Hub Dynamo efficiency comparisons?

[Brucey]

Having used old-style Sturmey Archer hub dynamos for over 35 years, I have always been of the view that the drag was negligible with the lights off and barely more than that with the lights on (only making 1.8W helps here I expect... ).

I have recently been looking (finally) at LED light upgrades for the old dynohubs and also at newer designs of hub generator.
Some of the (major?) annoyances found with modern hub generators (such as vibration being apparent when riding the bike) are largely absent in the older SA design, IME.

[Perhaps controversially, I believe that you don't really need 3W (especially with LED lights) for urban commuting. I've seen quite a few bikes (equipped with 2.4 or 3W hub generators) which are shipped with lower power LED front lamps. Although these lamps don't really cast enough light to see by in unlit areas (having just one 5mm white LED within them), few people bother to change them for urban commuting; yet the weight, parasitic drag (lights off) and cost of a 3W generator is still incurred.]

It seems that producing 3W or more with a modern hub dynamo probably 'costs' 6W of effort at 20kph, and still 'costs' about 5W at 15kph (at which there may not quite be 3W produced). For a reasonably fit cycling enthusiast this is about 5% of the total power output which isn't a big deal. Nor is the 1W or so drag with the lights off from a good generator. But for an untrained or casual rider at (say) 15kph, 5W is approximately 10% of the power available being used to drive lights which (in an urban environment) may not need to be as powerful as this in order to acheive their primary function. Note that even thought they are not perfect by any means, 'legal-in-town' battery driven LED lights consume a tiny fraction of this power.

So I'm interested in lower power lighting solutions for urban use as well as higher power systems for the open road.

I had hoped to find a comparison of generator drags that included an old-style SA dynohub as a reference point, but I've failed. Does anyone know of any data out there?

I found these;

many lights and generators reviewed;
http://swhs.home.xs4all.nl/fiets/tests/verlichting/index_en.html

a comparison of generator efficiency (2005)
http://www.bikequarterly.com/VBQgenerator.pdf

(from Schmidt data via our very own CJ ~10 years ago?) generators compared
http://www.myra-simon.com/bike/dynotest.html

I need to check the HPVA archives yet too.

Are there any more out there?

cheers

[CJ]

Even back when I comissioned that dynamo test, any surviving Sturmey Dynohubs were so old that I didn't think it was worth testing one. It might have more or less drag depending on how rusty its armature had got in the meantime and how it had been treated generally, so the data would be meaningless. A lot of 'Dynosaurs' have also lost power due to partial de-magnetising when the armature is removed for maintenance without simultaneously sliding a keeper into place. Who still has a keeper after all?

But if you have one knocking around, some modern LED dynamo lamps work perfectly well with them. For whereas the brilliance of a bulb falls off a cliff with reducing current, with an LED the output is near enough proportional to the input. I kitted out my son with such a setup on a college bike, built entirely from bits I had lying around in the garage, even the frame, even the spokes. We were especially sad when that bike was stolen and both still miss it much more than the other ones. A lot of love in that bike - irreplaceable.

You are right that we don't need 3W with the latest LEDs and there are 1.5W systems in the offing from Shimano etc. I expect they'll be on the market by winter 2013/14, so it could be worth hanging on. However the lights-off drag doesn't look likely to be any less than a good 3W.

As for vibration from magnetic 'cogging', I find I am hardly aware of it. Even with the lights on I don't notice the drag from the Shimano hub on my tourer. It's a non problem. So although I am currently preparing a hub dynamo article, the upshot is that even the worst ones are so very very good that only a true princess would tell the difference. You buy a more expensive generator because it looks prettier, weighs less and comes with a longer guarantee against bearing failure, not because it has less drag - but usually it does too.

[RJC]

Recent comparison, in German but the tables showing brightness & lights on/off drag are easy to understand:
http://www.nabendynamo.de/service/pdf/ar_09-10_2010.pdf

Other articles by Schmidt:
http://www.nabendynamo.de/service/links.html

[BeeKeeper]

This data from SON: http://nabendynamo.de/produkte/sondelux_en.html

Of course it is written to show their products in the best light (groan) but it does suggest that at 15 kph the power absorbed by the dyno is about 4W. Note the power output of their dedicated LED dyno is 3W.

The second graph shows light output against different sorts of light at different speeds. Unfortunately, it only shows results for their dynos but the curves seem to suggest there is plenty of output from about 10-15kph upwards. However, I am impressed by my new SON28 hub and Cyo Sensor front light as the small daylight running LEDs it has start to flash on an off when the bike is just being pushed along at walking pace. I guess they flash at all speeds but above a dawdle it is too fast for the human eye to detect.

[RJC]

A more recent test including the Shimano 1.5W & some Shutter Precision hub dynamos.
http://fahrradzukunft.de/14/neue-nabendynamos-im-test

[Brucey]

Thanks to all for your replies; much appreciated.

Lots of test data to look through!

Re older dynohubs;

I have several older dynohubs and they just seem to refuse to die.

For many years my 26x1-3/8" wheeled 'town bike' has had an 'FG' hub which I have converted to five-speed operation. This is being rebuilt right now, but will have LED lights when it is reborn I expect.

