Engineers want an idea to play with?
 

Dynohub, with battery/capacitor integrated (at the hub, the weight shouldn't be an issue) for maybe 5 minutes of 5V regulated USB power, and a line coming off with some ability to tap into it for parallel USB ports wherever they're needed.

Essentially, I want something that will charge some USB-rechargeable accessories, but not have the annoyance of them kicking out of charge mode when I stop at a light or sign and then back into it when I start off again. At least for my phone, that could be really obnoxious (losing the "screen always on when charging" for example - interestingly, the phone only seems to care that it's getting 5V; current determines how fast it will charge if at all, but 5V at <100mA keeps it in the charging mode so the screen stays on) and with all the GPS-based apps, it pretty much eliminates the need for a cyclocomputer if I could just keep the battery from dying so fast. Standardizing on USB means it can charge any of dozens of headlight/taillight options that are almost always cheaper than (or at least comparable to) direct dyno lights. (And having them use their own batteries lets them deal with any high current or odd voltage requirements internally.) Charging (ideally with passthrough capability so they're charging - or at least offsetting some discharge - during use) rather than driving them directly means you could charge someone else's lights for them, or use your headlight as a flashlight off the bike, charge a helmet mount light etc.

More fun; add an alternate power/storage option so that a larger USB power pack could be added to the system, sinking or sourcing as conditions dictate, and/or "shore power" to charge everything up from an external supply when parked.

Don't reinvent the wheel.

There are many studies out there of this exact endeavor.

The most notable one that I've read indicates that properly set up, you're looking at losing between 40 and 65% of your power. I realize you only want 100 milliamps, but you need to charge a battery to store that while the battery charges the accessories. Then you've got power loss in the generator itself, the converter, the voltage regulator. It just seems like something that's not feasible. Not to mention the weight of the battery.

I can assure you that if your idea worked as simply as it seems, there would be some little puny 100 mA finger crank that people could just charge their phone with while they watch TV. And that's just not the case.

Are you touring? If not, charging devices from a dynohub is kind of dumb. It's a lot of fuss and bother to save 4 cents a year in power to charge a $15 power bank that will charge your phone.

Also I'm not sure that it's correct that dropping from full charge to 100ma mode will not disrupt the charge cycle. phones typically negotiate or sense the charge current and expect to be able to keep drawing that.

I sure wouldn't want to use USB as a power distribution network on the bike if that's what you're proposing. USB connectors aren't waterproof or very durable in the scheme of these things.

If a little finger crank could charge the phone, I'd be able to run my house with a dynohub. Heck, I could arc weld with it on the last descent of my morning commute.

The ideal would be to add in some control to allow dialing the output (and thus resistance) up and down; max it out for regenerative braking while bombing down a hill with a stop sign at the bottom, but minimize it for climbs or sprints, while manually finding a happy medium in the flats. The battery could level that out a bit for the devices, especially if using a larger pack outside the hub, (and that's probably the only way you'd get practical use out of 1-2 minute shots of regen braking power) but I still like the idea of having a couple minutes of bare minimum power storage integrated into the hub so that feature (not having the devices switch out of charging mode at a stoplight) is always available.

Not at the moment, but it's a primary reason I'd want such a feature set. I've also put off some day trips and overnighters because my phone dies after a couple hours of running nav/comp apps. Those might be solved with a reasonably priced <2lb power bank, but nothing beyond that.

What the internals do with it is not the concern as long as it resumes actual charging once the current comes back up. AFAICT, it does that. What I want, and what my current one appears to do with low input current, is for it to leave the screen on, nav app up, and otherwise act like it's charging whether it is or not.

Well, most of the original developers are still alive, so torturing and then drawing and quartering them for their idiocy and as an example to others is an option I'd still like to explore, but it won't fix the existing standard. Perhaps the solution would be for the actual fixed sockets to be something that is waterproof and durable, then make various cables for it, including all the male USB (A, B, mini/micro A/B) for direct connection to devices and a female A pigtail for anything that requires a proprietary cable. (Thus allowing a waterproof cable to be made with the proprietary cable, the female A pigtail and some silicone tape.)

