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.

TPS2412: How to improve the turn-on speed on TPS2412 oring controller?

Part Number: TPS2412
Other Parts Discussed in Thread: CSD17570Q5B, TPS2413

Hi TI,

We have two VR for P5VA and P5VB and this design required redundant feature so we add two TPS2412 for oring control without active current sharing circuitry there due to PCB space limitation.

Then we found the Oring gate of P5VA keeps low state at no load or light load condition due to bad current sharing, only P5VB oring enabled.

If we perform the internal short at P5VB rail, the P5V (after OringFET) will drop to about 4.3V (5V-oringFET Vf) because the P5VA oring gate needs about 6ms~15ms ramping time depends on different loading.

After removing the bypass cap (0.22uF) on G/S pin of OringFET and short Rg (10 ohm), the ramping slop is improved from 1V/ms to 12V/ms but still cannot keep P5V within regulation.

Is there any way to improve the turn on speed faster by the setting of TPS2412? or it's limited by the selected OringFET?

Note: The OringFET is CSD17570Q5B * 2pcs

BR.

Stan.

  • Hi Stan

    0.22uF cap of G/S of MOSFET is huge.
    Removing it will definitely help to increase the turn on time of V5A gate.

    If you are not requiring current sharing, TPS2413 will perform better here than TPS2412.
    TPS2412 regulates the Vds of MOSFET to 20mV, but TPS2413 is ON/OFF hysteretic control.

    TPS2412 has turned off because it is trying to regulate the gate voltage at light load (V5B is slightly higher voltage than V5A).
    Replacing with TPS2413 will keep TPS2413 ON and avoid the drop in V5.

    Note that TPS2413 allows DC reverse current depending on the RSET setting and this needs to be considered if this is acceptable.

    Other options is use LM74700 which has a higher gate drive 11mA and gate control to regulate 20mV.

    Further increasing the Capacitance on V5 rail will help to reduce the drop till the V5A come back up.

    Regards,
    Kari.
  • HI Kari,

    Many thanks for your reply. Read from datasheet, the TPS2413 also requires Vac>10mV to turn on the gate.

    Because of the Vout deviation on 5VA and 5VB that caused by component tolerance and Vdrop by PCB trace.

    I think maybe we will face the same problem that either 5VA or 5VB cannot enable the oring gate even using TPS2413.

    Do you agree with my assumption? Or you're saying both TPS2413 will turn on the gate during first power on and keep both gate at ON state?

    BR.

    Stan.

  • Hi Stan,

    With TPS2413, once it is turned on it requires negative threshold to hit to turn off.

    Both the TPS2413 will turn on initially and then if the difference in power supplies is more than (20mV + reverse threshold), then 5VB will turn off.

    We suggest to use LM74700 or LM5050 because gate drive differences.

    In LM74700 and LM5050 gate is turned off by shorting Gate to source. Whereas in TPS2412 it is turned off by shorting gate to GND.

    When turning back ON, TPS2412 has to charge the gate all the way up to VIN + 12V. (TPS2413)
    With LM74700 or LM5050, gate can turn back on quickly since gate is parked at VIN level and it needs to go above few volts to turn on MOSFETs.

    LM74700 also has much higher gate drive of 11mA to turn on gate quickly.
    Increasing the capacitance at the 5V output will also hold the 5V from collapsing.


    Regards,
    Kari.
  • Hi Kari,

    Understood and thanks. I will try TPS2413 first and also study LM74700 and LM5050.
    Thank you.

    Regards.
    Stan.
  • HI Kari,

    I tried TPS2413 today but still got the same symptom. 5VA oring was off when 5VB oring ON during start-up. I sw this waveform when we using TPS2412 on our board.

    So the key is the start up period? Or we need to add an active current sharing design to regulate both VR has the same vout that could make sure TPS2412/3 can turn on both oringFET at no load?

    BR.

    Stan.

  • Hello Stan,

    Yes, key is the startup, having load sharing will definitely help. This will increase cost and space.

    So our suggestion would be to use LM74700 or LM5050 where the gate is not pulled all the way down to ground, instead gate is pulled to source to make gate-source 0V. And source is same as anode or IN pin, so gate will stay close to IN and the time to ramp up during turn on will be less.

    Could you consider this option? Additionally you may have slightly increase the capacitance after LM74700 or LM5050 to hold the output untill second channel is turned on fully.

    Regards,
    Kari.