Hub generators can turn draggy....

[Brucey]

I have used a somewhat beaten up DH-3N71 on my hack bike for the last six years or so. It has had a special cone fitted (to replace the original one, that broke) and it has done a lot of miles and seen a lot of weather. The weather caused the hub to go draggy once before, because water got in and the hub basically filled itself with rusty powder. The rusty powder 'set' into a solid-ish lump,(ironically whilst the hub was dry-stored indoors not being used, instead of being left out in the weather like normal) and prompted a strip-down that included repairing bent laminations and re-potting some parts of the assembly. That was two or three years (and five or seven thousand miles) ago.

The replacement cone is older than that and was an experiment that I was concerned would not work, since I had to grind a new profile on a cone meant for a different hub, that risked ending up with a cone that was the wrong shape or that lacked enough hardness. Well, It did better than I had feared; the hub bearings (over 12 to 15 thousand miles) have required adjustment about once every three thousand miles, and there have been occasional noises too, all of which suggest that the hub bearings are not entirely happy. However more recently the bearings have developed some slack more quickly than before, suggesting that the end is nigh.

I've had a bad winter and have not been out on my bike (barring utility journeys) anything like as much as normal; however last week I went on one of my few proper rides this year and I felt like it was very hard work, like I was creeping along, worse than I'd expected. I thought my tyres were soft, I thought my transmission had turned draggy, I thought the wind had turned mid-ride, I thought my brakes were dragging: I thought all kinds of things. I suspected the hub generator, too; I picked the bike up and span the front wheel; it seemed 'normal' apart from a bit too much free play in the bearings, and span easily. I got to the end of the ride and I felt done-in, and I hadn't gone at anything like my usual pace. I blamed my appalling state of fitness and wondered if I could have managed my bike-fitness better over the winter despite several things seeming to be against me.

Today I went for another ride, the same route on the same bike only this time I fitted a different front wheel, with no generator. By comparison with the same ride a few days ago it felt like a tailwind going both ways. I was still tired at the end of it but I had gone significantly faster under basically similar conditions, on the same bike; 'why?' I wondered....

Further examination of the hub generator showed me that maybe it had been a good part of the problem. Unlike a lot of other hubs the generator goes in the same way up each time, so that the connector faces upwards. This causes the cones to wear on one side only. In fact to even out the wear on the bearings, I'd turned the RH cone ~120 degrees a couple of years ago, so that the (potentially soft) RH cone wouldn't wear unduly on one side only.

However more recently the hub bearing has been wearing, faster and faster; when I tested the hub for drag the other day, the wheel was supported by the axle, and the balls ran on the upper (less worn) parts of the cones. The hub span freely. However when the loading was reversed, so that the balls ran on the (normally) lower sides of the cones, the hub became more draggy. In fact when displacing the axle as if there were a load of ~20kg on the hub, the axle became noticeably stiff to turn. I suspect that there is more rusty powder inside the hub, and that the stator has been rubbing quite hard against the magnet, causing the hub to become very draggy when in use.

If this rubbing turns out to have happened as I suspect, I am at once slightly relieved and I also feel like a bit of twit; I know perfectly well that the clearances inside the hub are tiny and that the slightest fault can cause the stator to rub against the magnet, yet I appear to have let it happen anyway....

I intend to strip the hub in the next few days to see what the true state of affairs is; watch this space for details.

cheers

[mercalia]

glad I have a bottle

[NATURAL ANKLING]

Hi,
Ah, now I don't feel such a twit when coming home yesterday and it felt like a flat back tyre..........its far worse if you find there is full compliment of air in tyre and you get back on blaming the time of week.

Huh, so I get my bike for its quick once over I do every ride so no surprises or flats in the early hours.........the rear is stiff.........very stiff.
I put it to fiddling with the clearance on the bearings recently.........which messes with the disc pad clearance..........ease off the pad and still stiff Oh check wheel is central and seated in dropouts......chest on rack.......clunk.......did not over do the QR as bearing clearance was close last time I mounted the wheel

Must tighten rear QR more, I normally go for 80 degrees and was more like 25

While I am at it there's brown sludge in the top A head race too.......all done now.

[Brucey]

glad I have a bottle

FWIW in the same time period as this hub generator has been 'on the go' (and BTW I obtained it already broken, as a bit of a project...) I have destroyed several bottle dynamos on other bikes that see a small fraction of the mileage.

cheers

[Gattonero]

glad I have a bottle

Which is inefficient, won't work well when it rains, will wear out the tyres, and won't allow you to fit some nice supple tyres. I'm glad they're a thing from the past!

[NATURAL ANKLING]

Hi,
Blimey......................no handbags....

[cycle tramp]

Hi,
Blimey......................no handbags....

Yes, i'm just glad the cons and pros of whatever lighting system didn't overtake this thread... its always good to read of repairs to any bicycle equipment. Please keep it up!
(For the record i do ride with bottle dynamos, and accept both their short comings and benefits, as i respect the choice of those who use dynamo hubs. Bicycle stands would be very boring if we all rode the same sort of bike)

[Brucey]

well the hub is in bits on my workbench and I am currently in chin scratching mode, wondering what to do. I think the immediate choice is obvious; bung in a spare wheel (I am lucky enough to have one). But to repair the old one it is not obvious what the best route is.

As expected the bearings were worn, and the generator had been rubbing, but not in quite the way expected. The electrical parts were not pretty but were no worse than I have seen them before.

