[SOLVED] Why low efficiency in Wireless power transfer systems ?

[nvt088]

Hi everyone.

I simulated and fabricated a WPT system using two loops (source and load) and two spiral coils (TX and RX) as showed in picture. The method is resonant coupling.
The simulated result is good. Efficiency at 1m is 80%.
However, my measured result is bad. Efficiency at 1m just 10%. I used Vector Network Analyzer for measurement.

Is there any reason for this low efficiency? Is it exact to use VNA for measuring efficiency because the input power of VNA is very low, 1W, right?

One more question, how to check up the resonant frequency for each loop and spiral coil? I want to make sure they have the same resonant frequency. Many thanks

 

[BigBoss]

I think your measurment environment is not proper because there may be some reflections,diffractions,scatterings from other objects in that environment.Especially metallic objetcs ( for instance your VNA ) will create some reflections with a phase changings and when these reflected,scattered waves have arrived to you receiver with different phases, they will probabaly create "beatings" and overall wave power will drop.
Instead, you should use antenna chamber to isolate these unwanted waves coming from different metallic and non-metallic objetcs.Rear reflected waves are also trouble.

 

[chuckey]

If you use a solenoid or coil to transmit your power, a half of the power goes out backwards. You need to reflect this power forward so it is phase with the forward wave, like a reflector in a yagi aerial.
Frank

 

[FvM]

In the shown setup, the instrument near the coils and the table frame are probably disturbing the field.

Generally speaking, it's rather unlikely that the real circuit has identical behaviour as the simulation. At least, you'll need to tune the resonant circuits to factor in component tolerances.

Another question is if your simulation models the coil geometry and e.g. skin effect losses realistically.

 

[vfone]

One more question, how to check up the resonant frequency for each loop and spiral coil? I want to make sure they have the same resonant frequency.

I see that you have the VNA on the desk. It is exactly what you need to find the resonant frequency of the coils.

How did you calculate 80% efficiency of power transfer at 1m distance? Sounds too good for me.

 

[D.A.(Tony)Stewart]

To capture most the magnetic field lines of force the rule of thumb is the loop diameter must be twice the separation distance. and then you only get one side of the loop forces. After that inverse square loss from dispersion prevails. Resonant frequency is easily to scan and calibrate with a tuned high Q capacitor. ( polyester, polyurethane, mica or teflon)

 

[nvt088]

I see that you have the VNA on the desk. It is exactly what you need to find the resonant frequency of the coils.

How did you calculate 80% efficiency of power transfer at 1m distance? Sounds too good for me.

@vfone:

I used formula: efficiency = |S21|^2 . 80% is a simulation efficiency.

Unfortunately, the measurement efficiency is just 10%

 

[FvM]

I used formula: efficiency = |S21|^2 . 80% is a simulation efficiency.

The formula is correct. Can you show how you achieved this high efficiency in your simulation circuit? I doubt that it's possible with real components.

To capture most the magnetic field lines of force the rule of thumb is the loop diameter must be twice the separation distance. and then you only get one side of the loop forces. After that inverse square loss from dispersion prevails.

Assuming the setup is working at e.g. 10 or 15 MHz, only a small part of the power fed to the coil is actually "dispersed" or radiated to free space. Most of the energy is circulated between the magnetical near field and the resonant circuit. The real power is mostly burned in circuit losses. It's not theoretically impossible to capture more than 50 percent of feeded real power on one side of the coil, but I agree that it's difficult with the sketched geometry.

 

[nvt088]

The formula is correct. Can you show how you achieved this high efficiency in your simulation circuit? I doubt that it's possible with real components.

In my simulation:
Diameter of loop (source and load) is 30 cm.
Outer diameter of spiral coil is 60 cm, pitch 1cm, thickness of wire 0.5 cm.

At distance 1m between two spiral coils, I change that distance between loop and spiral coil, so efficiency in simulation is 80 %.

I guess that the cable of VNA for measuring can impact on result, right? Do you have any idea about de-embedding VNA?. Because I need to take S11, S21 at the position of source and load loop (not at VNA) because the length of cable is 0.5m.

Many thanks

 

[volker@muehlhaus]

In my simulation:
Diameter of loop (source and load) is 30 cm.
Outer diameter of spiral coil is 60 cm, pitch 1cm, thickness of wire 0.5 cm.

Metal loss included, or ideal conductors?

 

[nvt088]

Metal loss included, or ideal conductors?

In simulation, I use copper for loop and spiral coil. So, HFSS take into account their losses already, right?

 

[volker@muehlhaus]

In simulation, I use copper for loop and spiral coil. So, HFSS take into account their losses already, right?

Yes, if you have defined the correct metal properties.

 

[FvM]

|S11| or |S21| aren't be affected by shifting the reference plane (port extension). But it's a simple operation for all recent VNAs.

Is the transmission system designed for 50 ohm source and load impedance?

 

[nvt088]

|S11| or |S21| aren't be affected by shifting the reference plane (port extension). But it's a simple operation for all recent VNAs.

Is the transmission system designed for 50 ohm source and load impedance?

Yes. I utilized 50 Ohm for source and load impedance. My HFSS result is almost same with CST simulation result. Just experimental result is low.

 

[volker@muehlhaus]

Can you show the measured and simulated S11 and S21 curves?

 

[nvt088]

Can you show the measured and simulated S11 and S21 curves?

 

[volker@muehlhaus]

Can you also show the measured S11 and S21?

 

[nvt088]

Can you also show the measured S11 and S21?

This is my measurement result. Do you have any idea to improve its efficiency? Many thanks

 

[volker@muehlhaus]

Do you have any idea to improve its efficiency?

You need to add snake oil.

Or you can start to troubleshoot the existing setup. Check the simulated and measured resonator Q factor, to understand the losses. Also, in simulation you can look at the simulated fields, and check if your lab environment (other metal etc) can be critical.