How do Pedelec systems work?

[frank9755]

By this I mean 'How much power do different pedelec systems deliver and how do the controllers decide when this power is delivered?'

I am asking in the context of hub motor powered bikes with proper pedelec operation (power when you pedal without using throttle) - rather than:
- Panasonic crank drive systems (eg Agattu), which I understand deliver more power at lower cadences, and
- throttle bikes, like Ezee and Powabyke, where the pedelec is just an override for legal compliance reasons, and which I don't really consider to be pedelecs.

We've got two pedelec hub powered bikes - a Powacycle and Wisper. Both work similarly as pedelecs giving a pretty good result. Essentially, when you pedal, you get a certain level of power assist automatically. (Both bikes also have throttles - but let's ignore the throttle for this discussion). I've never been able to work out from observation if the level of assist varies or is just a constant and, if it varies, what drives the variation.

Does it give more assistance at higher or lower cadence? Presumably how hard you pedal is not a factor, as the controller has no way of measuring that. Is bike speed a factor? The Wisper does have a very inaccurate factory fitted speedometer so must have pretentions to measure speed.

Does anyone know how they work?

Frank

 

[flecc]

As well as sensing just the rotation as on the eZee sensor, there is often a crude speed control set by the spacings on the pedelec sensor Frank.

For an example, on the type where a ring of magnets is on a disc mounted on the BB spindle, a Hall sensor measures their passing. The frequency of pass can give the controller a speed indication, which the controller can use to enforce the legal limit when in top gear for example.

Less frequently, that can also be used Panasonic fashion to determine a drop in power given at higher pedal speeds. If it's variable like that, it's practically certain to reduce power with speed, matching the law in the EU which requires phase down of power towards the limit. It could be made to work the other way, but I doubt that companies would set out to flout the export market law in that way.

Some alternative sensors are a ring of blades passing between a photocell and LED pair, and either of those types mounted within a hub motor, that last mounting much less common now though.

Changing the blades or the number of magnets on a disc will change the characteristics of those types.

P.S. If the pedelec sensor disc is in or on the hub motor, it can measure absolute road speed, not just top gear speed of course.
.

 

[mk1]

The Bionx system uses a strain gauge as shown in this video YouTube - BionX - Under the Cover.

 

[flecc]

Some others use strain gauges as an additional feature, an option on the UM36 model for example, and these determine power levels of course. They don't normally determine whether rotation is present though, Panasonic and I believe BionX being exceptions to that, though strictly speaking the Panasonic measuring system isn't a strain gauge.
.

 

[frank9755]

I realise I asked the question ambiguously! What I was getting at was what power do pedelec systems deliver, rather than what technical solution is used. Obviously the two are closely related so you do have to consider both sides - so thanks for those comments.

Taking this into account, I now think that my two bikes measure cadence but not speed, giving a set level of power at any given cadence. I suspect that the level of assist (ie power in Watts) decreases as cadence increases, which would give a large amount of power to pick up from a virtual standstill, and also explain the tailing off.

So Martin, for the Bionix, which is a sophisticated system, do you appear to get more (or less) assist the harder you pedal? Does the resulting pedelec feel 'right'? I recall some people complaining that the UrbanMover one was a bit odd.

 

[flecc]

Taking this into account, I now think that my two bikes measure cadence but not speed, giving a set level of power at any given cadence. I suspect that the level of assist (ie power in Watts) decreases as cadence increases, which would give a large amount of power to pick up from a virtual standstill, and also explain the tailing off.

Yes, exactly as expected if cadence measuring only for power Frank, probably mainly for EU law, the only large export market requiring pedelec.

A complaint with the early eZee system where the throttle didn't override the pedelec was too much uncontrolled power at take off when the pedal was pushed, but now the throttle tames that. I take it that's the same on the Wisper?
.

 

[frank9755]

A complaint with the early eZee system where the throttle didn't override the pedelec was too much uncontrolled power at take off when the pedal was pushed, but now the throttle tames that. I take it that's the same on the Wisper?
.

