Re: That Lowe's flashlight, again

Peter Cole wrote:
SMS wrote:
Peter Cole wrote:
SMS wrote:

A bridge rectifier is only 81.2% efficient (maximum with zero-voltage drop diodes)

You keep repeating this, but it's not true. You misunderstand (and misquote) a reference on rectification.

I think where you're confused is that you're looking at the AC input power and the full wave rectified power, but not the DC power. Indeed, with perfect diodes, you'd have 100% efficiency, but you won't have DC yet! When you convert to DC, using capacitors to filter out the ripple, you are at 81.2% (assuming perfect diodes and capacitors).

You are mixing up voltage and power. "Rectification efficiency" (the source of your 81.2%) is a measure of the output (DC) to input (AC) voltage. It has nothing to do with power losses/efficiencies.

After you rectify AC, you do have DC ("time-varying" DC). "AC" means "alternating current", which means the current is reversing direction. This can not happen (by definition) with a diode in the circuit.

depends on voltage and frequency. exceed reverse breakdown voltage and
you get? increase frequency and you get? zeners work by?



The principle of superposition states that a time-varying DC signal can be treated equivalently to a constant component plus an AC component, but that principle only applies to "lumped, linear, time-invariant" circuits (resistors, capacitors, inductors). Diodes are non-linear.

they are non-linear, but that has nothing to do with the function you're
trying to describe.




If you're able to run the LED directly from the the rectified AC power from the dynamo, without over-voltaging it, then you could get by
(put two LEDs in series, as you want to keep them under 4.5 volts).

Again, LEDs (and diodes in general) are (approximately) constant voltage devices. The ideal diode has a discontinuity in its resistance (slope of V/I), changing from infinite to zero when it passes through the forward voltage threshold. It is this discontinuity that makes them non-linear devices. You can't "over-voltage" a diode, LED or otherwise, you can only "over-current" it.

The only reason to stack 2 or more LEDs in a generator driven application is to improve the efficiency slightly by matching the generator load more closely to its designed load (incandescent) and making the real rectification power (loss) a smaller fraction of the load power.

An LED operating at 3.5V and 1.0A will be dissipating 3.5W.

direct current.

If the waveforms are time-varying DC (like FW rectified sinusoid), all those numbers are RMS (root mean squared).

the average numbers are rms, not peak. peak is relevant here, as with any switched mode device.


If the rectifier is a Schottky bridge, the voltage drop will be approximately 0.6V, the power 0.6W.

schottky junction losses are more varied than that - typically 0.15 - 0.45V.


The non-linearity of the diodes makes computing the actual power tricky, the shapes of the voltage and current waveforms are complex, but since both the rectifiers and the load are diodes, an efficiency calculation based on the ratio of the forward drops would be very close.

bullshitter. simply scope and integrate.




In reality, the almost all LED lights run off pure DC. Many flashlights use DC-DC converters between the LED and the power source so they can run off of a variety of different voltages.

The classical way to analyze a source/load circuit is to model both as circuits made from ideal components (equivalent circuits). The equivalent circuit for a generator is an ideal (sinusoidal, variable frequency, variable amplitude) voltage source with a series resistance (winding resistance) and inductance. The LED (and rectifier) equivalent circuit must be separately modeled before and after the transition (infinite resistance/zero resistance). The net is, that for most of the time (cycle), the generator is driving an equivalent 0 ohm load (short circuit), the current being limited by the series resistance and inductance. The inductive component being significant in that it limits the *rate* which the current can increase during the cycle, which limits its peak value. Since the voltage (internal ideal) and frequency both rise together with shaft speed, the net is somewhat self-regulating.

dude, where do you get this ***?



The reason incandescent bulbs may burn out at high generator speeds is that they have a bad dynamic resistance characteristic.

rubbish.


Their effective resistance goes up with power.

no, it goes up with filament temperature. there's a lesser inductive
effect for coiled filament, but that can be ignored for low frequency
applications like bike generators.


If you think of the series inductance and resistance as "ballast", the ratio of the ballast to the load changes as the bulb heats up, putting (relatively) more power to the load, increasing the efficiency (load to ballast power ratio). This is similar to solid state "thermal runaway".

that's bull***. increased voltage means a higher filament temperature.
higher temperature means creep and rupture. it's extremely
non-linear. "thermal runaway"? resistance /increases/ with
temperature, not decreases - wrong slope on the graph.



Batteries have relatively low internal resistance, so the equivalent circuit is an ideal (constant) voltage source in series with a small resistance. When driving a short circuit load (diode after transition) the currents can become large. For this reason, battery-driven LED circuits must be current limited.

no, they must be current limited to prevent led destruction.
destruction is a function of current density and temperature.


Additionally, it is often desirable (to optimize battery life) to operate at multiple, selectable power levels. For LED loads, this means variable current regulation. Boost converters may also be needed if the battery voltage is below the LED operating voltage.

more bull***. most leds are "regulated" by way of a switched driver.
this is primarily to limit current consumption and thus extend battery
life. it has a side-benefit of extending led service life also. hence
people just use drivers regardless of input.



Suitably paired LEDs and generators need no regulation,

eh????


either for current limiting or voltage boosting. Current limiters could be used to lessen the generator mechanical load selectably, but I don't really see the advantage.

that's because you don't know what you're talking about.


.