Avoiding power idler by using Alfine hub gear

[[XAP]Bob]

On my rapto (with 100 degree wrap around the idler) the net thrust could exceed that of a traditional chai line if my in-head diagram is correct - by around 50% (I.e probably less bad than having a stronger rider?)


As for noise - the rapto idler is loud in the bottom two gears, and quiet above that, despite the wrap being many more than three links... I suspect it's all to do with alignment against the side plates.

[Brucey]

Having given it some thought there's a much simpler explanation for noise being dramatically lower when there are at least 3 links in engagement. It is not to do with an arc being defined by three points. The chain rollers each generate a thrust into the pulley that is aligned with the complementary angle (i.e. that bisects the included angle) between any two links that are under tension. For the rollers that are arriving and departing, the included angle is varying and therefore this thrust direction is constantly changing.

If there are only links of this sort touching the pulley, the pulley is liable to be on the move all the time WRT the chain, with the chain never settled in the teeth.

By contrast if there are three or more links in contact with the pulley, the middle ones provide a constant thrust into the pulley, because the complementary angle is constant. Provided the chain fits the pulley well, this will tend to stabilise the chain 's position and running will be quieter.

In addition to the above there are of course always chordal losses in any chain drive. These arise because the momentary radius of action of the chain varies as each tooth engages with the chain on the tension run. The effect is largest with small pulleys/sprockets. The chain run under tension must flap up and down slightly, and it is unlikely that the energy contained within this motion is properly conserved. The losses that arise this way are (if the chain moves a similar amount) independent of load, thus chain drives can appear to be more efficient at higher loads if they are run at the same speed. In addition if the chain tension increases, the chain may tend to flap less in response to the chordal action, which may result in smaller losses.

There are also losses associated with the fact that the chain changes direction around each pulley. Again momentum is not conserved in this action. Losses of this type increase with more chain wrap around things (a rear mech for example) and also increase with speed of the chain.

Interestingly at low chain speeds the chordal losses as the chain enters the lower pulley are likely to be large. However at some speed the losses may become smaller; the reason is that the movement of the chain progresses along the chain length as a wave with a finite (and fairly low) velocity. As the chain moves faster and faster, the wave may fail to propagate any distance along the chain. This means that instead of the whole chain length flapping up and down, only a small proportion of the chain may move, which must surely consume a smaller proportion of the energy, relatively speaking, i.e. the losses may no longer be pro-rata with the speed of the chain in the same way.

cheers

[StephenW]

Hello all

Thanks for the interesting responses. I'll reply to a few points, and then add some new thoughts.

Replies

[XAP]Bob: I was thinking of vertical bending, rather than horizontal bending or torsion. I'm much more aware of vertical movement than the other two, although I'm sure they do happen too. I know the Raptobike has a reputation for a direct feeling through the drivetrain. Do you think this directness translates to hill-climbing ability, or is it more just a nice thing to have?

Upwrong: I have the Terracycle toothed idler. A previous owner must have fitted it. Have you ridden other similar LWB bikes to your Stratus XP which do have a power idler? Did you notice a difference?

Brucey: Your mid-drive sounds very interesting. How did you tension the front chain? Perhaps this kind of mid-drive could also be used on a FWD or even a FWD tandem?! I guess that this is something that the frame really needs to be designed around from the start.

I hadn't thought about the effect of the chain sagging. Do you think that a "clutch" rear derailleur could be useful in this case? I was also musing about some kind of floating chain support, where the chain runs between two wide, smooth pulleys. These are allowed to move vertically, but are held at their height by friction. If the amount of friction is correct, the device will stay in the same place when it is supporting the weight of the chain, but will move to a new position when the rider selects a different gear and the chain needs to be in a different position. I haven't yet done the sums to see if the force after a gear change is significantly smaller than the force from the weight of the chain.

I haven't fully digested what you are saying about toothed idlers and forces, but I will give it further thought.

