Sturmey Archer GH-6 dynohub magnet upgrade?

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
(hopefully not too off topic...)

The last couple of days I've been getting some performance data from my GH-6 Sturmey Archer dynohub (which I've become unhealthily obsessed with lately). I've got it set up in a motor-driven testing rig (more on that later...). I think I've sorted through just about everything that's been written about the GH6 online, including some insights from the (legendary) Jobst Brandt. To summarize, the GH6 can put out about 1.8W, supposedly reaches its rated output at about 20 kph/12 mph and has a saturation current of around 300 mA (6V @ 0.3A = 1.8W). One often discussed issue with the GH6 is the 20 pole Alnico ring magnet which can lose what little magnetism it has if the armature is removed from its centre without a keeper. It has also been said that this magnet can become weaker over time. My hub is a 1966, so an aging magnet could definitely be messing with power output. My testing suggests that my particular hub isn't saturating until 30+ kph.

This preamble brings me to my question: it seems like machining a simple adapter would allow the ring magnet to be replaced with a series of more powerful rare earth block magnets. This is kind of what I have in mind: http://www.etotheipiplusone.net/pics/pq2/pq2_54.jpg

Would replacing the old Alnico magnet with modern rare earth magnets in this configuration result in a significant increase in power output? Any obvious pitfalls I'm missing?
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
seems like a good idea.

The only issue that I can think of is the gap between magnets and the rotor. How close is the gap in the original design, and how close will it be with the new magnets?? The magnets will determine the max potential flux that you can put into the rotor, and the gap will be the primary factor that degrades that flux level.

Steve K.
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
The gap is pretty small. By awkwardly jamming my calipers in there, I'd guess it is probably 0.020". I'm thinking of upping that to 0.030" or so, just to be a little more forgiving. The fact that the new magnets are much stronger will hopefully make up for that. Thus far I figure I can fit 22 3/4" x 3/8" x 1/8" block magnets in an adapter: http://www.kjmagnetics.com/proddetail.asp?prod=BC62. The magnets have tolerances of 0.004" in every dimension.

Got something rendered now. Gonna have to become friendly with a machinist...

I also wonder what limitations are inherent to the armature design.
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
OK, looks like I'm not the only one who had this idea: http://fieldlines.com/board/index.php?action=printpage;topic=130083.0



L
ooks like an upgrade will get it saturating at about the same level of modern hub dynamos (500-600 mA), but the reported voltage increase is good news for my application. 3 mm gap seems large. By 'cogging' I think he's referring to the notchy behavior of dynamo hubs. I expect the stronger magnets are going to make the 'notichiness' considerably more noticeable.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
cool! A 3mm air gap, eh? That's huge! I suspect that it'll limit the open circuit voltage, since it's proportional to the change of flux over time. Not that most folks care at all about open circuit voltage, though.

The issue of core losses is a valid one, and I hadn't thought much about it. If the armature isn't built with thin laminations and the right material, you might just increase the rolling resistance by putting the extra flux through it.

At this point, though, I'm sitting here trying to figure out why the photo shows a handful of spoke holes that are clearly larger than the rest. What's up with that??

Steve K.
 

panicmechanic

Newly Enlightened
Joined
Nov 23, 2009
Messages
130
Location
Germany
Probably to attach the repeller hub, seems like a wind generator project. That might explain why he's so concerned about cogging, you can't have that if you want to harvest every small breeze in an area without steady winds.
Cheers,
Martin
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
a repeller hub? Is that to repel the crows that eat all of the corn?? :)

A propeller hub would make sense, though...
The subject of windmills/wind turbines is an intriguing one. There are a couple of large wind farms nearby, and I'm always fascinated watching them. Even with a modest wind at ground level (5mph or so), the turbine blades are turning. Of course, they are 100 feet or so above ground, so they get much better wind.

As much as I like the idea of a small wind turbine for home use, the poor winds at ground level really reduce their effectiveness. Perhaps the idea is more practical in a chronically windy area?

If the concern is cogging, then this would be a good application for a bit of electronics... the goal would be to only present a very light load to the dynamo at low speeds. Perhaps a "maximum power point tracker" would work? Probably not practical for a 1.8 watt dynamo, though. A simpler circuit that would just sense a speed threshold before hooking the dynamo up to the load... might be more appropriate?

regards,
Steve K.
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
Well, I think I'm going to have a go. Taking advice from Ben@gotwind, the magnets need to be the same in number and spacing as the original in order to be compatible with the existing stator, which makes machining a little more complicated. Here's my proposed design. The rotor adapter will be made of steel so that the magnets will stick to it, although I'll probably need to add a drop of epoxy. The adapter is designed to allow a slightly larger gap. The gap between stator and orignal magnet is only 10 to 20 mil. This design will be somewhere between 40 and 60 mil, depending on actual tolerances.

magnet%2520holder%2520diagram.jpg

magnet%2520holder%25203D.jpg


Now I need to get friendly with a machinist...
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
Isn't it sufficient to just do what he did? i.e. get some steel bar of the correct thickness and mount the magnets to it? Or in this case, just get a ring of the proper O.D. and I.D. and mount the magnets to it? I don't understand the need for the semicircular slots in the ring. Mostly, I have a hard time with the idea of throwing much money at a Dynohub....

Steve K.
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
I'm under the impression that the method Ben used would prevent the magnet cover plate from fitting in. The semicircular notches are there to make spacing easier. I don't want to throw much money at this (the magnets alone are $20 or so), and am hoping a friend of my father's can do the machining as a favour.
 

