3 - 10 MHz coil drive

[Hero999]

Yes, I get surprisingly good inductive coupling - on the bench I've had 90% coupling efficiency, driving the primary coil with a signal generator at 3 MHz.

The inductance of the coil I'm using was calculated using an online tool, not measured - maybe I put a number in wrong and it should be 100 nH, but it is a very tiny coil and 10 nH seemed intuitively correct to me. 3 MHz really is the optimal frequency (well, ok, it's really closer to 2.9 MHz) as tested with an RF sig gen and oscilloscope on the secondary coil.

100nH for a self resonant frequency of 3MHz still sounds too low. I would think it's either resonating at a harmonic or the inductance is much higher.

An inductance of 10µH with a parasitic capacitance of 280pF would give a self resonant frequency of 3MHz. If the inductance is 100nH to resonate at 3MHz the parasitic capacitance would need to be 28nF which sounds far too high for a 100nH coil.

If it's resonating at the 10th harmonic (30MHz) then the parasitic capacitance would need to be 280pF which is more plausible but I don't think that's the case because there won't be much power at the 10th harmonic.

I think your inductance calculations are incorrect unless you've added another capacitor to the secondary.

 

[Hero999]

No but an LC resonator can be excited by harmonics in a square waveform.

 

[Roff]

No but an LC resonator can be excited by harmonics in a square waveform.

Wow! You are fast! I left that post up for probably less than a minute. I deleted it when I figured out what you meant (before you posted your response).

 

[kellbengal]

You guys are certainly much more knowledgeable about this than I am. My design plan was simple:

1. maximise the number of turns of the coil in the space available
2. reduce the impedance on the secondary side
3. monkey around with the sig gen until I get maximum transfer.

I've gotten reasonable results thus far, but now I wonder if perhaps I'm only occupying a local maximum for the possible power transfer. This is a battery powered application, so I'm always keen to reduce waste power.

Unfortunately, I'm still waiting for all my circuit stuff to be shipped from Australia, so I can't test out the modified circuit just yet.

I'm still wondering if there is a better way of implementing the coil drive, though - I'm dangerously low on the pcb real-estate as it is.

 

[Hero999]

1. More turns in the same space means higher ESR so you won't gain anything efficiency wise. The only good thing about using more turns is the resonant frequency will be lower which means lower switching losses. You need to strike a balance between switching losses and resistive losses.

2. The impedance on the the primary side will be reflected to the secondary side.

Your best bet is to tune both the primary and secondary to the same resonant frequency, this should give you the best transfer at longer distances.

 

[kellbengal]

I can't do anything about the secondary side (because of the nature of that circuit) but I presume I'd tune the resonant frequency via some sort of variable capacitor?

 

[Hero999]

That would work.

I've done a similar thing to this before.

I used two ferrite AM radio aerials.

I wound the same number of turns round each, and added a fixed precision 1% tollerance ceramic capacitor and trimmer capacitor in parallel with both of them. I drove the primary using a single MOSFET connected to a signal generator and tuned both coils to get them both in resonance.

I could power an LED form about a meter away and a small neon lamp from a couple of cm away.

 

[kellbengal]

Sounds sensible. I'm not suffering for performance at the moment - the target circuit lights up four LEDs a mm or two away (which is all I need). I suspect there is a bit of capacitance in the secondary circuit simply because of long tracks threaded through the four LEDs in parallel.

The big challenge remains how to make the drive circuit small enough - it's supposed to fit onto something the size of an Aussie 50 cent piece... about the size of a wrist watch.

 

[Hero999]

Where are you going to get the power from?

A button cell isn't large enough to do this for any reasonable length of time.

 

[kellbengal]

Very small LiPoly - I have a great source of tiny ones.

 

[Hero999]

Are you planning to charge the battery via the coil as well?

 

[kellbengal]

Nope - the li poly cells gets charged via a plain ol' dc connection (probably USB with a fancy chip to control the current limit)