Forum du parfait patenteux - NormandNadon.com

I have a plan to try and prove what the real world loss of efficiency is with a NuVinci hub. First I will try to explain the factors that affect the efficiency.

NuVinci hubs are a tilting ball type transmission which works by using traction. This means there will be a rolling resistance similar to that of a train’s wheel on a rail. This resistance is proportional to the pressure between the rolling surfaces, the higher the pressure the more the resistance. It makes sense that the efficiency of power transfer from one traction ring to the other via the balls is dependent upon the clamping pressure. To avoid slippage the clamping pressure must increase as torque increases. Therefore the efficiency is inversely proportional to torque. If the torque transferred through the transmission is increased the efficiency will decrease.

There are three factors that can decrease the efficiency
1. Low cadence
2. Low road speed
3. Increased power
It is of course possible to increase the power without increasing torque by also increasing rpm but an increase in power tends to involve some increase in torque. There is also another factor that affects the efficiency of a tilting ball transmission and that is ratio. The further from the 1:1 ratio, the lower the efficiency.

So under what conditions will a NuVinci hub be least efficient? I recon when climbing the steepest hill as this is when all three things that decrease efficiency occur at once. To climb a really steep hill the lowest ratio will be needed but due to the min sprocket ratio it won’t be low enough so cadence will be low. A lot of power is required to keep moving and at a below ideal cadence will result in greatly increased input torque. Even at maximum power output road speed will be low resulting in increased torque on the output side as well.

What I’m considering doing is comparing with the help of a HRM the effort required to ride three different bikes up the steepest surfaced hill I know of.

The three bikes are
A NuVinci equipped steel framed 1990s maintain bike which is set up for commuting.
A derailleur equipped steel framed maintain bike set up for off road use and trailer towing.
A cheap derailleur equipped aluminium framed road bike.

The hill or the steepest section of it that I will use climbs 30m in a distance of about 100m giving an average gradient of 1 in 3 or 33%. Short sections of it must be nearer to 1 in 2 or 50% see attached photos.

I ‘m hoping to do the test in the next few months and the results will be posted here.

Seems like a good test but it will still leave some room for interpretation an subjectivity... But still, the most precise test for now!

Could you mount your nuvinci wheel on the same frame as one of your two derailleur bikes? This would make a direct comparison and more accurate results.
The other way would be to add weight to the bikes that are lighter to make them all the same weight and use comparable tires on all bikes. Also, make sure you have the same riding position on all bikes for the test... best to eliminate any possible variable.

I was thinking that the test would be a comparison between the different bikes and as I’m using a HRM not a power meter the data is never going to be very accurate.

However it would be nice to eliminate some of the variables. I realized when I thought about it that if I don’t need to shift the ratio of the NuVinci hub it’s easy to change wheels. Just bolt the wheel on and set the rear derailleur to the single sprocket.

(2013-04-22, 18:07)Oran a écrit : [ -> ]I was thinking that the test would be a comparison between the different bikes and as I’m using a HRM not a power meter the data is never going to be very accurate.

However it would be nice to eliminate some of the variables. I realized when I thought about it that if I don’t need to shift the ratio of the NuVinci hub it’s easy to change wheels. Just bolt the wheel on and set the rear derailleur to the single sprocket.

That would be even better...

Even better, fixed ratios on both setups: You don't change gear on the derailleur and you set the nuvinci to an according ratio... (Just a thought) But good luck on the steep hill!!!

We finally had what could be called a warm day last Friday so I thought I would do some efficiency testing. I loaded two bikes on to my trailer and used the third to tow it. Three miles towing 30kg (66lbs) of bicycles and 20kg (44lbs) of trailer got me well warmed up ready for the testing. By the end of the afternoon I had climbed the hill 13½ times and got some reasonable data.


Some more details about the bicycles.


First is the bike I use most of the time and is the one I have fitted my NuVinci Hub to. It’s a steel framed Raleigh mountain bike from the 1990s which I have owned since I was 15. I have fitted slick tyres and use it most for getting to work so I will call it the commuter.
Ratio during test: 34t chainring, 18t sprocket, NV hub internal ratio 0.5 = 24.6 gear inches
Approx weight: 18kg (40lbs)

Second is an old mountain bike I have modified for trailer towing. It has a third brake lever to operate the trailer brakes and a quadruple chainset as I added a 15t chainring which gives ratios down to 13 gear inches.
Ratio during test derailleur: 22t chainring, 23t sprocket = 25 gear inches
NV hub : 32t chainring, 18t sprocket, NV hub internal ratio 0.5 = 23 gear inches
Approx weight: 17kg (37lbs)

Third a cheap aluminium framed Raleigh road bike. This one doesn’t get much use as it isn’t very suitable for the roads where I live.
Ratio during test derailleur: 34t chainring, 28t sprocket = 33 gear inches
NV hub : 34t chainring, 18t sprocket, NV hub internal ratio 0.5 = 25.5 gear inches
Approx weight: 11kg (24lbs)

And now the data. I tried to maintain the same speed for the entire test but it proved to be difficult due to the noticeable difference in the amount of effort required to get each bike up the hill. Between each climb of the hill I rested for 5 – 10 min to allow my heart rate to drop and to swap bikes and change wheels.

The data clearly shows that NuVinci hubs are less efficient than derailleur gears (see graph), which I already know, the question is can the data I have give any idea of how much less efficient it is. Heart rate is not a good indicator of actual power output. I have the data on a spreadsheet at the moment and I‘m thinking about how it can be analyzed. I don’t think there is enough data, it would probably have been better to use one bike and keep swapping between derailleur and NuVinci.

Wow! Almost killing yourself out of exhaustion for science... That is what I call commitment!

This test puts the NuVinci hub in a bad position when it comes to climbing... Were you riding on the underdrive ratios mostly? They are know for really bad efficiency in underdrive due to the spinning effect...

For now, I am always on the overdrive side of the hub. I did not have the chance to try hard climbing yet, but I will probably make a long run to Rawdon, Qc. next week (super hard climbing, 105 meters elevation in around one kilometer (10 to 11%).

40 pounds of bike plus the rider needs more torque than my own setup at 27 pounds plus a 135 pounds rider on top! Results will be interresting.

Thanks Oran for your good work!

Towing the trailer home afterward got a bit tiring but I found climbing the hill to be a good form of training and felt fitter the next day.

With an average gradient of about 30% and nearer to 50% in places I had to keep the hub in full underdrive. The aim of doing this test was to make the NuVinci hub as inefficient as possible to see how bad it can be. Under more normal circumstances the loss of efficiency will be somewhere in-between.

I would estimate by looking at the graph that the loss of efficiency between derailleur and NuVinci is similar to that between a road bike and a heavy mountain bike with 2 inch tyres.

I’m still thinking of a way to get meaningful values for the loss of efficiency out of this data. If anyone has any ideas post them here.

I'll work on that tomorrow... I am a bit of an Excel Ninja!

You’re welcome to have a go, I have attached the spread sheet. I’ve done a few calculations and produced the attached graph which shows that more effort is required to achieve the same power output. The next stage is to link heart rate to power as this represents power input.

A more useful test might be to use one bicycle and keep swapping the wheel between derailleur and NuVinci. The important thing will be to maintain the same amount of effort for every test by trying to maintain the same heart rate and therefore similar power input. It should then be more obvious that there is a difference in power output and this should be the loss of efficiency.

What formula did you use to obtain your approx. power requirement?