Trying to wirelessly power an LED.

[md5fungi]

I'm part of a Senior Design project involving Wireless Energy Transfer. We are trying to power a standard red LED as a demo, via resonant inductive coupling. Our setup is as follows:

We have a 40.64 cm diameter transmitter coil (magnet wire) with 6 turns, hooked up to a signal generator. Our receiver coil has a 2 cm diameter, with 21 turns. We measured the respective resonant frequencies of these two coils, and used 522 pF of equivalent capacitance in parallel with our load (LED) to tune the receiver's resonant frequency to match the transmitter's.

We placed the receiver coil inside the transmitter coil on a table with our signal generator at the resonant frequency. The result was a very very dimly lit LED, and it became brighter when we moved the receiver coil outwards, i.e closer to the transmitter coil. When we look at the output waveform on an oscilloscope, we get ~ 5 V peak-to-peak, and when we rotate the receiver coil so that it is perpendicular to the transmitter coil, the amplitude drops quite a bit, which seems to indicate that resonant inductive coupling is occurring.

The thing is, we're not quite sure why the LED is barely lit. Adjusting the capacitance on our receiver coil does not seem to help, and changing the frequency only makes it dimmer. We are guessing that very very little current is going through our LED, which is why 5 V does not suffice to turn it on. We have tested the LED by hooking it up directly to the signal generator, and it lights up quite brightly.

Any ideas why we are getting such poor efficiency? Even when the receiver coil is very close to the transmitter the LED barely lights up...

Note: We got the LED to bright up very lightly when untwisted the twisted pair of alligator clips we used to connect the receiver coil to the LED. This would seem to indicate that the alligator clips were forming a loop and coupling with our other two loops, however, and that is not what we want.

This is our setup. We've tried this with and without that resistor, we just did that because we thought it might increase the Q:

 

[Bob S]

The thing is, we're not quite sure why the LED is barely lit. Adjusting the capacitance on our receiver coil does not seem to help, and changing the frequency only makes it dimmer. We are guessing that very very little current is going through our LED, which is why 5 V does not suffice to turn it on. We have tested the LED by hooking it up directly to the signal generator, and it lights up quite brightly.

Any ideas why we are getting such poor efficiency? Even when the receiver coil is very close to the transmitter the LED barely lights up.[/PLAIN]

Two possibilities.
1) You have a voltage divider, because the small coil is intercepting only a small fraction of the magnetic flux from the large coil.
2) You have a lot of leakage inductance from poor coupling.

See Smythe, "Static and Dynamic Electricity", 3rd edition, page 335; coupling of circular loops.
Bob S

 

[waht]

From what I understand, the main coil is connected directly to the signal generator via a twisted pair transmission line (crocodile clips).

For max power transfer, you need to do some impedance matching - the impedance of the signal generator, to the transmission line, and then to the coil loop where the energy is to be dumped, otherwise some of the signal is reflected from the coil back to the generator and thus wasting power.Chances are that the pick up coil is highly untuned in your setup, the breadboard adds capacitance, and you have it going between the crocodile clips. Try putting the capacitor banks together with the coil, and add a small trimmer cap to tune with screw driver. Make sure that the tank is resonating at the desired frequency by checking it using a sweep generator or a network analyzer if there is one lying around somewhere.Also, air core coils don't have a high Q, increase the Q by adding a ferromagnetic core.

 

[md5fungi]

Thanks for your responses. I will try your suggestions and see what I get.

 

[seb7]

Try a SEC exiter.

 

[famousken]

In the photo, I see you have a single led and the impedance matching capacitors. The problem I see is that your transmitting and receiving coils are both AC devices and an led is a DC device. The led will present a load only to one half of the voltage swing, which may lead to saturation of your coil, I would try placing another led in anti-parallel with the other led.

 

[Phrak]

Also, air core coils don't have a high Q, increase the Q by adding a ferromagnetic core.