The modified cone was worn as I had anticipated; the hard layer was almost completely ground off during modification and it has been spalling off (presumably due to subsurface fatigue) where the 14 x 5/32" balls have run. However the funny noises I had heard probably came mostly from the LH bearing which uses all standard parts; amazingly two of the 11 x 3/16" balls had shattered leaving only nine and a load of shrapnel. To my considerable surprise the original cone (the breakage of which is common) is still usable despite this and the cup insert in the hub (which fits into a ~26.95mm dia recess) is starting to break up.

So options are

1) give up (but it is not in my nature...)
2) machine hubshell/bearings etc to accept cartridge bearings
3) repair/renew as per previous modification
4) build with different cup and cone bearings

2) requires that two M11 threaded cones are ground up to make shouldered supports for cartridge bearings. Options here include 27mm OD, 13mm bore bearing, which can be ground to fit the hubshell.

3) is unlikely to be satisfactory unless a different M11 cone is used somehow. The LH bearing insert isn't a big problem; the back of it is accessible, so it can be knocked out and replaced with one from a standard shimano front hub (the dimensions appear to be the same).

4) could work with (again) different M11 threaded cones. It happens that the LH cone from some Alfine hub models is M11 threaded and it is slightly smaller dia (on the bearing surface) than the cone I used before (which came from another shimano hub dynamo which uses a bigger cup).

The alfine M11 cone is the wrong OD (so the seals will need revision) and the length is wrong too (which is more easily corrected). Measurements suggest that I can grind the bearing surface it by ~0.25mm in which case it will work with 5/32" bearings, or perhaps it will work 'as is' if I use 9/64" or 3.5mm dia bearings.

Choices, choices.... I shall mull on it.

Anyway, even with knackered (and loose, the rim would move over 1.5mm...) bearings etc the wheel span really freely in the workstand with no load (the working balls are the ones above the cone in this case, running on less worn parts of the cone) which shows me that the shimano 'sport' generators can have very low parasitic losses; once spun up to ~10mph the wheel took a minute or two to slow to a halt which means the average losses were a tiny fraction of a watt, not ~2W as reported by some.

cheers

[Samuel D]

Anyway, even with knackered (and loose, the rim would move over 1.5mm...) bearings etc the wheel span really freely in the workstand with no load (the working balls are the ones above the cone in this case, running on less worn parts of the cone) which shows me that the shimano 'sport' generators can have very low parasitic losses; once spun up to ~10mph the wheel took a minute or two to slow to a halt which means the average losses were a tiny fraction of a watt, not ~2W as reported by some.

Aye. I think testers sometimes measure the drag of bearings set with an overly high preload. (But the reported ~2 W was for a higher speed than ~10 MPH anyway.)

By the way, on a new DH-T780 hub I saw recently, the grease was no longer Shimano’s fluorescent green stuff but white. I wonder if that means anything.

[Brucey]

10mph is about 5m/s and the rim plus tyre (plus half the weight of the spokes) weighs ~ 1kg tops. Thus the kinetic energy of the whole shooting match is about 12.5J at that speed. Drag/parasitic losses are commonly reported to be about linear with speed so the average loss is about pro-rata with the average speed. So if (say) the wheel halves in speed in ~45 seconds then you have lost about 9J which is about 0.2 W average (at average speed of 7.5mph), which suggests that the drag at ~20mph will be around 0.5 or 0.6W. Of course this is an overestimate of the electrical parasitic drag, because in this test there are windage losses and bearing losses that will occur anyway. Given the present state of the wheel, in my case one would expect the latter to be quite a bit more than normal....

I think that with new hubs the shimano seals drag quite badly until they are run in or lubricated with a better lube (often the seal lips are bone dry). The preload in shimano DHs is usually enormous too. I have also (by simply greasing and adjusting a new DH-3N72) apparently reduced the parasitic drag to about half of what it was beforehand, and it will only get better over time, I'd say.

I've not seen it in a DH but they have been using a white grease in Nexus and Alfine IGHs for some years. Maybe it is the same stuff. If so, it is better than the green stuff in some respects , but still nowhere near good enough to resist any kind of wet/winter use. It dries out (separates) worse than the green stuff, too, or at least leaves more sticky crud behind....

cheers

[Brucey]

well the autopsy is complete and this is what I found in the RHS of the hub.

conked out cone

As I'd originally feared, I'd ground too much off the hard layer on the DIY cone and this is the result. The remaining hard layer wasn't really thick enough (so could have fatigued subsurface) and was probably full of stresses too. You can see that the hard layer had also been chipping off the edge of the cone too. This wear (which didn't extend around the full circumference, thus threw the wheel off-centre) was enough to cause the generator to rub inside.

Mind you it took over 12000 miles to get this way, so it hasn't done badly.

To use 5/32" balls inside cups that size the nose of the cone needs to be ~14.8mm or a bit less. After some digging around I found an M11 threaded Nexus/Alfine cone that was 15.1mm on the nose.

As an experiment I have simply fitted new balls to the LHS (despite the fact that the cup and the cone are both pretty worn) and fitted that M11 cone with ~3.5mm ball bearings. Because I didn't have any 3.5mm ball bearings to hand, I salvaged some from a used HT-II bottom bracket bearing. I was amused to discover that there are more 3.5mm ball bearings in my improvised arrangement than are found inside a single HT-II BB bearing, so I think it should be strong enough....


Anyway if this arrangement works I shall either leave it be or rebuild it using better quality parts; new cups, better quality balls, seals that fit properly, that kind of thing; as it stands it is below par in about half a dozen different respects. The worst part is probably the LHS cup, which has a fairly pronounced damaged area.

cheers