Qi Developer Forum. Circuit Design Considerations. Dave Wilson 16-February-2017

Transcription

1 WPC1701 Qi Developer Forum Circuit Design Considerations Dave Wilson 16-February-2017

2 Overview Getting Started Basics The Qi Advantage for Circuit Design Practical Design Issues Practical Implementation Issues Worldwide Agency and Government Requirements Do not worry if we go fast through some slides!! A lot of this material is for you to look at a later time. PAGE 2

3 Start: Consider Wireless Power Delivery Methods - Electromagnetic Induction - Magnetic Induction - Direct Induction - Magnetic Resonance - Resonant Transformer - Resonant Inductive Coupling Magnetic field radiates outward - not shielded Magnetic field mostly contained in between ferrite shielding - Capacitive Coupling - Radio Waves - Acoustic Waves PAGE 3

4 Start: Consider Wireless Power Delivery Methods - Electromagnetic Induction - Magnetic Induction - Direct Induction - Magnetic Resonance - Resonant Transformer - Resonant Inductive Coupling Magnetic field radiates outward - not shielded Magnetic field mostly contained in between ferrite shielding Makes sense only for applications where extra cost, difficultly, and - Capacitive Coupling - Radio Waves - Acoustic Waves complexity make sense PAGE 4

5 Start: Consider Wireless Power Delivery Methods - Electromagnetic Induction - Magnetic Induction - Direct Induction - Magnetic Resonance - Resonant Transformer - Resonant Inductive Coupling - Capacitive Coupling - Radio Waves - Acoustic Waves Magnetic field radiates outward - not shielded Magnetic field mostly contained in between ferrite shielding Makes sense only for applications where extra cost, difficultly, and complexity make sense Practical only for very specialized applications. Many issues!! PAGE 5

6 The Basic Air Core Transformer Idea Coupling coefficient k impacts mutual inductance M k influenced by turns ratio, ferrite area/thickness, and the space between the coils (k = ideally) V1 V2 PAGE 6

7 Study Likely Coil Size Ratios and Separation Distances Coupling between coils Distance (z) between coils Ratio of diameters (D2 / D) of the two coils Physical orientation Quality factor Ratio of inductance to resistance Geometric mean of two Q factors Uncoupled field has no losses Near field allows TX to see RX PAGE 7

8 Consider Coil Alignment Possibilities PAGE 8

9 Work Through All of the Efficiency Tradeoffs System Efficiency = hsystem=pol/pin Mag Eff Tx Eff Rx Eff PAGE 9

10 Start Circuit Design After Deciding What is a good operating voltage? What is the minimum / maximum power delivery required? How should the system respond to load changes? What about thermal design, EMI, and efficiency? How large is the good working area for power delivery? and should the design be single-coil or multi-coil?? Do a resonant type design or a non-resonant type design? PAGE 10

11 The Problem with this Design Approach... There is a truly infinite matrix of design decisions and tradeoffs before even starting a circuit design Following these rules requires a very deep understanding of physics, flux flow/capture, simulation/modeling of systems, and so-on A good design can take a very long time to make, and a fully working and tested system using discrete components may also be very expensive A completely new design based on the basic physics is unlikely to be interoperable with any other independently designed system PAGE 11

12 The WPC Qi System Advantage Best of both worlds solution using partially resonant magnetic induction Loosely coupled air-core transformer Useable air-gap working distance of about 3mm 8mm (depending on transmitter type) Partially resonant Transmitter LC tank I db/dt Partially resonant Receiver LC tank Air gap PAGE 12

13 Simplified Design Using WPC Qi Systems Qi certified Transmitter designs are fully worked out, including detailed specification of all operational parameters and the physical coil design Receiver designs are similarly worked out, but in a way that enables the designer to choose/modify/make a Receiver coil that perfectly fits the design Because Qi is a robust interoperability standard, IC makers have invested to make single-chip solutions that completely eliminate most of the design work PAGE 13

14 WPC Qi-Based Circuit Design Flow Transmitter Design Choose the pre-defined Qi Transmitter type that best matches the application Choose from an IC vendor a chip solution that best meets the requirements for cost, efficiency, EMC emissions, component count, etc. Follow closely the IC vendor guidelines for system implementation Receiver Design Choose from an IC vendor a chip solution that best meets the requirements for cost, efficiency, EMC emissions, component count, physical area/height, etc. Choose a receiver coil that best matches the application and IC vendor guidance Follow closely the IC vendor guidelines for system implementation PAGE 14

