Automotive power: Actual 15W system analysis

After providing readers with a background on wireless power systems in the previous article, we continue our discussion of in-vehicle wireless power charging by examining a 15W solution approved by CISPR 25. Expanding on the material impacts in wireless power transfer covered in Part I, Part II presents various magnetic shielding materials and thicknesses and their impact on system performance metrics, especially EMC compliance, efficiency, and temperature rise.

Taming the Beast

Using the Spark Connected 15W Automotive Tx Wireless Power System, aptly named “The Beast,” testing was performed to compare the type and thickness of shielding materials for the Tx coil. Four ferrites and one powdered iron-based material were tested and results were obtained for EMI emissions, efficiency, and thermal conditions. The Beast platform implements all of the EMI suppression techniques and efficiency improvements mentioned above. Results are for changes to coil shielding only. The results of this testing are provided in Table 1.

Table 1: Tx shield material and thickness and efficiency paired with a 50 mm x 40 mm, 8.22 uH Rx coil

In the Temperature column, shield angle (Cnr) and center (Ctr) measurements are provided for the worst case of 0.3 mm thick shield, as well as ambient + self temperature rise values. For iron powder material MS8, no monolithic shield thickness exceeds 0.3 mm. A single winding Tx coil with good Tx-Rx alignment at 7 mm distance is used and can be improved with thicker shield material.

Download our GaN and SiC technology eBook

Regardless of the shield material, all maximum efficiencies are obtained at a maximum thickness of 0.9 mm (using a 3-sheet stackup of MS8). This also coincides with thicker shields having higher inductance, which helps improve the coupling coefficient by increasing the mutual inductance (Lm) and is given by:

Where,

K – Magnetic Coupling CoefficientLm
– Mutual Inductance (also called “M”)
L(tx) – Inductance of the Tx CoilL
(rx) – Inductance of the Rx Coil

Table 2 also shows that the higher µ' material, labeled FT2, exhibits the highest efficiency across all thickness ranges. There are other loss sources and system factors that come into play, but the data does not address these. One of these is related to Tx coil inductance. Because the resonant circuit is tuned for the superior performing FT2 material and its respective higher inductance, its lower inductance compared to some of the other materials tested detunes the circuit and reduces efficiency. In design practice, care is taken to re-tune the circuit by adjusting the capacitance.

As currents continue to increase for higher power applications, it is important to understand the flux density saturation (Bs) value of the shield material. If saturation occurs, a percentage of the magnetic flux will escape from the back of the magnetic shield and reduce efficiency, as well as create an additional source of heat through eddy currents on any metal immediately behind the shield. This is similar to the thinner shield information in Table 2, meaning that a shield that is too thin cannot contain all the magnetic flux.

Looking back at the curve in Figure 1 from the previous article, another EMI test was performed by applying a metallized film on the Tx coil to help suppress the low frequency harmonic spikes that are prevalent in the curve. The thin metallization of the film does not absorb or attenuate strong magnetic fields very well, but it improves the E field associated with the even harmonics up to the 6th harmonic or 762 KHz. This new plot is now shown in Figure 1.

Figure 1: CISPR 25 Category 5 test 100 KHz to 30 MHz using metallized film

This updated graph shows that the Class 5 requirement has been increased to the 7th harmonic or 889 KHz. Improvements are still needed in the 1-2 MHz range as the quasi-peak values ​​are slightly above the limits in that range. However, the design has been certified to Class 4.

What's next for cars

Automakers have already begun their investigations and preliminary design work has begun for higher powers. The next target is in the 30-45W range, addressing tablets and low-power laptops. There is also a committee within the WPC to deal with this power range and the standard should be available in the next 12 months or so. The idea is that the in-vehicle Tx location could be behind the headrests of the front seats, or in a pouch for someone in the back seat. This would allow for continuous operation of devices that are powered wirelessly or have their batteries charged. For initial evaluation, the Spark Connected 30W Minotaur platform was tested under the same conditions as the 15W system. Table 2 provides the results for the highest and lowest performing magnetics.

Table 2: Tx shielding material, thickness and efficiency

In this higher power scenario, the Tx coil is now a custom coil as no WPC standard coil exists. This improvement is reflected in much lower AC resistance values, which improves system efficiency. In addition, the Tx-Rx coil distance has been reduced to 6mm, and the Rx coil, which needs to support 1.5A current, now also uses Litz wire to reduce its AC resistance. All of this leads to higher efficiency and needs to be part of the solution once 30-45W solutions become commonplace. It should be noted that with the thicker housings used on tablets and laptops, and the need to use Litz wire for each coil, 6mm spacing between coils may not be realistic. This may drive the need to use a "pot" core on each side, which consists of a magnetic shield with side walls around its circumference and a larger "puck" center elevated ferrite post. This brings the two magnetic pieces physically closer to each other and helps to concentrate the magnetic lines of force, thereby improving K and efficiency.

Although this power range is early in its development cycle, there is early discussion of increasing the power level to 65-90 W. With each step up in power, the design becomes more important in addressing EMI, efficiency, and heat dissipation.

in conclusion

As mentioned above, there are three key areas where wireless power systems face increased onboard charging power: EMI, efficiency, and thermal limitations.

The key areas discussed in this article are:

The results show that multiple components and techniques may be required to meet automotive CISPR 25 EMI requirements.

From a circuit design perspective, a push-pull topology with sinusoidal waveform and soft switching is an important approach to reduce EMI radiation and improve efficiency.

From a Tx coil perspective, the results show that the magnetic material (up to +5%) and thickness (up to +5%) of the coil play a significant role in efficiency and must be a key consideration for Qi EPP designs, even more so when power levels increase to the 30-45W range or higher.

It is also explained that system performance is not fully controlled by the car manufacturer or Tx wireless charging subsystem manufacturer. Therefore, the data provided is usually from a "best case" scenario.

Highlighted is the impact of the Z gap, or the separation distance between the embedded Tx coil and the Rx device (phone) to be charged, and that the targeted 5mm maximum no longer applies.

Higher power solutions become more complex and meeting EMI, efficiency and thermal requirements is not easy.

Installing a piece of TDK's transparent conductive silver film stack (silver alloy or Fleclear film) significantly helps to meet the CISPR 25 EMC limits below 1 MHz.

There are other target areas for possible locations of the Tx wireless charging subsystem. Some are: in the door interior area, above the center dashboard area, or embedded in the front seats for rear seat passengers. If there is no proper method to ensure that the Tx-Rx coil systems are oriented to each other, i.e. in a parallel configuration, then this may reduce the coupling coefficient (K) and efficiency, thereby increasing the temperature. We have a lot to look forward to.