Watt, wouldn't the addition of core material also increase the coupling? Wireless energy transmission is beyond my kin so I can't comment on this problem.

Would powdered iron core material also be an option?

 

[iulian28ti]

Hmmm... more turns and a core were the first things which came to my mind.

But... if you're using a high frequency, that core can become a problem.
As Phrak said, perhaps a powedered core will do the trick.

 

[Bob S]

two things:

1)At higher frequencies, ferrite cores become lossy, and reduce the Q of the circuit. if you use a ferrite core, if possible choose a frequency under ~ 1 MHz.

2) The impedance of the circuit at resonance is sqrt (L/C), so choose a lower inductance and higher capacitance. This will increase the resonant current in the coil. (Remember that the impedance of a simple antenna is near 72 ohms.) Using sqrt(L/C) over 300 ohms is too high.

Bob S

 

[waht]

Watt, wouldn't the addition of core material also increase the coupling?

not much, the magnetic core would have to be closed like in a transformer.

Would powdered iron core material also be an option?

yes, but as stated, different magnetic cores respond differently to different frequency ranges. you have to choose a type that works at the frequency you want to use.

 

[Bob S]

Originally Posted by Phrak
Watt, wouldn't the addition of core material also increase the coupling?

not much, the magnetic core would have to be closed like in a transformer..

Ferrite rod antennas are often used for AM reception 550-1600 kHz.
Bob S

 

[md5fungi]

Hi, I would just like to provide an update of our project.

We haven't had a chance to look into all of your suggestions, but the impedance matching on our transmitter made a HUGE difference. The LED lit up at what looks like full brightness.

We calculated our power transfer, and seems like we're supplying 92 mW to our transmitter and receiving 46 mW on our receiver at the closest range. The receiver can be a little more than a foot away from the transmitter before the LED turns off fully.

We aren't positive about our power calculations, however... When we hook up an oscilloscope probe (with the ground clip too), the power over our load drops significantly, and our LED turns off. Is there a trick to measuring the voltage of an ungrounded circuit? The best we could do was hook up two probes without the ground clips across our load, and take the difference of the voltages we got.

We are going to explore different antenna designs now; if you could maybe lend some insight regarding this oscilloscope probe, we'd be grateful. My best guess right now is that the probe's impedance is changing the resonant frequency of our receiver... which definitely makes it difficult to take accurate measurements.

 

[vk6kro]

A LED is a very non-linear load. That is its resistance varies a lot with the current that is flowing in it.
For measurements, you could try to get a metal film (non inductive) 51 ohm resistor and take your readings with an oscilloscope across this. Not as spectacular as the LED but more useful.

For low impedance drive and low impedance loads, you probably should be using series turned circuits, with the coil in series with the capacitor for the transmit and receive coils.

You will get best coupling between the coils if they are of similar diameter. About a foot would be OK.

What frequency are you using?

 

[md5fungi]

Providing an update after one semester of work, and seeing if anyone can find a solution to our problem.

Using equations for resonance and inductance, we have managed to light our LED up to a meter away using a big loop antenna transmitter and a spiral antenna receiver. We have also designed some external circuitry (i.e. amplifier, signal generator, etc.).

Our goal now is to get some accurate measurements as far as efficiency. We measured the power supplied by our transmitter and theoretically verified it. The calculation verified our measurement.

However, when we tried to measure the receiver power, we ended up getting more power out than we are putting in! We've tried measuring this various ways, using test loads (various resistors). Unfortunately, we don't know an easy way we can calculate this theoretically, because of factors such as mutual inductance.

One guess we have is that the Earth ground on our oscilloscope is messing up the measurement? We are fairly confident about our transmitter measurement, but we are pretty sure our receiver calculation is incorrect.

Any ideas?

 

[seb7]

Still recommend you look at the SEC exciter. As some claimed 100% power transfer using this method.

 

[vk6kro]

You haven't said what frequency you are using, but there is no risk of you getting over 100 % efficiency. Maybe 5% if you are lucky.

