Inductive coil circuit LED power

[edy]

Hi everyone,

I saw this instructable about building a circuit to induce a loop current to power an LED:

http://www.instructables.com/id/Easy-Wireless-LEDs/step3/Soldering-the-circuit/

I modeled it up in LTSpice and I'm not getting it. The bridge rectifier models fine, but I don't get a nice smooth DC source unless I use a smoothing cap. That aside, the author of the instructable says it can be powered by a battery.

The circuit is supposed to produce a high frequency current in the loop but my LTspice model is just not doing it. I tried using other npn transistors and nothing is producing any higher frequency than the AC input. I will try using a pure DC source as well but somehow I feel the circuit is not higher frequency, just increasing the rush current due to charge buildup and quick release through the transistor.

[T3sl4co1l]

Circuit?

[edy]

I've attached the circuit and the rest of the build is in the linked instructable. Hope this helps. I can upload my LTSPICE file as well.

[richard.cs]

Have you put a k statement into LTSpice modelling the coupling between the coils? Do that and tick the "start external d.c. supplies at zero" box and you can probably kick it into life.

[edy]

Thanks for the help. I did check off start circuit from 0 but did not use any k statements so that could be it... beginners trap, as Dave would say. I have to read up on it and will let you know if it worked and upload my LTspice file.

Just as a side note... I am not even interested in coupling any inductors... I am just looking at the frequency oscillating part of the circuit, not the coil with the LED. Should I be adding the LED-inductor loop as well? Is it affecting whether the oscillator works or not? Won't it oscillate regardless of the separate LED-loop element?

[edy]

I've attached the LTspice file. It seems odd, sometimes as I change the smoothing cap values it works other times it gives me open node errors. Can't figure out if I even need it or not. Either way, I can't get the oscillator to work.

Just rename file to ".asc" from ".txt" extension (I had to rename for it to upload).

[IanB]

It's not clear to me why that circuit should oscillate until I think about it some more. It looks like it should be a resonant tank circuit formed from the inductor and the capacitor. Possibly the RF emitter coil is not needed for oscillation and could be replaced by a low value resistor (e.g. 1 ohm)? Since the transistor is switching at radio frequencies it should a model suitable for RF use.

[edy]

I've tried a range of values on my LTSpice sim but cannot get any oscillation. The way I understand it the transistor drains the cap and it charges up quickly again, once it reaches a threshold it drains it again quickly, and so on. I've tried various values which affect the curves but not getting any higher than my AC input ripple.

[IanB]

I've also tried in LTspice and I can't get the circuit to oscillate either. I don't see where the feedback occurs that would be required for sustained oscillation.

[T3sl4co1l]

It's an emitter feedback blocking oscillator.

[IanB]

It's an emitter feedback blocking oscillator.

I was thinking along those lines, but I can't see how this specific circuit actually works. Can you elaborate on how the oscillations take place?

[Zero999]

That site should be called destrictables, since most of the circuits are poorly designed or don't work.

[T3sl4co1l]

It's like this, but rearranged slightly.



The k between sections around the transformer should be near 1 (maybe 0.9-0.99), and the k between those and the LED thing depends on distance, anywhere from, say, 0 (no coupling) to +/- 0.8 (coil right on top of the other, in phase or out of phase).

Note that the LED need not light in a given direction, because the peak voltages generated by the coil need not be equal.  The flux is what averages to zero (and thus zero DC voltage).

Tim

[IanB]

It's like this, but rearranged slightly.

That's not rearranged slightly, that's a different circuit!

What you've shown is like a classic Joule Thief blocking oscillator.

The published circuit is like this:



However, I think the published circuit is a fake (you've been Rickrolled!):

http://youtu.be/ythX0oG_9N8

http://www.instructables.com/id/Wireless-L-E-D/

[T3sl4co1l]

It's like this, but rearranged slightly.

That's not rearranged slightly, that's a different circuit!

Well, no.  Move the winding from supply to base, to supply to emitter.  Flip the phase, so B-E voltage is the same, and adjust turns ratio to account for the difference in C-E voltage.

(Grid-cathode coupled blocking oscillators were very common in tube TV deflection oscillators, probably because it's easy to injection-lock that topology.)

A circuit may look different while behaving identically.  Any circuit transformation that results in the same equations (or close enough given the operating conditions) has this property.  The simplest transformation is simply shifting ground, which immediately leads to a variety of oscillator circuits, for example (a few of which are named).  Swapping series elements (e.g., supplies and reactive components -- see above) leads to topologically different but functionally identical circuits, as does splitting or combining components (Colpitts has a split capacitor and single inductor, others use a single cap and tapped inductors or multiple windings).

Tim

[IanB]

Well, no.  Move the winding from supply to base, to supply to emitter.  Flip the phase, so B-E voltage is the same, and adjust turns ratio to account for the difference in C-E voltage.

Most importantly, your circuit has a feedback path through the coupled inductors that you have marked as primary and secondary.

The circuit under consideration here does not have such a pair of coupled windings. The inductor L1 is a standard packaged part and there is no indication that it is coupled to the antenna coil L2, nor that it should be connected with any particular polarity. The circuit description implies that that the oscillator consists of C1, L1, Q1 and R1.

Could that circuit oscillate, and if so, how?

[T3sl4co1l]

It's close enough that it'll be more like k = 0.1, or it may've been tested inside the coil (don't you just love it when people leave details out?).

It may also be weak enough feedback that it oscillates, rather than behaving as a blocking oscillator (or squegging).

Tim