drivetrain efficiency

[mig]

given that each is well maintained, new and so forth are there any statistics that show the relative efficiencies of different drivetrains? i.e how much input power is lost in each system?
derailleurs, hub gears, fixed gears, single speeds etc

[bobc]

This is stuff I've looked for over a long period for the greenpower racers. The only credible & useful data source I've seen is the HPVA article from 2001 which I've seen linked on here a couple of times; unfortunately this is now old & limited in scope.
The plain fact is that measuring mechanical efficiencies to an accuracy of a fraction of a percent is well nigh impossible. Add to that the fact that you can't believe anything a manufacturer says, be it shimano, rohloff or anybody else (they're all have an axe to grind) few independent data sources are funded well enough.
A chap called Tony something was trying to build the hardware to do these measurements for one of the greenpower teams. He was a proper old school engineer, I saw his kit and would have believed his results: he never published any, and that must be because he himself could not achieve consistency.

[NATURAL ANKLING]

Hi,
And if you bear in mind as soon as you set of on the road your exposed chain will lose effeciency before the end of your trip.
Fully enclosed chain sounds best but it will beforgotten because you cant see it.
Shaft drive is a nice idea but I would guess that a hub gearbox is needed so double hit on effeciency.
Belts dont seem to have taken off yet, I would gues that china and the 30 % more expensive chain manufactures in India are still below that cost of a alternative drive chain.

[squeaker]

This is stuff I've looked for over a long period for the greenpower racers. The only credible & useful data source I've seen is the HPVA article from 2001 which I've seen linked on here a couple of times; unfortunately this is now old & limited in scope.

This one?

Tests of efficiency for the derailleur type chain drive indicate that the overall efficiencies for the transfer of power from the front drive sprocket to the rear sprocket range from 80.9% to 98.6% depending on the conditions of drive operation. Primary factors affecting the efficiency include the sizes of the sprockets in the drive and the tension in the chain.

covers it nicely

There are some useful conclusions, though

[Brucey]

in HPVA archives there are also some other articles including some where a simple method of measuring static/slow speed transmission losses directly is trialled.

Using this approach the transmission is loaded at the input and the output in a more or less realistic fashion, and the measurement is taken (almost directly ) of the force that is required to sustain motion at a constant speed. The set-up is such that there is no net work between the input and output except for any losses that might occur.

I think that with a little effort this approach could be developed further and used to measure low speed losses in an accurate and reproducible fashion. However I do wonder if it will ever work well over a higher range of speeds, and that worries me; some of the losses are literally because the chain is flapping up and down and these will vary with speed in a way that may not be easy to predict or measure.

cheers

[bobc]

There are a few in ihpva archives e.g.
http://www.ihpva.org/HParchive/PDF/hp51-2001.pdf
but I was thinking of this one
http://www.ihpva.org/HParchive/PDF/hp52-2001.pdf kyle & berto
which covers a number of hub & hybrid possibilities as well as deraillieur
An important observation which has been very useful for us is that 12 or less teeth on a sprocket give a significant efficiency drop.

[Mick F]

An important observation which has been very useful for us is that 12 or less teeth on a sprocket give a significant efficiency drop.

The modern system seems to be small chainwheels and 11t cogs.
This equals poor efficiency and increased tooth wear.

In the "olden days", the smallest sprocket you could get was 13t, therefore you needed bigger chainwheels to get the high ratios.
In order to get my top gear of 12/53 that I have now, I would have to have had a 13/58

[Brucey]

this paper;

http://www.borgwarner.com/en/enews/Assets/2012_10_03_Engine%20Expo%202012.pdf

reports some extensive tests carried out on camshaft drive systems in modern engines. Obviously torques and speeds etc are not as bicycle drives but there are interesting results at low speeds that may be relevant. For high speed use you might well choose a belt but the low and medium speed results are interesting.

In a nutshell;

-a wet (oil bath) belt was slightly less efficient at low speeds than a chain, but similar or better at higher speeds.
- a dry belt was less efficient than a chain at all speeds.

Now the guys carrying out the tests may have had an axe to grind and could have skewed the tests accordingly; however they claim they compared like for like; they would say that, wouldn't they...? Note also that the chains were not roller chains which could be an important factor too.

cheers

[John Dalhart]

An important observation which has been very useful for us is that 12 or less teeth on a sprocket give a significant efficiency drop.

And yet, no one reports being able to 'feel' this significant efficiency drop, while at the same time many report being able to 'feel' the much smaller efficiency difference between ratios in an IGH.

JD

[Brucey]

I don't think that the differences in some IGH ratios are 'small' at all.

If you read the paper in question you will see what I mean.

cheers

[John Dalhart]

I don't think that the differences in some IGH ratios are 'small' at all.

But the differences in other IGH ratios are relatively smaller than the imperceptible drop when employing 12, 11T &etc cogs, which you would see if you read the linked report.

All the best,
JD

[Brucey]

I've read that article several times....

My point is that the IGH losses they measured varied from virtually nothing to over 15%; in some cases between adjacent gears in a single IGH.

A 12T sprocket is less efficient than a larger one (and indeed I don't ever remember using one that I thought was properly smooth...) but we are only talking single figures of % extra losses here vs (say) a 16T.

I think it is a 'false economy' to have a transmission that consists of smaller chainwheels and sprockets on the basis that it is lighter or something; the hill would need to be very steep before the weight advantage would eclipse the loss of efficiency even if the extra losses are only 1-2%. IME such transmissions are more highly loaded, wear out many times faster and offer only one (possibly) compelling advantage; if you use smaller chainrings especially, you can get get a larger total %age range of lower gears from a given set of mech capacities.

Even when running an IGH I would always recommend a chainwheel/sprocket of at least 46/18 unless the ratios made it impractical for some reason; the loadings are much lower and even chainwheels made in relatively soft materials have a much longer working life.

BTW many chain manufacturers have published documents which explain the nature (but not always the extent) of chordal losses. It is actually more complicated that you might think, being sensitive to the number of links in the chain, the size and spacing of the sprockets, and the gear ratio, all in a non-trivial way. Such data does not allow for the behaviour of a spring loaded tensioner like a derailleur, however; my suspicion is that there are extra losses in there (over and above simple pulley drag) which have yet to be allowed for.

In actual fact, I think a good IGH transmission can be more efficient on average than most derailleur transmissions; I wouldn't have used one for as long as I have done otherwise. However I also note that it is evident that almost any derailleur transmission you cobble together -provided it isn't worn out etc- can be more efficient than a bad IGH transmission.

cheers

[Tigerbiten]

All I know is on my bent trike ............
For the same gear inches, pushing my 38t chainring is easier than my 53t chainring.
Gears 8-14 are the easiest for the Rohloff followed by gears 1-7, 2.45x step-down from gears 8-14.
I get more drag when the Schlumpf HSD is in overdrive mode, 2.5x step-up.
So gears 15-21 are a little bit more work.
Running overdrive mode (2.5x step-up) in gears 1-7 (2.45x step-down) is a silly waste of effort.
But how much % loss each gear gives me is ..........