Anderas wrote:I am sure how LED behave when they get too much volts.
Bigger batteries are a solution only if there is a small resistor in the circuit.
When they get too much, they give you less light for the same amperage, forever. If you continue to give too much, they turn off completely after few minutes or seconds, depending how much too much you give them.
On the other hand, when you give them less than they need, they work just fine and they even get more efficient. But of course they give you a little less light then.
Luckily LEDs are less prone to burning out than normal diodes, the light emitting feature itself works like a (small) resistor.
I appreciate you are trying to simplify so I get the "like a (small) resistor" comment where in reality you cannot design the circuit treating an LED like a resistor. So we are AGREED on the above points.
However, we also have the known situation where he has more lights hooked-up and they are dim and when he hooks up fewer lights it looks good. He also has a lighted HQ board that is working fine with a similar circuit. If we assume the LEDs are not already permanently damaged (performance degraded), the question about LED behavior is:
Q: Is the light emitted purely a function of the current? Or if the voltage is wrong (too high), the efficiency for the same amount of current is compromised (temporarily)? I can imagine the possibility that the P-N junction leaks current and is rendered far less efficient by a large voltage and therefore a proper amount of amps actually results in less light (temporarily, not necessarily permanently).
If the above behavior is possible with 2-5x over-voltage conditions then adding a resistor will gain efficiency and result in the current battery powering more lights / the same lights for longer.
On the other hand, an LED output could be independent of voltage (and a function of current only) for 1-5x over-voltages and we can reasonably conclude from the dim LEDs that a resistor will not solve anything. So he can safely add a second battery in parallel to the first. In a final step, he can add resistance in the final configuration until the lights dim very slightly, then back off, to optimize battery life.
Like I said, I do not know much about LEDs.
It might be helpful if we could propose a simple non-destructive test with a second circuit using LEDS which have not been exposed to the raw voltage drop to show the proper efficiency which can be achieved? Like some number of LEDs in series where you remove one until they light up... Then you know the approximate proper voltage drop as battery voltage / number of LEDs.
Alternatively, providing the specs for the LEDs and battery would allow Anderas to quickly design the circuit.
Has resigned from the forum and would delete his account if he could.