Ideally, I would like to see something like this:

https://www.physicsforums.com/attachment.php?attachmentid=23162&d=1263862172

 

[md5fungi]

You haven't said what frequency you are using, but there is no risk of you getting over 100 % efficiency. Maybe 5% if you are lucky.

Ideally, I would like to see something like this:

https://www.physicsforums.com/attachment.php?attachmentid=23162&d=1263862172
View attachment 23163

We are using 3.00 Mhz.

The problem is our measurements are incorrect; we ARE getting over 100% efficiency with our measurements, which indicates something is wrong with our measurements. It is critical that we can record accurate measurements of the power transfer in our circuit. Here is the circuit we have set up:

http://docs.google.com/Doc?docid=0AW92ycw2vmFnZDR0ajY1OV8xNWdkcHI5NWM5&hl=en
http://docs.google.com/Doc?docid=0AW92ycw2vmFnZDR0ajY1OV8xNWdkcHI5NWM5&hl=en"

R1 is the 50 ohm output impedance of our signal generator. We are using a 1k test load (R2), and using the Vrms value over that load to then calculate the current in our transmitter. We use that current and the voltage supplied to get the power supplied by our transmitter. We have verified this measurement theoretically.

On our receiver, we are measuring the Vrms over our 1k load (indicated), squaring it, and dividing the resistance to get the power at the receiver.

Doing these measurements, there is more power over the receiver than the transmitter is supplying, and we are trying to solve this problem.

We kind of get the feeling that grounding isn't causing the measurement to be thrown off, that we are simply are missing something. Is there a radiative power equation we need to be using?

 

[md5fungi]

Still recommend you look at the SEC exciter. As some claimed 100% power transfer using this method.

Is there any documentation this? As much documentation as possible would be nice since this is for a senior project, and we need to be using reliable sources.

 

[vk6kro]

The power you are transmitting has nothing to do with R2. That is just in series with the transmitting coil.
You would have to measure the current in R2 but the voltage across L1. This will be mostly reactive but a very small percentage of the volt-amps fed to L1 would be radiated.
This is because the coil is too small to be an effective antenna at 3 MHz.
I'm not even sure you could estimate the radiated power like this.

To get more power radiated you could remove (or reduce) R2 as it is just limiting the power radiated.

Also that 3 pF probably needs to be a variable C as stray capacitance would add up to more than that.
The circuit may work better without it.
L1 is very nearly self resonant so you may have to remove some turns to get a better tuning action.

How did you measure L1? Do you know its dimensions and number of turns?
I have a program somewhere that can estimate the inductance of such coils.

 

[md5fungi]

The power you are transmitting has nothing to do with R2. That is just in series with the transmitting coil. You would have to measure the current in R2 but the voltage across L1. This will be mostly reactive but a very small percentage of the volt-amps fed to L1 would be radiated. This is because the coil is too small to be an effective antenna at 3 MHz. I'm not even sure you could estimate the radiated power like this.

I realize the power being transmitted has nothing to do with R2; R2 is ONLY USED to measure current. We're not actually measuring the power transmitted, we're measuring the power SUPPLIED by the signal generator. We get the voltage by measuring the voltage across L1, C2, and R2... which is the same voltage being supplied by the signal generator (or alternatively you could take a BNC cable and hook it up to an oscilloscope to see what the supply voltage is).

Regardless, the power supplied by the signal generator should be greater than the power of the receiver load.

I still have not gotten any helpful suggestions regarding why we are failing to measure the power on the receiver correctly.

Isn't apparent power in = apparent power out? Or am I mistaken?

We are simply comparing Vrms*Irms on the transmitter and Vrms*Irms on the receiver.

How did you measure L1? Do you know its dimensions and number of turns?
I have a program somewhere that can estimate the inductance of such coils.

We measured the value of L1 based upon its dimensions and number of turns. Using L1 and the resonance equation, we solved for the needed capacitance to resonate at 3 Mhz.