Don't quite understand that (isn't pedelec a difficult concept to describe in words!), but with the Wisper (and Powacycle) what happens is you pedal and, after about a second, the power cuts in quite noticeably - but not at excessive levels. It doesn't start completely from zero speed, avoiding the wheelspin that might happen if it did. With either bike I wouldn't normally use the throttle.

 

[flecc]

The original eZee pedelec was completely isolated from the throttle Frank.

When switched to pedelec, the full motor torque/power acted immediately the pedal was pressed, there was no hestitation, and that was too much on their more powerful motors, scared the life out of many. A to B heavily criticised it on the early bikes.

Now that the throttle acts in pedelec mode as well, the amount of power can be fed in as required, and there is also the hesitation you describe. Therefore it seems as though it's the same as the Wisper system?
.

 

[frank9755]

Wisper, Powacycle and others I know of are like the old Ezee system - but there is no problem with excess power on start-up - perhaps because the motor is not as powerful.

I only know of Powabyke which works like the new Ezee bikes. Personally, I don't like that system as you have to use the throttle all the time!

 

[flecc]

I agree, the eZee pedelec has never really been satisfactory, always seeming to have something of the afterthought about it, probably because it was!
.

 

[fishingpaul]

frank how does the powacycle compare to the wisper,is there a noticeably large difference in power and hillclimbing abillity, which models do you have.

 

[frank9755]

Wisper 905se and Powacycle Windsor.
The Wisper is a good bit more powerful - faster on the flat as well as uphill. I've never ridden either up a steep hill but my wife felt the Wisper was almost as good a climber as a Powabyke when we were in the Kent / Sussex Downs the other weekend.

 

[jimmyengland1000]

For an example, on the type where a ring of magnets is on a disc mounted on the BB spindle, a Hall sensor measures their passing. The frequency of pass can give the controller a speed indication, which the controller can use to enforce the legal limit when in top gear for example.
.

On the Wisper, I thinkt the full power kicks in when assistance starts. It does so after a few rotations of the crank. Does what you have said above mean that by removing a magnet or two, the controller could be fooled into thinking that the road speed is less than it really is... and therefore the motor would continue to give assistance above the legal limit??

 

[frank9755]

Interesting thought...
For those who aren't familiar with it, on the Wisper (mine at least) the pedelec gives assistance to around 15mph, although the throttle can give assistance to higher speeds.
My ideal configuration would be to get the pedelec to give full throttle assistance. I suspect someone who knew what they were doing could possibly achieve this by re-jigging the controller wiring - connecting the pedelec sensor to the throttle connection. I suppose this mght lead to the Ezee-type problems which Flecc mentioned (too much power on start-up).
Another way to do it might be to replace the freewheel to give a higher top gear - adding an 11-34 megrange in place of the 14-34 which is fitted to mine.

 

[jimmyengland1000]

Interesting thought...
For those who aren't familiar with it, on the Wisper (mine at least) the pedelec gives assistance to around 15mph, although the throttle can give assistance to higher speeds.
My ideal configuration would be to get the pedelec to give full throttle assistance. I suspect someone who knew what they were doing could possibly achieve this by re-jigging the controller wiring - connecting the pedelec sensor to the throttle connection. I suppose this mght lead to the Ezee-type problems which Flecc mentioned (too much power on start-up).
Another way to do it might be to replace the freewheel to give a higher top gear - adding an 11-34 megrange in place of the 14-34 which is fitted to mine.

Full pedal assistance to the same level as the throttle (about 20mph on mine) would lessen the range though, eh?

I was thinking of removing some of the pedelec magnets on mine, as an experiment, but then I thought that it might make the pedal assistance more delayed, since the motor kicking in is dependent on a few revolutions of the crank being detected. One thing I have come across is a sensor disk on the wrong way round..... it made the motor spin (in the correct direction) only when the pedals were rotated backwards!