15 WHAT COULD POSSIBLY GO WRONG?? Actually there are a few things, and system circuit designers are still very important This is the really practical stuff!! PAGE 15

16 WPC Power Transmitter Basic Topology Primary coil (L p ) + serial resonance capacitor (C p ) DC-to-AC Inverter: e.g. half bridge (shown below) or full-bridge Power level is controlled by changing transmitter operating frequency, operatin duty cycle, and/or bridge supply voltage. Power is controlled by the Receiver which is the master of the transmitter Multiple coil solutions function the same as single-coil with the best coil selected by the transmitter before beginning interoperation with the Receiver Power Conversion Power Conversion Freq + - Half Bridge C p L p Freq + - L m C m Multiplexer L p Single Coil Multiple Coils PAGE 16

17 Load Secondary coil (L s ) WPC Power Receiver Basic Topology Serial resonance capacitor (C s ) for efficient power transfer Parallel resonance capacitor (C d ) for required alternative method to detect presence of a Receiver device by making 1MHz resonance Important Note: The C d value must be chosen to work with the selected Receiver coil while on a non-operating transmitter to accurately decide the capacitor value. This will help transmitters that use this method to work well. AC-to-DC Rectifier: full bridge (diode, or switched) + capacitor Output switch for disconnecting the load C s Power Pickup Unit L s C d C PAGE 17

18 WPC Communication (Modulation) Receiver modulates load by method of Amplitude Shift Keying (ASK) Switch on/off modulation resistor (R m ), or Switch on/off modulation capacitor (C m ) Transmitter demodulates Receiver information by Sensing primary coil current (I p ) and/or Sensing primary coil voltage (V p ) Note: Demodulated amplitude ranges from about 800mVpp to 10mVpp Transmitter modulates the frequency of the coil power signal to send information to Receiver Receiver measures Tx coil drive frequency to demodulate Transmitter information Transmitter Receiver C p C s Modulation Modulation + Load C d C m C R m - L p I p V p Power L s PAGE 18

19 Start Parity Stop WPC Communication (Data-Format) Speed: 2 Kbit/s Bit-encoding: bi-phase Byte encoding: Start-bit, 8bit data, parity-bit, stop-bit Packet Structure Preamble (typically 4-11bits) Header (1 Byte) Indicates packet type and message length Message ( Byte) One complete message per packet Payload for control Checksum (1 Byte) 0.5ms b0 b1 b2 b3 b4 b5 b6 b7 Preamble Header Message Checksum PAGE 19

20 Now Finally the Actual Very Important Circuit Design Work PAGE 20

21 Current-Sense circuit should closely follow IC vendor recommended Example of Circuit Design Special Cases Precision Current-Sensing Type Resistor Bulk and bypass capacitors should closely follow IC vendor recommended Single-Chip Transmitter with Integrated Power Transistors Caps Must Be COG Type Main Coil should be approved type in certified product Coil peak detector and demodulation circuit all closely the same as IC vendor recommended PAGE 21

22 Design Checklist: Component Selection (1) If the transmitter uses external power transistors for the bridge driver, these must be carefully and fully designed-in IC vendor may specify gate drive characteristics, switching time, etc. Typically the designer can choose the device on-resistance for the best cost decision Passive components tied to the power transistor may need to be adjusted according to the IC vendor recommendations depending on the properties of the selected transistor The transmitter coil is best selected by choosing a coil that is used in an actual design that has passed all WPC requirements. Alternative coils could have issues to be checked carefully, such as: Improper inductance value for the specified transmitter type Higher than expected coil resistance (could cause FOD or Guaranteed Power issue) Thinner than specified ferrite material (could cause FOD and/or Guaranteed Power issue) Low quality ferrite material that could be lossy, become easily saturated, etc. COG capacitors in the power circuit are a must and should not be substituted. For example, using instead an X7R type in a Tx-A11 design will add about 400mW heat loss into the capacitors. This can cause FOD and Guaranteed Power problems and could cause failure of the capacitors. This is from the partial resonance in the LC tank making a big circulating current PAGE 22

