Post Go back to editing

question about the resonant freq of LTC4125 with LTC4126

Hi ADIExpert

I have a system using LTC4125 as a wireless TX and LTC4126 as a wireless RX

I have a question about the resonant freq discrepancy between RX and TX

is there rule or constraint for the RX LC and TX LC resonant freq?

This was my test flow 

1. first a TX coil:6.8uH, Tx cap: 0.1uF, (calculated resonant freq: 193kHz) is tested.

this is the auto resonant freq (around 203Khz) measured without any RX attached

2. then, I put a RX coil: 13.3uH, RX cap: 0.039uF( calculated resonant freq: 221kHz) on the TX

I measured the waveform (freq:226K) as below with power transferring from TX to RX without problem 

3. Finally,  I tried to reduced the  RX resnoant freq to match Tx's one by changing the cap from  0.039 to 0.047uF

but the circuit doesn't work, the auto resonant freq seems could not be found (TX freq has gone beyond the max resonant freq)

May I have any advice why this happened?

also, if I keep the cap 0.039u, the resonant freq is 10% higher than the TX setting. if there is any issue may cause or just lower efficiency?

thank you

Ken

Parents
  • one more question about the charging termination scheme what condition the LTC4126 may present done charging status?( is it C/10 or other kinds of termination method)

  • Hi Kencheng, 

    It is normal to see higher resonant frequency when you have a Rx on top of the Tx coil. The LTC4125 auto-resonant will re-adjust driver to work with this higher resonant frequency. It is okay to use 0.039uF. 

    However, I am not sure why your 0.047uF is not working. In our DC2770A-B demo board, we used 0.047uF with a 12.6uH inductor and it works well. Please see this link below for details. 

    https://www.analog.com/media/en/technical-documentation/user-guides/DC2770A-B-KIT.pdf

    Typically for 7.5mA charge current with LTC4126, we would recommend to use simple one switch transmitter as shown in DC2663A demo kit. The LTC4125 could be less efficiency when paired with LTC4126. 

    https://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/dc2663a-kit.html

    LTC4126 uses timer based termination. The termination timer is set at 6-hour. 

    Thanks,

    Wenwei.

  • Hi WenWei

    thanks for the reply, however here comes a new question when the charging is almost finished and  LTC4126 is operating at CV state

    from the below waveform, we can see the the charging current has dropped from 50mA to 22 mA because of CV mode, the battery voltage is around 4.17V

    however, we can see the TX side randomly failed to detect RX (can be observed by the stat pin of LTC4125 as well as the VCC pin of LTC4126)

    1. case1, charing current drops to around 22mA

    2. case2, charging current drops to  2.5mA

    the stat pin goes high more frequently

    so I just want to confirm if the condition is normal and how to explained the TX failed to detect RX. Is it because the demanding charging current to battery is too small?

    from the VCC voltage, it doesn't look like ΔVUVLO


    thanks

    Ken

  • Hi Ken, 

    As the charge current decreases, it is likely the FB voltage differences on each step during optimum power search is changed. If LTC4125 fails to trigger the exit condition, it will not detect the Rx, and will not provide the power to charge the battery. 

    Here are all the exit conditions for the LTC4125: 

    1.ΔVFB(n) > 2(Δ VFB(n-1))+1bit

    2.VFB(n-1) > VFB(n)MIN + 2 bits &
    # of Steps > 2

    3.ΔVFB(n) > [VDTH]7bits &
    ΔVFB(n-1) < 2 bits &
    # of Steps > 2

    4.VIMON > 0.8V

    I think increasing capacitance on VCC on LTC4124/LTC4126 could be a good attempt to improve the detection during CV region. 

    Thanks,

    Wenwei.

  • Hi Wenwei

    we have tested several caps value 0.1u/1u/10u/20u on Vcc pins

    however it didn't improve the detection of CV region

    is there anything I can check for this issue? like adjust the FB network divider ratio of the LTC4125?

    thank you

  • Hi Kencheng, 

    It might be better to understand why the exit condition was not triggered first. Do you have a zoom-in picture of the FB pin when it fails to detect the RX? 

