This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS2546: How to control the power switch before TPS2546/IN

Part Number: TPS2546

Hi,

There's a request to supply TPS2546/IN differently with 2 scenarios as below.

  1. When no device is connected to USB port, TPS2546/IN is supplied by a 5V/100mA LDO.
  2. When a device is connected to USB port,  TPS2546/IN is supplied by a 5V/3A BUCK.

The question here is, what can be used to control the switch between these two power sources?  Considering EN pin is enabled by default, how fast should TPS2546 become capable of supplying the current that the device needs?  If the switch from LDO to BUCK is too slow to supply the current in time, what will be the concerns here? 

We're originally considering to use STATUS pin to choose LDO or BUCK, I'm afraid it's too late.  (P.S.: TPS2546 will be configure in CDP mode for this case.)

Thanks!

Antony

  • Randy,

    The TPS2546 will act as a low impedance switch in this case, so the two devices will have to be diode or'd together on the input side of the TPS2546 with a make before break type operation so the supply never droops. They can use the ILIM_SEL pin to insure that the down stream power consumption does not brown out the system when switching.

    The most difficult thing I can see in this system is the control logic which I do not believe is as simple as using the STATUS output.

    Regards,
    Chuck
  • HI Chuck,

    It looks to me there's something need to be fulfilled as below.

    1. EN should always stays high.
    2. IN should always have a valid 5V even during the switching of two power sources
    3. There should be a mechanism to control to choose the power source.

    As i understood, CTL0/CTL1/CTL2/ILIM_SEL would stay the same value during this scenario.  Besides, I can't think of any good mechanism to choose the power sources under such scenario.  

    Do you have any thought to fulfill this especially the control of these two power sources?  It could be acceptable if there's a need to add some external circuit to accomplish it.   Please let us know if you have any idea about it.

    Thanks!

    Antony

  • Antony,

    The three conditions you stated are required, but if ILIM_SEL does not shift, then there is now way to prevent the connected device from browning out the system.

    Here is what I see as the requirements for the power control system:

    1. Determine that a switch between the two power sources is necessary. This may be the Status pin from the TPS2546 if all conditions desired.
    2. Enable the new power source (Both power sources are enabled at the same time)
    3. Wait long enough for the new power source to be fully operational
    4. Change the current limit to the new target.
    5. Wait long enough for the current limit to settle (This is almost instantaneous)
    6. Disable the appropriate power source.

    The most difficult thing about this will be determining when to switch. A simple RC time constant driven into a R/S or JK Latch can set the delay for each change of state.

    Regards,
    Chuck
  • Hi Chuck,

    Now customer consider to implement this feature differently, and decide to choose the Power Wake feature described in datasheet (8.3.6.1 Implementing Power Wake in Notebook System).  We have some questions at this moment,

    In datasheet "8.3.6.1 Implementing Power Wake in Notebook System", the Power Wake feature is implemented with .       CTL1/CTL2/CTL3/ILIM_SEL configured as "0011"--> Power Wake Limits.  Customer try to implement this during S5 mode, but found the STATUS behavior is not as expected.  It's always low even when no device connected.  In parallel, customer implement it with these 4 pins as "0111" --> Non Power Wake Limits, the behavior looks normally.  Here're some related questions.

    1. Regarding the test result, is it because "Power Wake Limits" is much smaller, and there could be some small current even when nothing is connected??  So, STATUS keeps low at this moment?
    2. Can customer implement "8.3.6.1 Implementing Power Wake in Notebook System" with these 4 pins configured as "0111" instead of "0011"

    Thanks!

    Antony

  • Antony,

    The table you have shows that the Power Load Management function in Figure 33 is in place for the first 3 modes and is functioning as expected.

    1. in the first half (0011), if more than 45mA drawn, then status will be low and the port will be in charge mode. It does not make use of the USB data lines in this state, so it is possible to have the bus powered on a disconnected port.

    2. In the second half (0111), Load detection based on USB datalines so their state will determine the discharge/charge connections, so it will transition to high in 2 conditions to tell they system that it is ok to switch to a lower current regulator because the draw is below 45mA.

    If the customer has a MCU to interpret the results and switch the appropriate regulator, then they are able to use 0111 for their system.

    Regards,
    Chuck
  • Hi Chuck,

    For either case of "0011" or "0111", both are DCP_AUTO with data lines disconnected.  Why did you mentioned the load detection is based on USB datalines for "0111"?

    For case of "0111", the load detection is based on "non Power Wake limits" which is around 700mA.  Does it mean STATUS will be driven high when the currnt drawn is less than 700mA?

    THansk!

    Antony

  • Antony,

    If you look at figure 8.3.6.1, it shows using the 0011 case for Power Wake with a controller.

    You table makes sense for this case.

    The LDO should be selected to keep VBUS present on a cable that is disconnected (less than 45mA) (No Pllug)
    The DC-DC converter is selected when the current requirements are above 45mA (50%-60% charged Iphone)
    The LDO should be selected to keep VBUS present on a cable that has a light load (95%+ charged Iphone)

    The Status line is low only when the DC-DC is required in 0011 mode.

    When you select 0111 mode, the TPS2546 expects some other element in the system to manage the power mode, so status is always held low to turn the DC-DC on.

    From what you have described, your customer should use 0011 mode for their system and then manage how to switch between the LDO and DC-DC with a simple MCU or other mechanism to handle the power source switchover.

    Regards,
    Chuck