LTC4155 not charging.

Welcome.

We are using LTC4155 on our prototype PCB.
We have two problems:
1) LTC4155 is not charging battery (0mA) most of the time
2) LTC4155 sometimes kicks in and charge battery at 75mA current, even if we set 1000mA charging current and no input limits

We are using 1,5A 5V wall adaptor as power supply. Measured voltage on WALLSNS and VBUS is 5,35V. Almost no current drawn from PSU.
We are using 3,7V 4000mAh Li-Poly battery with 10k NTC. Actual battery voltage is above 3,8V (measured on BATSNS).

Register readings:
(0x0 means REG0, 0x1 means REG1, etc)

0x0: 10000000
0x1: 00000110
0x2: 00001100
0x3: 00000001
0x4: 10100000
0x5: 10000000
0x6: 00000000

As you can see, on REG3, LOWBAT is set to 1, so probably that is reason why there is no charging at all.

We are waiting for your help.
Regards Mateusz

  • 0x0: 10000000
    0x1: 00011111
    0x2: 11111100
    0x3: 00000001
    0x4: 10100000
    0x5: 10000000
    0x6: 00000000

    This is correct reading, sorry.

  • 0
    •  Analog Employees 
    on May 22, 2019 3:39 PM

    Hi Mateusz,

    Since you emailed me as well, I answered you there. I am copying my advice here for anyone who sees this. In general, please choose one support resource or the other. Let's keep this conversation here in case it helps others.

    Initially, you confirmed that you had only assembled one board:

    In general, I recommend building up at least 3 boards or so to ensure this is not an assembly error. I would not rule that out at this point.

     That said, the circuit layout is definitely going to cause some problems. The biggest issue is the grounding due to the fact that this was done on only two layers. I recommend that you use 4 layers and follow the demo board example layout.

     C20, for example, should be placed with short routing between the VBUS and the GND pad. The trace from VBUS is pretty short, but look how far it needs to go before getting back to GND.

     

    A few general tips:

    - If possible, dedicate a whole layer to GND. We generally use layer 2. This becomes important when dealing with power.

    - All power components should have good return paths to the IC paddle.

    - Do not route anything directly below the inductor, shield it with GND.

    - Do not use thermal relief on power components.

    - Caps that input to the switcher (C20) should have very short routing between the switcher power supply (VBUS) and the power GND return (the IC GND paddle).

     

    For further info:

    1. I recommend you read this: https://www.analog.com/media/en/technical-documentation/application-notes/an136f.pdf
    2. Consult the Board Layout Considerations section of the LTC4155 datasheet.
    3. Reference the demo board layout files.

    Regards,

    Zack

  • Thank you Zack.

    Are you sure that there are no other "hard" mistakes?

    I do nothing and not LOWBAT has dissapear and now my PCB is again charging only at 75mA.
    I know there was 75mA issue with LTC4155 reported here, but no clear resolve.
    Look at the 0x4 register. 


    Actual readings:
    0x0: 00000000
    0x1: 00000110
    0x2: 11111100
    0x3: 01000000
    0x4: 10101000
    0x5: 01111100
    0x6: 00000000

  • 0
    •  Analog Employees 
    on May 22, 2019 5:38 PM in reply to MateuszKusmierz

    The VC capacitor is way too large, so that may cause some issues, but I would guess it would still regulate.

    I can't be sure of anything until a second board behaves in the same way. I recommend you assemble a second one.

    Bad grounding can cause noise which can cause logic issues, so it is possible that this is the reason for your bad 0x04 reading.

    Regards,

    Zack

  • Zack, thank you. I checked if grounding was ok, and it was not!
    Probably assembly mistake as you said.
    Ground pad was not soldered properly. And that is why chip was seeing low voltages everywhere.

    We repaired it, and it works PERFECTLY.

    We will redesign PCB also.

    Thank you!