 

[flecc]

On the Wisper, I thinkt the full power kicks in when assistance starts. It does so after a few rotations of the crank. Does what you have said above mean that by removing a magnet or two, the controller could be fooled into thinking that the road speed is less than it really is... and therefore the motor would continue to give assistance above the legal limit??

If a bike's controller senses speed from the pedelec disc, yes. I don't think the Wisper does, judging from what David Miall has said about controllers in the past.

If the controller kicks in power after some rotations, removing some magnets could delay that though, meaning more pedalling before power arrived.

Another snag with actually removing magnets on some designs is that the controller might cut out again with slower pedalling, thinking pedalling had stopped with fewer pulses coming through. Rather than every other magnet taken out, say 5 from 10 leaving only 5, it might be necessary to use another disc with 8 evenly spaced magnets for example. This sort of thing is easier to change with photocell/LED reading and a rotating vane ring, a new vane ring easily made.
.

 

[frank9755]

Having checked the wiring diagram, I suspect the controller gets its speed information from the motor. There is a connection from one of the Hall sensor wires to the speed indicator - and no connection to it from the pedal sensor.

In which case neither removing magnets nor changing gearing would raise the pedelec cut out speed.

In an ideal world, I'd like to get pedelec assist right up to the full speed of the motor.

Before the pedelec problems with the Wisper got resolved, I made a workaround by taping the throttle fully open - allowing me to use the switch on the throttle to have either fully 'on' or 'off' - without the wrist pain from constantly holding the throttle open. I would start with it off and switch on when the bike was moving, getting 'pedelec' operation up to the full assist speed (c.18mph on mine). This works very well, but I wouldn't recommend it to anyone else (and I didn't post it on here for fear of encouraging anyone to copy me) as it clearly wasn't going to win any health and safety awards.

Thinking about it, I suppose what I have come round to saying I would like is the Ezee-type variable assist system (aka toaster control) that they put on the Forza - but I think it only needs 2 or at most 3 settings rather than the 8 I think that had.

Anyway - if anyone works out a way to re-wire the controller to give assist to the full throttle speed - please let me know!

Frank

 

[flecc]

Having checked the wiring diagram, I suspect the controller gets its speed information from the motor. There is a connection from one of the Hall sensor wires to the speed indicator - and no connection to it from the pedal sensor.

That's as I suspected, the controller regulating maximum speed direct from the motor, fitting with previous things David mentioned about the 905e and se.
.

 

[Beeping-Sleauty]

Less is more

On the Wisper, I thinkt the full power kicks in when assistance starts. It does so after a few rotations of the crank. Does what you have said above mean that by removing a magnet or two, the controller could be fooled into thinking that the road speed is less than it really is... and therefore the motor would continue to give assistance above the legal limit??

Yes, removing or covering some of the magnets works as you suggest, this came up in a thread exactly a year ago - Pedelec sensor rings - had a bit of fun with this, nearly pedalled my little hairy legs off. But this was on a brushless hub on a Eco Brand Whirlwind, so not sure it will work on a Wisper.

beeps

 

[rooel]

This thread is of interest to me using the Sram Sparc system, as there are definitely no pedal sensors in or around the cranks or the chainwheel, nor have I been able to find anything of that kind inside the motor hub on the rear wheel. Inside that hub there is a five speed mechanical hub and two 100watt motors mounted on each side of the axle which drive the hub (hence the wheel) through a system of linked nylon gears, and an electronic circuit board which controls speed and cut-in/cut-off for the motors.

The supply of power to the motors comes through that circuit board, and somehow or other the electronics sense when the wheel speed reaches 15mph, that speed being transmitted through the same gear system as the motors use to transmit their assistance to the rear wheel. It is definitely not a regenerative system but it seems that the speed at which the wheels turn and also the speed at which the external rear sprocket is turned by the pedals is transmitted back through the motors (even while they are still driving) to the circuit board which then controls the motor output or cuts it off at the legal speed limit.