Induction Charger Design

bulb,

A 2W bulb in this sort of small enclosure will get friggin' hot. To the point where it'll discolor the PVC, make a smell and so on.

Have you ever touched a 20W CFL after it ran for 1/2 hour? Plus the size is huge.

LEDs are an option but for 2W continuous output, meaning with a reasonable MTBF, it gets much more expensive than ye olde inductive charger. I clearly would try to tackle this case inductively. Maybe even going to 6.78MHz ISM. Might have to be 13.56MHz because AFAIK 6.78MHz isn't legit in some areas.

No, ours is bigger :-)

Two solutions:

a. Time-gate the charge oscillator. Requires collision detection and the firmware guys might throw a coffee mug at you, so maybe not a great method. Unless you buy them a lot of beers :-)

b. Use 13.56MHz ISM for radio and 27.12MHz ISM for power so minor harmonics won't interfere with the radio. 27.12MHz isn't that great for radio anyhow in many places because of CB radio being in the same band. Then there is 40.68MHz ISM but I don't know if that is as widely allowed throughout the world as the other two.

Of course, if you use ISM in the UHF range for radio you'll be so far away from the shortwave ISM bands that interference should be more manageable.

If you are using 35 gauge wire, have you worked out what the resistive voltage drop in the wire is at 500 or 100 turns ?. At 200 to 400 mA, you are probably losing all the volts in the wire resistance.

I would start with a 20 to 50 turn secondary of 26swg as a starting point, with frequency of 25 to 100Khz and adjust the primary, resonated, to suit...

Regards,

Chris

On a sunny day (Sat, 14 Nov 2009 08:56:34 -0800) it happened Joerg wrote in :

bulb,

Well, I think the receiving end, so the photocell needs to be in a small enclosure, not the adaptor that it fits in, or in this case fits next to, perhaps.

Well, not sure, but a lot better then the glass of a 20W filament bulb.

I would like to avoid any radio frequency use, couple of kHz already worries me. Did I ever tell you how I blew up some nice RF transistors by just having those in a *metal* (cigar) box next to the tank coil of my 500W linear (4 Mz or so)? I never figured out how it could have damaged those transistors, but they were killed.

There are things like pace-makers, anything that could tune to your transmitter, and maybe cause damage, or kill somebody. It is that MIT wireless power idea. That other link mentioned here, working at several hundred MHz, seems unhealthy too.

Light is safe to some level, and you will not have damage claims. The little pot core idea also works very well, but the halves need to be close.

The grass is greener, wait, no I cannot explain that one. We sure could use some here, been raining now for days.

bulb,

You can easily get 30% quantum efficiency by shining an IR LED on a photodiode. Maybe 10% in energy terms.

Cheers

Phil Hobbs

The latest results... and I hope this NG allows an attached picture, cause there is one included with this post, a schematic of my driver and a snapshot of the 2 iron cores.

I can resonate the primary, and I can resonate the secondary, but as soon as I attach the load the primary resonance goes to hell in a hand basket. Once the load is attached to the secondary I can use most any value of capacitor across the primary (actually 2 caps, one across each half of the primary) with very very little to zero effect.

The thing that does make a difference is resonating the secondary. I started with a .0047u across the secondary and got an immediate boost in output voltage, at around 50KHz. As I lowered the frequency I got a peak V around 30KHz. I think the peak V was around 2v. I tried .01 and tuned again. Higher peak, at around 20KHz. Kept going to a higher cap and retuning... I'm now at .1u across the secondary with peak V of 5v @2.8MA at 12KHz. (I'd kinda like to stay above 20KHz because one of these things will some day have a loose winding and the customer is going to complain about the buzzing mosquito sound.)

I feel that the air gap between the cores can be as little as .06", the thickness of the pendant case plus the thickness of the cavity walls in the base. One of those cores would be mounted below the base of the cavity and the other of course would be in the pendant at one end.

I'm still not liking this very much. Finding a core like you see in the picture won't be easy, and probably not cheap.

This was supposed to be easy, but I'm finding it quite a challenge.

bulb,

enclosure,

I assume the OP wants a charger base that doesn't double as paper weight :-)

Sure, but still too hot IMHO. CFLs do not fare too well unless properly vented, not so suited for charger bases.

me.

killed.

transmitter,

If a pacemaker gets upset by such fields the designers of the thing should have their heads examined. Or better yet the FDA should put a padlock on the factory doors and order a major recall. Just imagine what would happen if a dude with such an inferior pacemaker would walk past a car with a shortwave ham radio transmitter in there?

unhealthy too.

I do not believe in such schemes. For small power levels, yes, but not to supply major loads.