New Stuff

I went out for a ride on Sunday afternoon and removed the idler halfway around. When I took the idler off, I flew along! It was great! There may be a psychological advantage from having a quiet bike (and a psychological advantage is also not to be sneezed at), but I am certain that there was a genuine increase in efficiency. I can think of two reasons for this:

1. Friction resulting from idler.
2. Reduction in frame flex due to forces from chain being more aligned with frame tubes.

I feel that no. 2 is quite important.

In this instance it seems that the losses from the chain sagging were less than the losses which were avoided by removing the idler.

Consider this M5 m-racer. I have never ridden this bike, but I find it hard to imagine that it would climb well in a small chainring. The load from the chain is not aligned with the direction of the frame tubes, meaning that they are loaded in bending rather than compression.

m5.jpg (43.86 KiB) Viewed 1638 times

Consider this "stick bike". It seems to me that such a bike would be ideal to run without an idler - the "stick" would be under almost pure compression, no vertical bending forces. Unfortunately, by the looks of things, the chain might rub on the chain stay in certain gears, due to the dropouts. I suppose that on a steel frame these could be modified. I think some kind of cover for the chain as it passes under the seat and the rider's right leg could be a good idea, also to stop the chain flailing around when riding over bumps. This could be a piece of square plastic drainpipe.

[Brucey]

re the chain tensioning; the mid-drive hub is mounted in dummy dropouts and can be moved back and forth slightly to allow the front chain to be tensioned. Possibly the efficiency is no better than a good idler, but the scheme has other advantages.

Re the chain sagging; a clutch mech typically has a damper that acts on the upper pulley when it tries to turn backwards. This will damp the action of the lower run of the chain when it tries to flap around and that helps the whole chain to flap about less. I doubt that such a device would rally calm the flapping around that could result on a chain run that is as long as that on a recumbent though.

cheers

[StephenW]

Hi all

I realised that the boom had been cut down by a previous owner. Because I am quite tall, I have the boom extended fairly far out. This meant that not enough of the boom was inside the frame, causing undue flex. I was able to buy a spare full length boom from someone on BROL, and this has fixed the problem. The old boom had been slightly squashed out of shape Oops!

I've been thinking a bit about two mysteries: firstly, Jan Heine's "planing" idea, meaning that a more flexible frame seems to climb hills better, and secondly that recumbents supposedly don't climb hills well. Are these two things related? I have an idea that they might be.

My idea is that the abdominal muscles contribute quite a lot to cycling. A flexible upright frame might allow the hips to move sideways relative to the pedals and the abdominal muscles to be used more effectively. Since the hips are more fixed in position on a recumbent, it is harder to use the core muscles effectively.

According to Heine, flexible chainstays are not good. It is a flexible top tube and seat stays that are desirable. This fits with my experience - I don't think sponginess in the drivetrain is beneficial at all. Even a relatively flexible upright is likely to be stiffer than many recumbents in this regard I suspect.

What I think would be best, is a drivetrain which is as stiff as possible, combined with the necessary sideways flex to allow the core muscles to contribute to the effort of pedaling. I can think of three ways this might be achieved:
1. Front wheel drive. The front part of the frame, as far back as the power idler, can be made very stiff. Rearwards from this point, the frame can be made more flexible.
2. Stick bike with no idler. In this case, the chain is loading the frame in almost pure compression. This gives a stiff drivetrain. The stick could have a certain amount of sideways flex.
3. A seat which consists of two parts: a back which is firmly fixed, and a bottom part which is flexibly mounted so that the hips can move a bit. Perhaps it could be able to swivel, with a spring pulling it back to the centre.

What do you think?

[Brucey]

my take on the planing thing is that there may be a real mechanical inefficiency there with a less stiff frame, but that it may be more than made up for by giving the rider feedback that allows 'a better pedal stroke' to be achieved. Whether this translates to a recumbent or not remains to be seen.

cheers