Bandgap

Enlightened
Joined
Aug 14, 2002
Messages
284
Location
London
How about NOT* going for the strongest magnets you can find.

In my simple understanding, any flux beyond saturation in the claw poles adds to cogging without adding to output power.

Also, really, should you be spending $20 on magnets for a heavy outdated steel design when you don't need to spend much more for a modern Shimano generator that will weight much less.

Don't get me wrong, I am a great fan of Dynohubs and own more than one, but I stopped experimenting with them once SON dynamo hubs became available.

Although....
I could understand replacing the magnets in a classic bike with a Dynohub if the Alnico magnet has been ruined.

If you have a Dynohub with a good magnet, then even with its 1.8W output it should give lots of light with an LED headlight.

Expecially if you run it at a nominal 12V (four leds in series, at its simplest), where it will be a 3.6W generator.

Steve

*Thanks Marty ;-)
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
How about NOT* going for the strongest magnets you can find.

In my simple understanding, any flux beyond saturation in the claw poles adds to cogging without adding to output power.

Noted. I was thinking 1/8" thick N42 magnets, but 1/16" might be enough. I don't have any experience with Neodymium magnets, but understand they are almost an order of magnitude stronger than Alnico. Finding a magnet that isn't too strong might be the actual challenge here.

Also, really, should you be spending $20 on magnets for a heavy outdated steel design when you don't need to spend much more for a modern Shimano generator that will weight much less.

I kind of hate to admit this, but I just counted and realize that I own 7 contemporary hub dynamos (2 have yet to be built into wheels) plus 3 old Sturmey GH6s. Obviously, this isn't about being practical, sensible, sane, etc. etc.... ;)

Don't get me wrong, I am a great fan of Dynohubs and own more than one, but I stopped experimenting with them once SON dynamo hubs became available.

Although....
I could understand replacing the magnets in a classic bike with a Dynohub if the Alnico magnet has been ruined.

If you have a Dynohub with a good magnet, then even with its 1.8W output it should give lots of light with an LED headlight.

Classic bike upgrade is my main aim. I've got a couple of projects in mind for classic Raleighs.
 

BrianMc

Enlightened
Joined
Nov 4, 2009
Messages
940
Classic bike upgrade is my main aim. I've got a couple of projects in mind for classic Raleighs.

My guess was right, and combining modern functionality in a classic design in a restore-upgrade is much in line with my 1980 Mercian as a 3 x 10 with modern sealed hubs, deep rims, and cable routings. :thumbsup:


If all you need the scallops for is to place the magnets accurately, and if weight is not an issue, and if the cost of the scallops is high, then, could you might consider simply making a stiff paper template to mark the ID with permanent marker X ing out the interspaces?


BrianMc
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
Had an adapter machined out of 3/4" steel. Used J-B Weld to attach 1/4 x 3/4 x 1/8" Neodymium block magnets. Took a while to fit the thing into the hub shell; needed an extra spacer ring to stop the stator from rubbing. It's definitely more notchy than the stock magnet (and far more than modern hub dynamos) but in my test rig its saturating at 800-900 mA at around 30-40 kph. Even at speeds below 20 kph its putting out more than 400 mA. No idea what the drag would feel like. I'll have to put it on a bike and take it for a ride. Wonder now if I could get away with using 1/16" magnets which might put out less power but would hopefully have less no-load drag.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
wow! a turbo-charged Dynohub! As fun as it might be to stick with the existing arrangement and get the extra power, I suppose there's no point in getting any extra current that you don't need. Might be a good match for a LED headlight built from XP-G's, though.

It would also be interesting to know if the notchiness translates into actual drag once the wheel is on the bike. My experience with SONs is that they feel very notchy as a hub, and somewhat notchy when handling the wheel, but I don't notice it at all when it is on the bike.

Good to hear that progress marches on!

Steve K.
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
Just checked the notchiness of a cheap Novatech hub dynamo. It cogs more than I remember! The Dynohub with the new magnets is definitely harder to turn, but not by much. Also, I think the cogging effect is exaggerated by the fact there are only 20 poles compared to the Novatech's (presumably) 28 poles. As you say, Steve, what feels notchy when turning the hub in your hand doesn't necessarily translate to huge amounts of drag on the wheel.

Running the upgraded magnet with Cree XPGs is exactly what I have in mind!

Still would like to see if 1/16" magnets will have a similar power output with less drag. The 1/8" might be overkill (and they're more expensive). I'll have to see if I can convince my machinist acquaintance to make an adapter that will take 1/16" magnets...
 

minisystem

Newly Enlightened
Joined
Jul 12, 2011
Messages
96
Here is a more detailed account of how the magnet upgrade seems to be working. I've plotted current of a red and white LED in series against speed. The series Vf of the two LEDs is about 5V so you can infer power from it but I didn't actually measure Vf, which goes up modestly as current rises. The data was acquired on a motorized testing jig. I'm using an Arduino to log the current at different speeds. Speed is calculated using zero-cross detection to measure the frequency of the pulses coming from the hub. My jig has crappy speed control so all I can do is ramp it up and slow it down and catch the current at each speed. At each speed I collect four data points and plot the average. Error bars are standard deviation of the mean.
GH6 magnet upgrade data.jpg


I should mention that the whole contraption includes a buck converter running at 100% duty cycle, so the losses inherent to that should be taken into account (ie. the curve would be slightly different if I just connected two series LEDs directly to the rectified output of the hub).
 
Last edited:
Top