rhm

the equipment:
Headlight: B&M Lumotec IQ Cyo Premium Senso Plus
Dynamo hub: Shimano Alfine DH-S501
Taillight: not sure, I think it's a Spanninga Pixeo

the setup:
The wires are connected with soldered-on 'banana' connectors pretty well encased in heat shrink tubing with dielectric grease. Where possible, the wires run under the fenders, attached with tubular tire cement and pieces of gorilla glue.

the problem:
The lights stop working for no obvious reason. Sometimes they work, sometimes they don't, and even when I get them working again, I don't know what I did to make them work.

Typical situation: before leaving the house, I pick up the front of the bike and spin the wheel, and both lights work fine. But when I get on the bike, the light starts flickering and then goes out. When I stop and pick up the front wheel and spin it, now the light sometimes works and sometimes it doesn't. I check the connection at the hub (pull it off and push it on again), and now the light works... or maybe it doesn't. Pull apart the banana connectors and put them back together again, repeat the test... and the light works, or maybe it doesn't. If I repeat these steps, sooner or later the light works again, and I get on the bike again, and all is good. Or maybe the light starts to flicker and then it goes out again. After it goes out completely, it may still give a bright flash of light every time I go over a bump, but remain dark when the road is smooth.

This has happened several times over the last month. Several times I've missed my train after losing time futzing with the light. Something similar happened on a couple long rides I did in April; in both cases I got the light working eventually and it worked fine all through the night. But I never did diagnose the problem. After the first such occasion, I replaced the wire and the banana connectors, and thought I had solved the problem; but it recurred.

This morning, after stopping to play with the wires three or four times, on a whim, I tightened the quick release on the front hub. That is, I released the QR (it was not loose), tightened the nut a quarter turn or so, and locked the QR lever down again. I made it really tight this time, so it was hard to push the lever into position.

And this seemed to solve the problem. The light worked all the way to the station, though I had already missed my train.

So, my question is:
Does it make any sense that tightening the QR should solve the problem? What's going on here? What do I need to do, to make this setup reliable?

zacster

It is possible that tightening the QR fixed a loose connection internal to the hub, but it is also possible it was just luck. Have you tried jiggling the wires at the light? It is more likely loose or broken there. It could be internal to the light itself too. What you need is an independent source, such as a borrowed wheel, and also using your wheel with a borrowed light. I have a similar setup with a Shimano 3N72 hub and the same light, and would be willing to do that if you'll come to Brooklyn since I see you are local. Is it a 700c or a 26" wheel? Mine is a 26 mtb wheel, but it may not matter.

Check your wiring for shorts too, including the taillight. I have a DVM that can check.

Reply here if interested and we can go on PM for more details.

rhm

Thanks for that offer! It is indeed a 26" wheel (the MTB size rim, with a Compass RTP tire), so your offer would work; but though I'm in NYC every day, the bike doesn't get closer than the Hamilton, NJ, NJTransit station.

Yes, I've tried jiggling the wires at the light, as well as just about everywhere I can think of. Checking for shorts is really difficult if the problem is intermittent.

Trying the bike with a different wheel is a good idea. I have a 650b wheel that I can pop in there; the brake won't work, but if I can diagnose the problem, then I'll be on my way to fixing it.

steelbikeguy

My rule of thumb is that wiring and connections cause more problems than anything else, although that rule was developed on aircraft and earth moving equipment. Photos of the connections might help us evaluate them.

A handy troubleshooting method is to sequentially replace each part of the system with a known good part. i.e. swap out the dynamo for a known good dynamo and see if that fixes it. If not, swap out the headlight. If that doesn't do it, swap out the wiring. This will almost always identify the faulty part. Sometimes just removing and reinstalling the part fixes the problem... but possibly only temporarily.

A meter can help too. Without spare parts to swap in and out, this might be the easiest method to find a bad part, but is probably not well suited to finding an intermittent problem.

Sorry I can't offer more help...


Steve in Peoria

noglider 

I agree that wiring is mostly to be the problem in general, though your case, the QR thing might mean otherwise. I would replace all the wiring without troubleshooting the existing wiring. I assume you're using thin wires, so there could be a chafed section that offers high resistance, and jiggling that could change things. The chafed section won't be visible, because even intact wires are very thin.

Trevtassie

If tightening the QR did actually fix it, it may be a broken connection in the hub, the Shimano hubs have the "earth" connected to the bike chassis. So there may actually be a connection being made because a QR can measurably compress an axle (which is why you set wheel bearings with the QR done up). Alternately , it could be the "positive" (hah AC, but you know what I mean) contacting the chassis, and the QR pressure uncontacts it. Best way to test would be with a multi meter, during a fault time. The earth terminal should be connected to the axle with zero resistance, the other terminal should have some resistance between the other terminal and the axle.

rhm

Right, that's what I've been thinking (emphasis on your first word). What worries me, in this scenario, is: the lighting system, both headlight and taillight, "should be" electrically isolated from the frame/fork. The tighter QR may have made a better electrical connection between the axle and the fork, but if the lights are truly isolated, as they should be, this wouldn't matter... right?

By chance, USPS brought me a new hub today. So I'm busily building a new wheel that may perhaps solve the problem, assuming there's a problem in the hub.

More later. Thanks, all, for your suggestions so far!

unterhausen

Failing to read the whole thread, I have had similar problems and fixed them by redoing the wiring into the connector.

Trevtassie

Yeah, the lights are above earth in a wiring sense. You've replaced the wire and whatever connectors, so that leaves the hub. If you have a multi meter you could check for continuity between the terminals and between the terminals and the hub body. It could either be an open circuit or a active wire thats grounded and ungrounds when the QR is tightened.

znomit

Try flipping the wires to the hub. You might have issues with the earth being wrong.

unterhausen

traditionally, lights have not been isolated from the frame. The hub is grounded to the frame, that I know. Headlights nowadays seem to be isolated, but taillights often are not. The Pixeo is not. In fact, some versions come without the ground lug.

rhm

I'm familiar with the old lights that had one wire, using the frame / ground as the other wire. That's pretty straightforward with incandescent lights, but becomes more troublesome when you use LED's, especially when it's not clear whether a given headlight or taillight has an internal bridge rectifier. Most modern lights seem to have an internal bridge rectifier, even if they have clearly labeled contacts (positive and ground, for example).

But as I mentioned, I've now built a new wheel, with a Shutter Precision hub, which came with a new plug. So far, no problems. Fingers crossed.

Had I known that hub was coming so soon, I probably would have held off on starting this thread. It wasn't supposed to get here until June!

Tourist in MSN

The SP is not grounded to the frame like the Shimano hubs are. I learned that the hard way, I started with an SP hub on one bike. I assumed all hubs were not grounded to the frame. Then I got a Shimano hub and put it on another bike, and wired that bike the same as my bike with the SP. After a few hours of total confusion, I realized that the grounding was my trouble.

You already figured out that the Shimano and SP plugs are interchangeable, I mention that in case other readers do not know that.

rhm

Followup... since building the new wheel, the problem has not recurred, unless you count a bad dream I had last night (slept poorly, stressing about tomorrow's 600k randonnee).

noglider 

Good luck, @rhm! I still can't fathom how one can ride such distances continuously, but I'm also intrigued and might try it one day.