Yep, pot core halves do work well in such applications. You can easily bridge several millimeters, enough for even the toughest ABS enclosures. The Philips Sonicare doesn't even use a pot core, probably it would have cost 1/2 cent too much ;-)

Ask your politicos to deliver on their promise that some global warming will materialize :-))

If you want 25mW net to come out at the rechargeable that's still a fairly big LED. I think the magnetic solution can be cheaper, provided you find a low-cost factory to wind the coils. We had to do that for a project lately and did a few prototype coils using a fishing rope spooler that had a built-in counter. It's amazing how fast that goes. The coils look like factory-built, all perfectly similar.

I am curious as to why a primary lithium battery would not work in your application? Is this a high duty cycle application? Reason I ask is that US Part 15 transmitters have a very tiny duty cycle. A lithium primary battery might be better use of your fractional cubic inch.

It doesn't allow attachments :-(

Maybe you can post it to a company web site area or free picture server.

The trick is to series-resonate (not parallel) the primary side _and_ have the oscillator adapt to changes in series resonance. Those will happen when the gap moves plus there will be lots of Al-value tolerance in the cores, easily 30%. In a nutshell a free running power oscillator that uses series resonance. There are also chips for this but that gets more expensive.

Can't see the picture but ferrite cores aren't expensive (in production qties). If you need small quantities to experiment with you can buy a nice collection here:

They also have rods and all sorts of other shapes. These are small qty prices, once you know what you need and are ready to place bulk orders it's going to be a whole lot cheaper and there are tons of other suppliers, US as well as foreign.

I sort of like those projects, things that are considered voodoo by many other engineers.

Warning: Be very careful when machining ferrite in any sort of manner. It is brittle and razor-sharp stuff can fly off at high speed.

Wingsy wrote: one end.

Is the core iron or ferrite ?. At a few 10's of Khz, a ferrite core would be better idea. From people like:

Both of whom do a very wide range of ferrite components.

The critical thing will be the air gap between the cores in relation to the core transfer area. Someone can correct me here, but the distance / transfer characteristic is not linear, possibly inverse square law or something like that. A better way might be to wind the receive loop around the inside sides of the box, ie, maximise tha flat area at both sides and thus avoiding a small transfer point. Perhaps use a flat ferrite plate between the two coils.

I looked at the schematic and a better way as someone else said, would be to use a complementary driver, emitter follower, then use this to drive a series resistor and cap to the tx coil and ditch the centre tapped configuration. As a series resonant circuit, you can adjust the cap value to get a higher or lower voltage in the coil, more or less sinusoidal. The series resistor lowers the effective q and thus makes it less sensitive to the load conditions, though you need to optimise it all with the two coils adjacent.

Oh yes, remember to use the thickest wire that will fit in the box, consistent with getting enough turns in place...

Regards,

Chris

I haven't met with any success in resonating the primary. I can do it if the secondary is unloaded but as soon as I attach the load, I can substitute a very wide range of caps (even none) across the primary with no effect.

Resonating the secondary has helped, quite a bit. I'm now up to 5v @

In most cases it probably would, but one requirement for the pendant is that it must periodically respond to a ping from the base. If it ever fails to respond the base issues an alert signaling that the pendant has wandered off. When that feature is enabled the battery will drain over a period of a couple of weeks (or more I hope). When not enabled the pendant will still consume some small amount of power, as it must wake periodically and perform a battery test. (Hey that kinda sounds silly ... you drain the battery just so you can test it! Anyway, battery testing is in the requirements and if I can help it, I don't want to make an argument for taking away stuff they think they need.)

Sounds like you have a transponder, not just a transmitter, so this is not an "I have fallen and can't get up" device and more of a "grandpa has wandered off again" product! The only downside I see is that the user may put it in the charger and then "forget" to wear it again.

If one MOSFET is "On", and the other is "Off" then there is +9V from the center tap to one side of the winding (the one with the "On" MOSFET.

Ignoring the diodes, what will the voltage be on the "Off" MOSFET drain? Considering the diodes, what's gonna happen?

Best regards, Spehro Pefhany

Snip

Would not some type of capacitive interface work for a small application like this? The device could have two halves of pair of capacitors molded into its case and the charger could have the other halves molded into a depression where the pendant rests. Depending upon charging current requirements, there might still need to be a switcher to provide HF AC source, but perhaps the parts count could be reduced in a low current application such as this.

"Some Wog Wanker "

The device needed is a "photovoltaic cell " or "solar cell".

..... Phil

It's both. For the ones who may wander off there is an aid handy to swap out their pendant with the one in the charger whenever a low battery is detected.

If my math is right, each of the pair of capacitors would be around

Dimension of the plates = .25"x.5", separation = .06"