Once again, at this point we don't care how well our circuit works, we just want to be able to measure it correctly.

 

[kessy]

Hello, i m new here and i like it, its interesting, and i want to try this. its amazing and this type of http://injectionplasticmould.com/electronics_products.html and it useful in future. you have this type more plans please share with me.
Thanks.

 

[vk6kro]

I used LTSpice to do a simulation of this setup.
I had to make a big assumption that the two coils had a coupling coefficient of 10 %. This is probably on the high side but I will rerun it if you feel it might be different.


https://www.physicsforums.com/attachment.php?attachmentid=23169&d=1263908551

According to this, the signal generator produces 5 volts at 4.3 mA and the receiver tuned circuit produces 1.5 volts across 1000 ohms producing a current of 1.5 mA.
This means that the maximum power in is 21.5 mW and the power out is 2.25 mW.

This would be about 10 % efficiency.

I have put my actual values on your diagram so maybe you could do a check to see where you differ from this.

If you can't find an error, maybe you could try joining the Earth's of the two circuits together. See if that makes any difference.

 

[seb7]

Is there any documentation this?

Sorry, you would need to do a little digging. The reason I find this very interesting as a number of people who have managed to duplicate the SEC are claiming to be measuring over 100% efficiency.


This may help:
http://www.panaceauniversity.org/Spatial Energy Coherence By Dr Ronald Stiffler.pdf
http://creatorguy.com/

http://www.scribd.com/doc/21832427/Replication-of-Dr-Ronald-Stiffler-s-Near-Infinity-Light-System

 

[md5fungi]

Vk6kro, thanks for the info. I will read over your post, compare it to what I'm getting and get back to you.

Seb7, I will take a look at your links, but a claim of over 100% power efficiency violates the first law of thermodynamics and if such a circuit truly existed I think the entire world would know about it by now.

 

[md5fungi]

I used LTSpice to do a simulation of this setup.
I had to make a big assumption that the two coils had a coupling coefficient of 10 %. This is probably on the high side but I will rerun it if you feel it might be different.

It looks like your attachment is broken. Can you reattach it or link me?

Also, the 3.00 Mhz frequency is ballpark, I believe it resonates at something more like 2.926 Mhz or something...

Regardless, I'm interested in what equations you used to find the voltage across the receiver coil. You used a coupling coefficient? How is this calculated and how can I use this to find the power transferred?

If this is found using transformer equations, does the same apply to a resonantly coupled circuit?

 

[Phrak]

I don't intend to read all the posts here to see if this has already been commented upon but...

In viewing your photograph, md5fungi. The axis of pickup coil is not aligned with the axis of the transmitting coil. You do realized that the two should be coaxial, right?

 

[md5fungi]

In viewing your photograph, md5fungi. The axis of pickup coil is not aligned with the axis of the transmitting coil. You do realized that the two should be coaxial, right?

Didn't quite realize that, but we're using a larger spiral antenna for our receiving coil now. From how we hold the spiral receiver relative to the big loop transmitter, it seems to behave how you'd expect a magnetic field to behave.

 

[vk6kro]

OK, I hope you can see this one.

I used the computer simulation program LTSpiceIV. So, no equations, just some voltages and currents for you to compare with if you like.

You said you wanted to find out why you are getting over 100 % efficiency, so comparing these voltages might give you a way of doing that.

I described it in the same way as you did, comparing the received signal level with that out of the signal generator.
However, 18.5 mW of the 21.5 mW out of the generator are actually used up in the 1 K resistor, R2. This leaves 3 mW into the transmit coil producing 2.25 mW in the receive resistor, which is pretty good efficiency

The simulation program gave a double hump in the response. The main peak was at 2.2 MHz (from the transmit coil) with a smaller one at 3 MHz (from the receive coil).

