2014-03-21

Measuring the performance of the wireless power demo kit

When talking about wireless charging, one of the questions that always come up is: what are the distances this will work for? I did answer this question already for another wireless charging kit, so I was curious what the Qi (or WPC) standard had ins store.

I’m interested in the maximum distance that is suitable for transferring a significant amount of power. Especially I wanted to know how well the transmit and receive coils need to be aligned with each other. After all, we are talking about a kid and its toys - one cannot expect millimeter precision when trying to charge to a toy over night.

The test setup

Note: for updated measurements, including efficiency, look at this update.

So I started with designing the test. First I removed to acrylic plate on top of the transmitter coil, to get the distance between transmit and receive as small as possible. Since the receive coil is mounted in the plastic cage together with the receiver board, I left it there (its fixed with some tape, which looked quite firm). That meant the minimum distance between the coils was about 1 mm. The acrylic plate I had removed was about 2.5 mm thick, so adding it back again gives a distance of about 3.5 mm as second test point. Last one was a foam sheet I had laying around, which was about 9 mm thick (so I had 10 mm as maximum distance)

As next step I looked at the load I wanted to test. First data point should be with no load at all, to get kind of a base line. Then the LED board coming with the kit seemed like a good target, even though I don’t know how much power it needs (and I need to admit I forgot to measure….).

To get some more real life numbers, I went with load currents of 100, 250 and 500 mA as data points, to simulate different types of load. My expected load current is about 350 mA (according to my current plans), so I should get an idea how this will behave. For these loads I used some resistors, and measured output voltage and load current with my DMMs.

This setup then looks like this:

Load test setup

Each test started with the receiver coil centered on the transmitter, so the transmission starts reliably (which gets signaled by the LEDs). Then I shifted the receiver outwards until the measured voltage started to drop (or in case of the LED board, the LEDs started to flicker visibly). This distance is called “Offset (charge)” below.

Then I removed the receiver altogether (to stop the power transfer completely), and put it onto the transmitter again - but not centered, but with a side-wards shift. I repeated this multiple times until I found the maximum distance the coils could be shifted, while the power transfer starts up reliably (meaning I got a 5 V output) (this is called “Offset (startup)” below).

The results in numbers

Here are the raw numbers:

  Load   Distance   Offset (charge)   Offset (startup)
0 mA 1 mm 13 mm 13 mm
0 mA 3.5 mm 14 mm 13 mm
0 mA 10 mm 13 mm 9 mm
LED board 1 mm 17 mm 8 mm
LED board 3.5 mm 18 mm 9 mm
LED board 10 mm 16 mm 8 mm
100 mA 1 mm 13 mm 12 mm
100 mA 3.5 mm 15 mm 13 mm
100 mA 10 mm 14 mm 8 mm
250 mA 1 mm 14 mm 12 mm
250 mA 3.5 mm 15 mm 13 mm
250 mA 10 mm 14 mm 9 mm
500 mA 1 mm 13 mm 12 mm
500 mA 3.5 mm 15 mm 13 mm
500 mA 10 mm 13 mm 8 mm

Its interesting to see that putting the coils directly on top of each other actually loses some performance. With a 3.5 mm distance the coil center could be shifted a little bit farther away than otherwise. Its also interesting to see that the amount of power drawn does not affect the performance - Qi employs adaptive control over the power and can so adapt to changing load conditions.

Some other observations, and conclusions

While conduction the tests, I also made some more observations about the behavior of the boards:

  • When a power transfer has been established and the coils get shifted against each other, the output voltage drops shortly and then stabilizes again at 5 V. This was especially noticeable with 500 mA load - I guess this is the adaptive power control at work
  • The adaptive control also means that when the power transfer gets interrupted due to the distance of the coils, the need to moved closer together to restart the power transfer. This gets more noticeable with a larger vertical distance (but not with load).
  • Since the receiver coil is rectangular and not round, shifting it to different directions yields different results. I shifted to the long side for my tests, but using the short side reduces the distances where the transfer still works.
  • Without any load, the voltage dropped immediately to zero when close to the maximum range. With some load it just started to drop slowly, while still delivering current into the load.
  • 10 mm seems to be the maximum vertical distance between the coils - even one millimeter more and the transfer will not initiate
  • Having a larger vertical distance means the allowed offset between the coils gets lower
  • The LED board behaves differently: its DC/DC-converter allows it to run with less output voltage, so the allowed offset is larger. On the other hand, it doesn’t seem to like startup into load, so the distances here are noticeably smaller.

Given that the results depend on the direction the coils are not aligned, it seems I should try a round receiver coil to get better results. But even then I need some visual indication that the power transfer has started, so my son can see that the toy is positioned properly.

Judging from what I have seen with the LED board, it might be a good idea to prevent starting into heavy load (esp. when involving switch-mode regulator). So maybe I will add a startup delay for charging.

It also seems that, as long as one doesn’t rely on the maximum vertical distance of 10 mm, drawing even the full load of 1 A doesn’t reduce the performance in terms of proper alignment. Should I decide to need more current than currently planned, I should test that though…

Next steps

My next step in this project is to decide the final design. So my next post will explain my basic ideas and the challenges, and the ones after that will detail these a little bit more. There are several ways to overcome the obstacles, so I have different routes to shine light to. Having done that, I can try some of my ideas with prototypes to verify that they really work as intended. Having finished that, a PCB can be designed and manufactured, and then the mechanical work can start.

Stay tuned!

Posted by Hendrik Lipka at 2014-03-21 (Google)
Categories: electronics projects contests