Mark J. wrote:
On 4/2/2021 11:59 AM, Ralph Barone wrote:
sms wrote:
On 4/2/2021 9:00 AM, jbeattie wrote:

snip

I don't love it. I would love a dyno light with a solid 800 lumen
output, a little more upward spew and a stand light that was stronger
than the light on my give-away key chain from Wells Fargo -- and one
with a battery so I could use the light off the bike. We
transportational cyclists often need a light for use off the bike. A
flasher would be nice for dusk and dawn, but not required. And while
we're wishing, how about something lighter and more efficient than a
bunch of magnets whirling around. There must be some other way of harvesting electrons.

Such a light would be wonderful but it would be a stretch with a 6V/3W
dynamo, even at higher speeds where you can get more than 3 watts out of
it. Some LED makers are claiming 300 lumens per watt, at least in the
lab, but 200-250 lumens per watt are what is available commercially at
this time.

A 12V/6W hub dynamo (or even a 9V/4.5W hub dynamo) would make dynamo
lights with sufficient intensity more practical, including a beam
pattern where some upward spew would be possible. DRL flash capability
is trivial to add, as are internal batteries to be able to use it off
the bicycle. But there is just not much of a market for any of this.


You seem to have it stuck in your head that the internal impedance of a hub
generator is some immutable quantity and not a design parameter. Why not a
6V/12W hub dynamo? Hell, if you were willing to do frequency dependent
series capacitor switching, you can get a lot more than 3W out of a 3W
labelled hub.

And yet virtually all the commonly available bike dynamos come out with
a half amp nominal design. (So that wattage is half of voltage). I'm
told their coils "saturate" (or something like that) at a half amp.

I've only been using dynamos over a span of 50 years. I've owned 6V/3W
and 12V/6W generators, and once I saw an 8V/4W claim. If there was a
big improvement to be had with some other amperage design, I would think
someone would have tried to market it by now. Dynamo design might have
some complications, but surely they are well understood at this date.

Mark J.

Pretty much all dynamo design dates back to a time when electronics didn’t
exist. Since light bulbs don’t put out much light at low voltage and tend
to die at high voltage (I seem to recall bulb life being inversely
proportional to the fourth power of voltage), it was important to try and
find a way to produce a constant voltage from a variable speed generator.
The solution that was found was to build the generator with a large leakage
inductance. This inductance provides more and more resistance to current
flow as the frequency (ie: speed) goes up. Since the voltage also goes up
with speed, the two effects cancel each other and you get a reasonably
constant output into a constant resistance load. This became its own
standard (with a 6V output because that was the voltage of three lead acid
cells or four carbon cells).

It is the adherence to this 6 V 500 mA standard which is inhibiting
progress. With modern electronics, it is possible to produce a regulated
output voltage from a variable input voltage. If you have this, there is no
longer the requirement to have so much leakage inductance to regulate the
voltage. The output of the generator also doesn’t have to be 6 V AC, and
if we look at modern LED lights and lithium ion batteries, 6V doesn’t make
sense anymore. 5V actually makes more sense in terms of LED voltage drop,
lithium ion battery voltage and USB powered accessories.

What would make sense to me would be a generator based on modern brushless
DC e-bike motor design, with a maximum output somewhere in the 25 W range,
but normally limited to around half of that (because how much drag do you
want from your dyno hub?). Generate at a slightly higher voltage (for more
efficient rectification, but not so high that the wire in the rotor is too
fragile), then regulate it down to 5V DC for distribution. Add a
controller that figures out how much to ask the generator to produce and
intelligently divides up that power between the battery and the connected
loads and you have a nifty scheme. Unfortunately, the market is to small to
repay the poor shmuck who ends up re-engineering the entire ecosystem.
Maybe somebody with deep pockets (Shimano?) can lay out the standard, then
others can join in.