The coupling coefficient is the ratio of the actual mutual inductance present, to the maximum mutual inductance possible between the inductors involved.
I will leave you to look it up if you are interested, but it is a useful way of describing how closely two tuned circuits are coupled together. This is useful because it affects the frequency response and also the power transfer.

 

[Phrak]

Didn't quite realize that, but we're using a larger spiral antenna for our receiving coil now. From how we hold the spiral receiver relative to the big loop transmitter, it seems to behave how you'd expect a magnetic field to behave.

I see. It was just a word to the wise.

And how do you determine that your transmitter and reciever are both resonant at the same frequency as well as their respective Q?

 

[md5fungi]

Once again, vk6kro, thanks for the information. You have been a huge help.

I'm going to try spicing the circuit to see if I get similar results, and hopefully this mutual inductance/coupling coefficient is the missing piece of the puzzle to get our calculations to turn out.

Phrak, the capacitors in the diagram (values could be off, they were off the top of my head) are used to tune the capacitors to a certain frequency to resonate at. We didn't worry much about Q when we started doing this (our primary concern was to get calculations to figure out properly), but that can be easily calculated.

 

[vk6kro]

Hi,
I'm not sure if this is enough to reproduce my results, but it may save you some work:
Turn off the phase plot and get both scales linear.

Version 4
SHEET 1 880 680
WIRE 464 96 336 96
WIRE -80 128 -128 128
WIRE 32 128 0 128
WIRE 192 128 112 128
WIRE 336 144 336 96
WIRE 336 144 304 144
WIRE 368 144 336 144
WIRE 464 144 464 96
WIRE 112 208 112 192
WIRE 112 208 32 208
WIRE 304 224 304 208
WIRE 336 224 304 224
WIRE 368 224 336 224
WIRE -128 240 -128 208
WIRE 32 240 -128 240
WIRE 192 240 192 208
WIRE 192 240 32 240
WIRE 32 272 32 240
WIRE 192 272 192 240
WIRE 336 272 336 224
WIRE 336 272 192 272
WIRE 464 272 464 224
WIRE 464 272 336 272
FLAG 32 272 0
SYMBOL voltage -128 112 R0
WINDOW 123 24 114 Left 0
WINDOW 39 0 0 Left 0
SYMATTR Value2 AC 5 0
SYMATTR InstName V1
SYMATTR Value SINE(0 5 3000000)
SYMBOL ind2 16 112 R0
SYMATTR InstName L1
SYMATTR Value .0017
SYMATTR Type ind
SYMBOL cap 96 128 R0
SYMATTR InstName C1
SYMATTR Value 3e-12
SYMBOL res 176 112 R0
SYMATTR InstName R1
SYMATTR Value 1000
SYMBOL res 16 112 R90
WINDOW 0 0 56 VBottom 0
WINDOW 3 32 56 VTop 0
SYMATTR InstName R2
SYMATTR Value 50
SYMBOL ind2 352 128 R0
SYMATTR InstName L2
SYMATTR Value 6.96e-6
SYMATTR Type ind
SYMBOL cap 288 144 R0
SYMATTR InstName C2
SYMATTR Value 422e-12
SYMBOL res 448 128 R0
SYMATTR InstName R3
SYMATTR Value 1000
TEXT -136 96 Left 0 !;tran 0.000001
TEXT -176 48 Left 0 !.ac lin 50000 1e4 5e6
TEXT -88 -24 Left 0 !K1 L1 L2 0.1

 

[Mike_In_Plano]

I worked on this problem about 12 years ago for a trans-dermal power supply (medical implant). On the primary side, the problem is one of getting sufficient flux linkage in the area of the pick up coil - N, I, and geometry.

For even a few 10's of mW at an inch away, the current in the primary side gets very high - on the order of an amp. To reduce losses on the primary side, the current in the primary recirculated through a parallel resonant tank circuit. Silver mica caps proved to have low loss for this application.

