't more efficient for most riders who get shorter ranges, but the example given plus the evidence of my own Q bike shows how close hub motors can now be to the efficiency of drive through gear bikes, making the design of new complex drive through gear systems not worthwhile. The fact that hub motors are cheaper to make is a red herring, there will always be manufacturers willing to exploit a market niche if one exists, but the failure of the Panasonic system to gain anything like the acceptance enjoyed by hub motors after nearly seven years of trying proves to other manufacturers that it's not worthwhile following that example or that of the many other drive through gear bikes that failed before it.

I believe most riders share my point of view, that what matters is a bike that does the job well, regardless of any tiny increase in current used, and on that basis, hub motor bikes comprehensively outclass any drive through gear bikes yet seen. That's why my Twist sits in the garage as a reserve bike only, although it's a better cycling only bike, while it's my hub motor Q bike which works daily for me and does a far better and faster overall job than the Twist.

Therefore from everything said here, the only commercially valid reasons to design a cycle motor drive through gear system now are for either heavy towing or load weight carrying, or just possibly for very high speed use off road. There is a theoretically valid reason for providing a two speed Hall motor drive as I've previously explained, but as shown above, the practical advantage is so small it's scarcely worth bothering.

Choice, Hub Motor or Drive Through Gears.

First it must be stressed that these are pedal assist bikes as will have been clear in the first section of this article, the rider's input an important part of the concept. If someone is unable to contribute at least 100 watts minimum of cycling power, an e-bike is not the right vehicle for them, though they might cope with some models in a very flat area. Many bikes demand that the rider pedals anyway before they will work, and this could become the law before long.

So will a gearless hub motor e-bike do your job? In nearly all cases yes, but for the very steepest hills of around 1 in 5 (20%) or steeper, and/or with heavy riders, the choice may be limited or not exist at present.

For a lowish powered rider capable of 100 watts of cycling power and up to 70 kilos weight, eZee's most powerful models, the Forte and Forza will climb 1 in 5 hills at a low speed, but on steeper extreme hills or where the rider is heavier, the rider will need to be a bit fitter and stronger.

Alternatively they could choose one of the very few drive through gear bikes on the market which will climb almost anything, but only with good rider input and rather slowly. Currently the Gazelle Easy Rider is such a bike, rather slow and low powered, but with very low gearing to handle climbing. The couple of others use the same motor unit and are similar, but more expensive.

For all other circumstances with less extreme hills, the choice widens considerably, and even quite low powered hub motor bikes like the Powacycle Windsor/Salisbury models will climb the usual hills met without any problem given a moderate rider input.

The eZee Torq is currently the one exception to these general indications of hill climb ability. Designed for the rider with more sporting inclinations and with less emphasis on comfort, it only works at it's best with a fairly strong rider who is prepared to work with the bike to achieve it's maximum performance. For that sort of rider it's a very fast and rewarding ride.

Many thanks to Flecc for permission in reproducing this article.

If you would like to discuss this topic further then please visit our Technical Forum