My 'training' /light touring bike has a front dynohub which (in my use of ~30 years) has lived out of doors and done many tens of thousands of miles. It still isn't draggy, rusty, or appreciably reduced in output, despite being ~56 years old now.

One of the few things I don't like about the front dynohub is the weight of it. Not the dead weight, but more where it is on the bike. With a nice springy fork, it seems (to me) to ride differently than a lighter wheel. I plan to build another set of similar wheels with another modified FG (or perhaps just an AG) hub in, so that I can compare the effect of an alternative solution (with the generator at the rear) more closely. [It is of course a fair bit heavier than a modern hub generator; the magnet and stator alone weigh more than Shimano's current, er, 'current offerings' (groan... ).]

When maintaining older SA dynohubs there are a few tricks that can help; first and simplest, the stator and magnet can be removed together as a pair. Second, a spare stator (from a scrap hub) works very well as a keeper.

Between these two, I have found that it allows enough access so that any rust particles can usually be removed without risk of demagnetising the magnet further.

Incidentally, I have an idea which will (quite simply I hope, if it works...) recover/boost the magnetism in an old dynohub magnet.

CJ, If you would like an original front dynohub for your test comparison, I have a (~1980 I think) SA front dynohub built into a spare wheel, which has seen almost no use (I need to dismantle this wheel at some point BTW). I am more than happy to loan this (as a wheel or a bare hub for comparison purposes in your article) if this might be of interest.

cheers

[Brucey]

A more recent test including the Shimano 1.5W & some Shutter Precision hub dynamos.
http://fahrradzukunft.de/14/neue-nabendynamos-im-test

google-trans-mangulated linky

http://translate.google.co.uk/translate?hl=en&sl=de&u=http://fahrradzukunft.de/14/neue-nabendynamos-im-test/&prev=/search%3Fq%3Dhttp://fahrradzukunft.de/14/neue-nabendynamos-im-test/%26hl%3Den%26prmd%3Dimvns&sa=X&ei=VkOBULviO8K80QWK44DoBA&ved=0CCcQ7gEwAA

v interesting....

cheers

[james01]

I have recently been looking (finally) at LED light upgrades for the old dynohubs

http://www.reflectalite.com/LEDpage.html

If I were the lucky owner of an old SA hub I'd certainly want to check out these Reflectalite LED bulb upgrades. The SA hub is specifically mentioned in their narrative, this could be a simple and cheap way of enjoying modern performance from the original lighting system.

[RJC]

Surprisingly the lights-off drag of Shimanos new 1.5W dynamo is higher than their 3W 3N80.
Maybe they are trying to improve this, hence the delay in launching this dynamo. I think it
was first announced at Taipei 2011.

[Brucey]

Surprisingly the lights-off drag of Shimanos new 1.5W dynamo is higher than their 3W 3N80...

it is also slightly higher than the drag when the lights are on, isn't it?

-not the first time we have seen that from a shimano hub generator IIRC.

cheers

[Colin Jenkins]

Hello Brucey.

I have an old (1974) Sturmey-Archer GH6 Dynohub in use. It has covered about 45,000 miles from new and I have serviced it (new bearings & grease) 3 times. Easy to do, good quality engineering, and there is absolutely no need to separate the armature and stator for this job. Unfortunately, that is easily done by the unwary and leads to reduced output. Be aware of this if seeking to buy a second-hand one today.

The GH6 delivers 6volts at 1.8 Watts and the filament lamps of the day were designed to work at this power level. They always seemed ok to me in times gone by, and still are in road-lit urban areas, but modern high-powered car lighting over-powers them on unlit roads.

I decided I needed to upgrade the lights, but my old GH6 & Wolber rim will soldier on for many more thousands of miles and I really didn't want to just scrap these old and well-sorted friends.

Logic indicates that modern LED lights may not demand as much power as filament lamps, so I gave this a try by installing B&M LED lamps front and rear to see if it works. It does. There is a caveat: the LEDs are not as bright as they are on my other bike which are fully-driven by a 3 Watt SON dynohub. But they are brighter at all times than the original filament lamps, so the upgrade was effective. Stand-lights are a bonus here too.

The Sturmey-Archer & B&M combination works well-enough that I have no plans to replace my Sturmey-Archer Dynohub with a modern one - it's still doing a good job.

Regarding drag - I have always been unaware of it. Lights On or Off, Sturmey-Archer or SON - never an issue. I think the magnetic "cogging" effect felt at standstill cancels out on rotation to virtually nothing.

I hope this helps.

Best regards

Colin.

[interestedcp]

This 632 page pdf file has a lot of raw measurement data on several hub and side dynamos (and dynamo lights). Many of the magazine articles about hub dynamos you see on the net, have used Mr. Olaf Schultz' data.
http://www.enhydralutris.de/Fahrrad/Beleuchtung.pdf

It is in German, but if you are only interested in efficiency, then just look at the tables under the symbol " η" (wirkungsgrad=efficiency). There usually is a table for both 12 and 24 Ω resistance.
E.g., table 2.32 on page 72 shows the SON20 has a η (efficiency) = 74.83% at 11,03 km/h at 24 Ω. Symbols are defined in the beginning of the document.