23 Design Checklist: Component Selection (2) If the transmitter uses an external current sensing resistor, this is very important for the FOD and other measurements. So a true current-sense type resistor should be used with the accuracy as specified by the IC vendor Bypass and bulk capacitors that on both the Transmitter and Receiver should not be made less than or different from the IC vendor recommendation. And if these are low-cost ceramic type, then the derating of the actual capacitance value should have careful engineering attention so that the effective capacitance value is the same as what the IC vendor recommends. (Note: The IC vendor may already have taken into account such derating assuming a particular type of capacitor is used.) Additionally, capacitors can make acoustic noise, and this can vary depending on the construction and material type of the capacitor. If there are any demodulation passive components in the transmitter or modulation components in the Receiver, the type and tolerance recommendation of the IC vendor should be followed. These may be critically balanced for best performance, and changes can cause unreliable communication issues between Rx and Tx. PAGE 23

24 Design Checklist: System Physical Layout (1) Most applications require high power, high efficiency, good thermal performance, and low EMI. The circuit designer and layout designer should be careful to follow all of the normal rules for good design of these kind of issues. And they should note that because of the partial resonance that the circulating current can easily be 2x or more higher than the average DC current flowing in some paths of the system. The most common unexpected cause of an EMI failure is accidental series inductance between the power transistors and their supporting bypass and bulk capacitors. Even a wire as short as 1cm has enough inductance to cause a severe ringing in a power circuit. This type of failure will typically show up as a strong emission in the 30MHz 100Mhz frequencies. If there is a signal diode for the purpose of recovering a demodulation signal on a Transmitter design, this diode should be placed near any other demodulation passive components which should placed very close to the IC demodulation function pins. The reason for this is that the demodulation analog signal at times can be as small as about 10mV, and so it is easy to lose this signal because of noise that is added from long layout signal paths. PAGE 24

25 Design Checklist: System Physical Layout (2) It may be desired to place the Transmitter coil very close to the Transmitter coil PCB ground plane. This can be allowed provided that the Transmitter coil shield material is high quality. If the shield is less good, then there can be trouble with FOD calibration and/or passing Guaranteed Power tests. It should be avoided in general to place friendly metals in the Transmitter closer than about 10mm. At a closer distance, such metals will have more and more effect on the FOD adjustments and will reduce system efficiency. If total losses are too much, then the Transmitter could also fail the WPC temperature-rise limit for the temperature testing. The Receiver coil can also be sensitive to close placement to the ground plane of the Receiver PCB. And in some Receiver designs, it may be a must for the shield material to be very thin and this could affect FOD, efficiency, and heating of the Receiver unit. The Receiver coil can also be affected by friendly metals in the Receiver that are closer than about 10mm to the Receiver coil. FOD, efficiency, and heat issues can be more when metals start become closer than 10mm PAGE 25

26 Design Checklist: System Physical Layout (3) If there is an external current-sensing resistor in either the Receiver or Transmitter circuit, this should be connected to the device pins using a Kelvin style connection. In this method, the layout is such that no current flows through the two wires used to measure the resistor voltage. This is very important for accurate current measurement. PAGE 26

27 Kelvin type connection path between sensing resistor and IC pins Example of Circuit Layout Special Cases Precision Current-Sensing Resistor DC Power Supply path to Caps <15mm. Important for Good EMI Performance!! Power Path to Main Coil MUST be short and low loss at 200kHz Single-Chip Transmitter with Integrated Power Transistors Main Coil Demodulation circuit all placed close to the IC demodulation pins PAGE 27

28 Do Not Forget the Power Supply is Part of the System! The power supply is coupled directly to the power delivery bridge circuits. If bad, then power delivery could be unstable. Most importantly and easy to miss: Power supply noise interferes directly with the load modulation communication signal from the Rx. Real example from customer issue: Unlucky power supply noise repeating at 248us with almost the same period as WPC 250us communication signal! Preferred power supply noise: Periodic noise <10mVpp 400Hz-20kHz Bigger noise such as 100mVpp is OK at greater than 20kHz! No sudden change of DC level faster than about 10mV/ms. Slow change is OK! Clean communication 50mVpp 500mVpp WPC communication signal inside Tx cannot be decoded! PAGE 28

29 Worldwide Agency and Government Compliance WPC certification and specifications do not address worldwide requirements for EMI/EMC, efficiency, materials, etc. And these specifications can be strongly different depending on the country, or in cases, even depending on a smaller region inside a country. Designers must have good knowledge of all such requirements where they plan for their product to be sold. Examples: CISPR-22 FCC Part-15, Part-18 EN (magnetic emissions) California Green efficiency requirements Regional Green materials requirements for safety Regional Green materials requirements for recycling PAGE 29