    Here are the conditions to trigger the detection: 

    1.ΔVFB(n) > 2(Δ VFB(n-1))+1bit

    2.VFB(n-1) > VFB(n)MIN + 2 bits &
    # of Steps > 2

    3.ΔVFB(n) > [VDTH]7bits &
    ΔVFB(n-1) < 2 bits &
    # of Steps > 2

    4.VIMON > 0.8V

    You might need to find some way to help LTC4125 fulfill one of these conditions in order to detect the Rx. 

    Adjusting the FB network might help but without knowing the exact behavior of the FB pin waveform, it might be difficult to decide what need to be done. 

    Thanks,

    Wenwei. 

  • Hi Wenwei

    this is the test result, FYIXLS

    from the test result, it seems that the no charge state has even high deltaVfb(n)

    however it stills cannot find the exit condition?

    1.ΔVFB(n) > 2(Δ VFB(n-1))+1bit  ( looks fine from the waveform?)

    3.ΔVFB(n) > [VDTH]7bits &
    ΔVFB(n-1) < 2 bits &  ---> (2bits means 2 x 39mV when Vin=5V, DTH=Vin)?
    # of Steps > 2

    note: DTH pin is tied to Vin with a 100k ohm resistor without any voltage divider

    I also attached the schematic again 

    thank you

Reply
  • Hi Wenwei

    this is the test result, FYIXLS

    from the test result, it seems that the no charge state has even high deltaVfb(n)

    however it stills cannot find the exit condition?

    1.ΔVFB(n) > 2(Δ VFB(n-1))+1bit  ( looks fine from the waveform?)

    3.ΔVFB(n) > [VDTH]7bits &
    ΔVFB(n-1) < 2 bits &  ---> (2bits means 2 x 39mV when Vin=5V, DTH=Vin)?
    # of Steps > 2

    note: DTH pin is tied to Vin with a 100k ohm resistor without any voltage divider

    I also attached the schematic again 

    thank you

Children
  • Hi Ken, 

    It seems like when LTC4125 is not able to detect the LTC4120, the first criteria is not met. The delta VFB of step #5 is not greater than 2x delta VFB from the previous step (step#4). Please see this picture before from your excel sheet. 

    It seems like the VFB is not quite flat and has some fluctuation at the very end of step #4. 

    The reason why there is a big jump in the case when it is able to detect the LTC4126 Rx is that on that big jump step, LTC4126 reaches a level where the receive power is enough to charge the battery, and the LTC4126 wireless power manager starts to shunt the Rx tank to ground periodically. Because of the start of the shunting event, the Rx reflected impedance on the Tx side will change significantly. The voltage level of FB on LTC4125 basically reflects the impedance of the Tx side, combining with the reflected impedance from the Rx side. As a result, when reflected impedance changes significantly, a significant change can be observed on FB pin of LTC4125. 

    In your case when LTC4125 fails to detect the LTC4126. On step#4, it seems like it is in a boundary condition. At the beginning of step#4, the LTC4126 wireless power manager has not started to shunt the Rx tank yet. However, for some reason, by the end of step#4, LTC4126 is able to provide just enough power and the shunting event starts partially, creating some oscillation on FB voltage.  

    One idea that might help to improve the detecting is just to use half bridge on LTC4125,as suggested in this article below. It might help to create bigger difference in the step where the LTC4126 reaches enough power. 

    https://www.analog.com/en/technical-articles/complete-efficient-100-ma-wireless-charging-solution-with-optimum-power-transfer-and-safety-features.html

    In the half bridge setup, the SW2 of LTC4125 is not used. The other side of the Tx tank can be directly connected to ground. 

    Using LTC4125 in optimum power search mode to work with very low power Rx could be a bit challenging as the jump in VFB steps is smaller for the low power Rx. We would usually recommend just the simple LTC6990/6992 to drive a one switch transmitter for LTC4126. As the power level is low, the heat should not be an issue. 

    Thanks,

    Wenwei.

  • Hi Wenwei

    we tried the half bridge setup.

    A huge delta Vfb can be observed from the waveform below

    however LTC4125 still failed to detect the Receiver

    Did it not satisfy the condition of 1.ΔVFB(n) > 2(Δ VFB(n-1))+1bit?

    I attached the full test waveform,please check the page 0118AXLS