He does mention the SA GH-6 but hasn't measured it for several reasons, among these that it wasn't street legal in Germany at the time (not 3W).

Apparently there is a new German standard that agrees with your reasoning that less than 3W makes sense for Urban riding. For the moment only Shimano has a product for the new 1.5 watt standard, the 1N70 hub dynamo. Not really something that I drool over, but I understand bicycle manufacturers that sells really cheap low end commuter bikes; cost cutting is crucial in that segment, and the new 1.5 Watt standard may result in even cheaper dynamo systems (and lighter too) that is more than adequate for the casual commuter.
Even my old halogen bulb B&M Lumotec Oval is probably more powerful than 95% of all the commuter lights I see daily here in Copenhagen, so a 1.5 Watt dynamo system with LED's would be a major improvement for most commuters here.

[Brucey]

This 632 page pdf file has a lot of raw measurement data on several hub and side dynamos (and dynamo lights). Many of the magazine articles about hub dynamos you see on the net, have used Mr. Olaf Schultz' data.
http://www.enhydralutris.de/Fahrrad/Beleuchtung.pdf

He does mention the SA GH-6 but hasn't measured it for several reasons, among these that it wasn't street legal in Germany at the time (not 3W).

Apparently there is a new German standard that agrees with your reasoning that less than 3W makes sense for Urban riding.

Even my old halogen bulb B&M Lumotec Oval is probably more powerful than 95% of all the commuter lights I see daily here in Copenhagen, so a 1.5 Watt dynamo system with LED's would be a major improvement for most commuters here.

Thanks for the link, much appreciated; I'd found the (same?) data on web pages already [e.g. index here http://www.enhydralutris.de/Fahrrad/Beleuchtung/node449.html but the pdf is very useful;

I think the 1.5W standard is a sensible appraoch for a wide variety of utility cycling purposes; it will still be about the same power -and a good deal more light- than was common with either BS battery lights or dynohubs in the past.

The target (or dream) for such a hub generator would be lighter weight, (say) 0.5W no load drag and 75% efficiency at 1.5W output, i.e. 2W under load.

The GH6 delivers 6volts at 1.8 Watts and the filament lamps of the day were designed to work at this power level. They always seemed ok to me in times gone by, and still are in road-lit urban areas, but modern high-powered car lighting over-powers them on unlit roads.

The Sturmey-Archer & B&M combination works well-enough that I have no plans to replace my Sturmey-Archer Dynohub with a modern one - it's still doing a good job.

Regarding drag - I have always been unaware of it. Lights On or Off, Sturmey-Archer or SON - never an issue. I think the magnetic "cogging" effect felt at standstill cancels out on rotation to virtually nothing.
Colin.

Colin's experience tallies with mine although I have not invested in a B&M light as yet. I (perhaps naively) am wondering if a) I need to for town use, and b) if I can perhaps do better.

On a) I've installed a higher powered LED into an inexpensive 'be seen' dynamo LED light and revised the circuitry to suit. This is fitted with a rear light that illuminates at a walking pace. This is already far better than the old lamps, and probably better than a 3W system driving tungsten bulbs. Since there is no data out there I may try to measure the drag -and therefore infer the efficiency- of the hub itself.

but b) doing better though? My logic is this; a typical white LED presently has a forward voltage of 3.0 to 3.5V. Most generators to the '3W' standard produce 0.5A into a 12 ohm load. Some generators and some lights produce/consume more current when driving a lower impedance load, but not all. A typical result might be that the generator consumes 5-8W of power, and produces about 3W. At the generator, the voltage output might be 5-6V, 0.5A. At the LED, (say) 3.3V 0.5A, i.e. 1.65W, so about half the power is lost somewhere in the rectifier/ regulator. Overall the system is only about 20% efficient in terms of delivering power to the LED device. Even a very efficient system still might struggle to do better than 40% overall.

Thus I harbour a hope that a really efficient '1.5W' system might consume 2W and produce nearly as much light as a present 3W system, if it can be made well enough.

cheers

[RJC]

I believe Busch & Muller's single LED lights like the Cyo have a current doubler/voltage halver switch mode converter
so it will take in say 7V @ 0.5A to drive the LED with 3.5V @ 1A.

[robc02]

I gave this a try by installing B&M LED lamps front and rear to see if it works. It does. There is a caveat: the LEDs are not as bright as they are on my other bike which are fully-driven by a 3 Watt SON dynohub. But they are brighter at all times than the original filament lamps, so the upgrade was effective. Stand-lights are a bonus here too.

+1 (D'lumotec / Seculite driven by FG hub - slightly lower output than GH6)

The fact that the lights are brighter when driven by a modern hub (Shimano in this case) suggests this system is at or slightly beyond its limit - still a good economy option, though. However, if good quality lights are developed for the 1.5W Shimano hub they could be an even better option. I suppose the quality/performance will just depend how much effort manufacturers' put in to what might be seen as a budget system.