Lasersaber strikes again. A joule thief king ?

[TinselKoala]

TK and All:

Have a look at the LM3909:

https://www.youtube.com/watch?v=87xMbWM8qz0

Just about everything said in that clip was my experience also.  I built my first LM3909 circuit in 1977.

MileHigh

Heh... think of what 99 cents would buy you in 1977... it's equivalent to about 5 dollars today!

[TinselKoala]

If you look at the data sheet of the LM3909 (is attached) you find on the first page:

Low current drain, averages under 0.5 mA during battery life.

The 0.5 mA power drain is about what I achieved with my circuit, see:
http://www.overunity.com/14591/lasesaber-strikes-again-a-joule-thief-king/msg402222/#msg402222

The road to very low power draw:

- 100 Hz flash rate
- very short pulse
- low LED current (rather dim LED)

In my circuit the pulse is already very short but happens at about 5 KHz, the LED current is already low (dim LED). By having more windings (thinner wire) on my core the frequency can be lowered. How can the pulse be made shorter?

Remarks:

The LM3909 uses a peak LED current of 45 mA at 1 Hz (slow) or 1 KHz (fast) to have a bright flash.

The supply current of a NE555 is between 5 and 15 mA (5V to 15V), therefore it is of little help.

LaserSaber's special core and the 2N1304 transistor seem to operate at less than 10 µA (at about 12 V) on average.

Greetings, Conrad

You can choose your power draw based on how bright you want your LEDs to be. If all you are concerned about is getting a glimmer of light, where you can see the little wires inside the LED.... then you can make your power draw very low, as we have seen, and if you don't use dissipative elements like resistors in your circuit you can make this tiny glimmer for a long time.

But what if you want enough light to actually see something, or read a label? Show me a JT that operates on less than 10 uA at _one volt_ input not 12 ... and is bright enough to read by ... please.

[lasersaber]

But what if you want enough light to actually see something, or read a label? Show me a JT that operates on less than 10 uA at _one volt_ input not 12 ... and is bright enough to read by ... please.

I am working on it.  It might be possible.  When I reach best possible performance I plan one making a super low current draw flashlight.


I did have significant progress since the Mother's Day circuit: http://youtu.be/6B79UJGoNJE  That circuit was using around 25uA.  Using a different germanium transistor and a bridge rectifier I was able to reduce the current to 15uA while maintaining, if not increasing, the brightness.  I have since burned out all my germanium transistors doing crazy tests.  I have never had one burn out under normal operation.  It usually happens while trying to recapture energy with a rectifier.


Here is a link to my latest video showing DC vs SJR testing: http://youtu.be/5EkXNRWAi1Y

[conradelektro]

You can choose your power draw based on how bright you want your LEDs to be. If all you are concerned about is getting a glimmer of light, where you can see the little wires inside the LED.... then you can make your power draw very low, as we have seen, and if you don't use dissipative elements like resistors in your circuit you can make this tiny glimmer for a long time.

But what if you want enough light to actually see something, or read a label? Show me a JT that operates on less than 10 uA at _one volt_ input not 12 ... and is bright enough to read by ... please.

The whole exercise (trying to light a LED with an electrolytic capacitor for many minutes) is of course a little silly. There is the hope of finding some "energy gain" in a high impedance core used in a certain way. This hope is of course also silly, but why not?

I think that Steorn tried that and failed and many others looked for some "energy gain" in the brake down of magnetism in a core. The idea is that one gets more electricity back when the magnetic field brakes down than one used when magnetising the core. But nobody ever could show that conclusively.

Do not look for any use, look for the fun in trying something strange! And do not take it too seriously.

About my experiments: I could not put a copper strip between core and coil former because the core has to be insulated (otherwise it would shorten the copper strip). Insulation plus copper strip are too thick. It has to wait till I get the 2N1304 because I want to try it with the coil without copper strips first. I have the suspicion that the copper strips are of little help.

Greetings, Conrad

[conradelektro]

I am working on it.  It might be possible.  When I reach best possible performance I plan one making a super low current draw flashlight.

@Lasersaber: did you ever use a coil without the copper strips in this circuit (with your special pot core)?

I saw that my electrolytic capacitors (one 4700 µF and 50 V, the other 4700 µF and 25 V) have a leakage current of about 30 µA. This means that I can never go below a 30 µA power draw when using these caps.

Greetings, Conrad