I took a nine-hour one-day ride last summer and used a one-pound external battery. I used it for charging a few things, one at a time. I had a USB-chargeable headlight, a tablet, and a smartphone. It was more than enough. I think the rating on the battery was 20,000 mAh.

The dynamo-powered B&M USB-Werk and Tout Terrain Plug 3 both have small built-in cache batteries for that reason.

The USB-Werk is about $55 ordered direct from Germany at current exchange rates without VAT which US residents don't pay. bike-discount.de has $22 flat-rate shipping to the US. Combine the two and you'll still save $50+ over what a US distributor will charge you.

I run a USB-Werk with my SON28 hub to keep my Edge 800 charged on long rides. The combination isn't enough to keep my Samsung Galaxy S5 charged except in Ultra Low Power mode which precludes running apps - outside that it bounces in and out of charge mode and still continues to loose battery even with the screen off.

You can also combine a B&M E-Werk with their external cache battery. That combination will be much more efficient than a USB charged USB battery.

I wouldn't want that integrated with my hub. Keeping the electronics separate allows running a hub with a good track record - people are only getting 10,000 miles out of some of the new hubs which is completely unacceptable when that's you're annual mileage. It allows separate replacement if something goes wrong.

I have a Shimano dynamo hub + in-steerer forumslader. I have no problem recharging my S5. It takes a little over an hour to fully charge the S5 (I mean S5 battery starting St close to zero) and an additional 45 minutes to restore full charge of the cache battery.

The forumslader generates 6W @ 20kmh.

(10k on a hub?!)

Isn't it what happens naturally with a generator? Almost no output at low speed, up to some fairly high output at high-speed.

A number of people have reported 10,000 miles (16k km of the factory quoted 20k km) as their Shutter Precision bearing life at which point it's time to remove all the spokes because they require disassembly for bearing replacement.

SON28 Neu hubs need disassembly for replacement too, although Schmidt claims 50k km (30k miles) and there are reports of more.

Is this a rainy Sunday job requiring $10 of parts or are we talking LBS?

Maybe with a Shimano hub, but the others, no. Generator hubs are not for the faint-of-heart to open, as Sturmey-Archer of old and Shimano are the only ones that do not require any real specialized tools to open. For the others, opening the hub as a non-authorized repair center voids your warranty.

The existing standard is entirely fit for purpose. Taking a product off the shelf, deciding that you want to use it for an entirely different purpose than it was designed for then whining about it not being fit for your new purpose is a bit off the mark.

Let's be honest, USB kind of sucks for its intended purpose too. Connectors that only fit on the third try (after you've flipped them over twice), micro-USB connectors that wear out quickly, unpatchable security issues, etc. Granted, it sucks less than the assortment of connectors (serial, parallel, PS/2, MIDI, etc.) that it replaced, but it's still not as good as it could be.

Micro USB is a huge step up from Mini at least. They're tiny because people demand they be tiny (perhaps indirectly by demanding razor thin phones.

Mini is only rated something like 1000 to 5000 insertion cycles. Micro is rated at 10,000. The tiny size does make them prone to damage by clumsy users, but there's not much to be done about that. They made the right decision by making the male end flimsy rather than making it sturdy so the connector inside the device might break instead. Easier to replace the cable than the socket. Also they put the retention springs on the cable in micro rather than in the socket like the mini, so when they wear out, you just buy a new cable, not a new phone.

I really don't have any trouble inserting USB cables. Hell of a lot less than I do most other connector types anyway.

Yeah; tell that to all the failed USB cables I've tossed out and the laptop I'm about to give up on because two of the three ports are worn out and require jiggling any device plugged into them until it's recognized.

OTOH, it has become the default standard for pretty much everything that can be charged with 5VDC. That means if you want to charge normal consumer devices, it's what you're going to be using.