30 Supplemental Information PAGE 30

31 Foreign Object Heating WPC Compliance Test Three Reference Test Foreign Objects are Defined in Detail Object #1: 15mm dia steel disk with integrated thermocouple Object #2: 20mm dia aluminum alloy disk with integrated thermocouple Object #3: 20mm dia aluminum foil disk with integrated thermocouple Test frames and spacers are also defined for placing/holding test objects on Tx surface Example Cross Section View 17.9mm dia for comparison Example Plan View 15mm dia object Synopsis of Test Procedure Follow various specified placements and sequencing If Tx refuses to go to power transfer with the object present, this is passing If Tx terminates power transfer within various times/metrics, this is passing If Tx continues power transfer, but object temperature remains below limit, this is passing Pass if objects #1, #2 < 60-C heating and #3 < 80-C heating PAGE 31

32 FOD Power Loss Method Step 1 Transmitter determines actual power available to Rx Measure input power Pin Account for all known Tx losses Pptloss Determine power available at Tx coil surface Ppt PAGE 32

33 FOD Power Loss Method Step 2 Receiver determines actual power available from Tx Measure Pout Account for all known Rx losses Pprloss Determine power available at Tx coil surface Ppr PAGE 33

34 FOD Power Loss Method Step 3 Transmitter determines if there is an unexplained loss Receiver reports Ppr as Received Power Transmitter calculates Ppt-Ppr If more power is lost than allowed, then FOD PAGE 34

35 Q-Measurement Method Normally, for efficiency, the Tx has a very high Q, which is the Quality Factor. When the coil is very low resistance and losses in the Tx PCB, COG capacitors, and transistors is very small, then the Q-factor will be very high When there is any loss of energy from the Tx coil such as by foreign metal, friendly metal, or power taken by the Rx, then this reduces the Q of the Tx. In the Q-Measurement method, the Tx uses information from the Rx to decide if the Q-factor with the Rx present is the amount expected. If there is some FOD, the Q-factor measurement will be lower than expected and so Tx will decide there is an FOD case PAGE 35

36 Circuit Design Requirements for FOD Measurement Tx and Rx must be able to accurately measure their power (or voltage and current) at the correct time and with sufficient bandwidth. These parameters will vary with different IC designs and measurement methods used by those ICs. Typical implementations have some low pass passive circuit to smooth the measured value to avoid aliasing. Additionally, for Q-measurement, the Transmitter must accurately determine its peak coil voltage at the correct time and with sufficient bandwidth. Typical implementations have some low pass passive circuit to smooth the measured value to avoid aliasing. PAGE 36

37 Calibrating the Tx FOD for Loss Method Calibration of the Tx for the loss method is done by: 1. Step through the entire power range from zero to maximum 2. Measure the known losses of the Tx by comparing the available output power to the actual DC input power 3. Make a table of the known losses inside the Tx 4. Use the loss table to determine the available output power for all possible cases Note: For this to work, the output must be measured very accurately, and the only way to do this at this time is to use a specialized test tool such as made by Nok9. PAGE 37

38 Calibrating the Rx FOD for Loss Method Calibration of the Rx for the loss method is done by: 1. Step through the entire power range from zero to maximum 2. Measure the known losses of the Rx by comparing the available output power to the actual power available from the Tx coil 3. Make a table of the known losses inside the Rx 4. Use the loss table to determine the actual power available to Rx based on the measured output power from the Rx 5. Add a loss allowance so that there is some margin for error when the Tx does the loss calculation to determine FOD or not Note: For this to work, the power available from Tx must be measured very accurately, and the only way to do this at this time is to use a specialized test tool such as made by Nok9 LOSS ALLOWANCE TABLE FROM WPC SPECIFICATIONS PAGE 38

39 Calibrating the Q Method The Tx must know from its memory the Q-factor of its overall design. This can be a self-calibration, but it must be done at a time when it is 100% guaranteed there is nothing on the Tx surface. (So this cannot simply be done in the field.) The Rx must know its effect on the Nok9 MPA1 Tx Q-factor. This must be measured in the lab on the Nok9 tool at a time it is 100% guaranteed there is no foreign object present. The Rx must remember this value in its memory and report this value during the startup sequence with Tx PAGE 39