bulbs to LEDs

On 13/08/13 15:04, TimC wrote:
On 2013-08-13, James (aka Bruce)
was almost, but not quite, entirely unlike tea:
On 12/08/13 16:43, Trent W. Buck wrote:
James writes:

On 06/08/13 18:41, Trent W. Buck wrote:
I suspect the responses are all gonna be "I hate dynamos! Batteries are
not that hard, just HTFU" so I haven't bothered to dig up hub specs &c,
but I'll do so if there's interest. The bulb claims to be a Philips
HPR60, 3.6V 2.4W.

Interesting that the bulb is rated for such low voltage.

What is the rating of the hub dynamo you have?

Most dynamos and the required bulb are rated something like 6V, 3W.

Now you've got me freaked out, because I tried a "HPR53 4V0.85A JAPAN"
in it, and it didn't light up *at all*. (A replacement bulb finally
arrived from LBS, and that's working, hooray.)

Ah, well that's good. Perhaps the light housing has a voltage limiting
zener diode or similar, to protect the 3.6V bulb from the full 6V. Or
perhaps the original lighting circuit had the front and rear bulbs wired
in series, so that the 6V was shared across two bulbs. I can't tell
from here.

Dynamos are current sources (not voltage sources). LEDs are current
sinks (not voltage sinks).

That's one way to look at it. Dynamos have sort of regulated current
output due to the inductive impedance increasing proportionally with
speed (X = wL).

And yes, the I-V curve of LEDs has a fairly flat voltage independent of
current, once a threshold of current is reached.

Your 3W LED will typically have about 3 or less volts across it, and
if you wired it up to a non-current limited 12v power supply, it would
blow.

Absolutely. However I have 4 LEDs, each with about 3.5V across at up to
1A, configured as a full bridge rectifier and connected directly across
my 6V 3W dynamo. (Two LEDs in series, in parallel with two LEDs in
series but connected to conduct in the opposite direction. And yes the
reverse polarity breakdown threshold of each is about 5V, well above 3.5V.)

If you hooked up an appropriate resistor in series with your 3W
LED, and then the un-current-limited 12V power supply, then you'd end
up with a voltage divider and the appropriate current flowing through
the LED that caused it to have a ~3V voltage drop (with 9V and most of
the power being dissipated by the resistor). But in the ideal case of
having a current source, then the LED will just sink as much current
as the source supplies. Both will match their voltage if you've got
no resistor or anything else in the circuit. But of course, you don't
presumably know the actual current limit of the dynamo - at least I've
never seen it documented before.

I measured the output voltage and current from my dynamo at home at
various speeds with a fixed resistor load, and recorded and graphed the
result. It gave me a good understanding of the output characteristics.
There is also a set of test results for various dynamos he
http://www.myra-simon.com/bike/dynotest.html. It's also instructive
to consider the electrical model as an AC voltage source proportional to
speed, and a resistive and inductive element. Note that the inductive
reactance increases linearly with speed like the internally generated
voltage. This is what limits the current.

But given they are ~3W, it should be
more or less right. The LED won't work perfectly at 2.9W and blow
catastrophically at 3.1W.

LEDs of different types *have* to be wired in series (the current
source will just double its voltage output in order to maintain the
same current going to 2 equal LEDs in series).

Only if the LEDs can both handle the same current.

You can get away with
very similar lights being in parallel, until they age. Then one will
start taking more current, until the power supply starts overpowering
it, and there'll be thermal runaway and that LED will completely short
out, rendering you with a Dark Emitting Diode.


Well, my home made headlight has been in regular use (3 times a week)
for the past year. I've descended some pretty big hills as well. Zero
defects thus far.

--
JS