The secondary side was not tuned to resonate with the primary. Instead, it's parallel capacitor served to work with it's leakage inductance to obtain maximum power transfer. Since the coupling between primary and secondary coils would change with the application, the optimum value for the secondary capacitor would vary. Our pragmatic fix for this was to optimize the value for the poorest reasonable coupling - thus you get an improvement in transmission range at a cost in maximum power transfer.

I've attached a couple of drawings to communicate the basic idea and to give some idea as to why the power transfer varies with different orientations of the secondary.

Have fun,

Mike

 

[md5fungi]

vk6kro, I spiced the circuit using LTSpice and also Multisim. The results were satisfactory. I still, however, cannot measure the power correctly.

We've done all sorts of things to measure the power. We're calculating Real Power by using Vmax values and using the equation P = 1/2(Re(VI*)). Besides this, we've tried apparent power and RMS calculations.

We simplified the calculation by building a voltage doubler/rectifier on the output and measuring that as a DC value. No matter what we do, we measure more power on our receiver coil. We also made sure we are picking up no stray signals.

We really need a solution to this problem; we're not sure what's happening, we've tried multiple oscilloscopes/probes and approaches to the problem.

 

[vk6kro]

If you can be sure your resistor is 1000 ohms, you would measure the peak to peak value of the RF signal (that is the total amplitude of the waveform) divide by 2.828 to get RMS and then using V*V / R, calculate the power.

However I very much doubt that your resistor would be 1000 ohms at 3 MHz. I have an antenna analyser which can measure such things and I have seen resistors give very peculiar values at radio frequencies.

The best I have seen are surface mount resistors. At such low power, just a single 1000 ohm resistor would probably be OK. You have two 1 K resistors and I suggest you check or change them both.

Some metal film resistors seem OK. Others are really bad. I can't think of a way of testing them unless you have access to an antenna analyser.

Have you tried having two oscilloscopes, one on the transmitter and one on the receiver?
If you had an oscilloscope on the transmitter and then moved it to the receiver, this could remove a load from the transmitter which might then transmit more power than when you were doing the measurement.

A simple RF probe for an oscilloscope is to use a half wave voltage doubler
http://dl.dropbox.com/u/4222062/RF%20probe.PNG

This gives a reading close to the peak to peak value of the input signal. The capacitors I have shown as 22 pF should be as small (in pF) as possible to still get a good reading.

 

[md5fungi]

If you can be sure your resistor is 1000 ohms, you would measure the peak to peak value of the RF signal (that is the total amplitude of the waveform) divide by 2.828 to get RMS and then using V*V / R, calculate the power.

However I very much doubt that your resistor would be 1000 ohms at 3 MHz. I have an antenna analyser which can measure such things and I have seen resistors give very peculiar values at radio frequencies.

The best I have seen are surface mount resistors. At such low power, just a single 1000 ohm resistor would probably be OK. You have two 1 K resistors and I suggest you check or change them both.

Some metal film resistors seem OK. Others are really bad. I can't think of a way of testing them unless you have access to an antenna analyser.

Have you tried having two oscilloscopes, one on the transmitter and one on the receiver?
If you had an oscilloscope on the transmitter and then moved it to the receiver, this could remove a load from the transmitter which might then transmit more power than when you were doing the measurement.

A simple RF probe for an oscilloscope is to use a half wave voltage doubler
<circuit>This gives a reading close to the peak to peak value of the input signal. The capacitors I have shown as 22 pF should be as small (in pF) as possible to still get a good reading.

Wow, these are both very good points.

It's a good point that the resistors may behave differently at high frequencies, and possibly have a significant self-inductance.

I think we have access to antenna analyzers that will help us make some impedance measurements at HF.

It's also a good point that our oscilloscope may be loading our circuit, and that our circuit may deliver more power when it's not hooked up. I think we tried using two oscilloscopes simultaneously before, but then again, I'm not sure we were measuring power correctly.

Thanks for the help again, vk6kro, this is probably the most helpful response I've gotten so far.