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.

  • TI Thinks Resolved

TPS7A92: parallel operation for increasing output current

Genius 10235 points

Replies: 6

Views: 280

Part Number: TPS7A92

Hello,

The output current is 3A and we want to use two TPS7A9201s in parallel due to some reasons.

TIDA-01232, ballast resister solution, seems to be able to achieve it. Is TIDA-01232 good way for this purpose?

Best regards,

Toshihiro Watanabe

  • Hello,
    I tried to calculate the ballast resister value. Vref has 1% error. resister assumes 1% tolerance. I think Vout = 5.14V would be max voltage. When we tried to balance the current two LDO, the ballast resister seems too big, like 1.4ohm.
    Please give me your advice if I can think correctly.
    Best regards,
    Toshihiro Watanabe
  • In reply to Toshihiro Watanabe:

    Hi Toshihiro-san,

    With the app note equations, I calculated a ballast resistor of 0.3 ohms, but it was still too large to be useful. The author of TIDA-01232 also calculates a large ballast resistor, but then uses measurements to justify a smaller EVout value and calculates a smaller ballast resistor.

    If we do the same as TIDA-01232 equation 5 and use Vref tolerance = 0.094% and 0.1% feedback resistors, I calculate EVout = 6.37 mV and Rballast = 29 mohms. Now the voltage drop is 0.044 V across the ballast resistor. This will work as long as the actual output tolerance is close to 0.094% rather than 1%.

    Larger ballast resistors allow LDOs with wider output tolerance to work together in parallel, but a value of 0.3 ohms is not feasible because the voltage drop across the ballast is too high to maintain good output regulation.

    The 1.4 ohms you calculated would definitely be too big. At 1.5 A through each ballast resistor, the voltage drop would be 2.1 V!


    Further detail and calculations from app note:
    From TIDA-01232 equation 1 and 4, VoutNom = 5 V, Vref tolerance = 1%, Vref = 0.8 V, ERfb = 1% (1% resistors),
    EVout = 66.8 mV.
    Rballast calculated from equation 2 with EVout = 66.8 mV, Ilim (typ) = 2.6 A:
    Rballast = 304 mohms (too large)

    If 1.5 A is flowing in each Rballast = 0.304 ohms, there will be a voltage drop of 0.456 V and 0.684 W dissipated in each ballast resistor. This does not look like a good design solution.

    If you look at TIDA-01232 equation 5, there were some measurements used to replace the EVout calculation to achieve a much smaller Rballast. Instead of the datasheet 1% Vref error (8 mV), the author used 0.094% (0.75 mV).

    Best,
    Eric Preiss | Validation Manager | Linear and Low Dropout Regulators | Texas Instruments

  • In reply to Eric.Preiss:

    Hello Eric-san,

    I think TIDA-01232 is based on slva250 application. If TOLRFB or TOLVREF is 0.01(=1%) to use, I can get 0.268ohm(near 0.3ohm) as your calculation.

    As you mentioned, the Author seems to measure data and find a correlation then use smaller variation but we don't have time to do this. 

    One question. Denominator of equation 5 in TIDA-01232 is 2x Ioutmaxsingle - Ioutmaxtotal. Do you think it becomes 3x Ioutmaxsingle - Ioutmaxtotal when 3 device current share?

    In this case, ballast resister becomes 0.1ohm and each current is 1A, Vdrop_ballast is 0.1V, which is still a bit high drop but may be acceptable.

    Anyway, I will ask if the customer can separate the load by two then use one LDO per each blocks. If possible, no ballast resister and it's loss can be removed.

    Best regards,

    Toshihiro Watanabe

  • In reply to Toshihiro Watanabe:

    Hi Toshihiro-san,

    Would the customer consider using a single TPS7A85A (4 A LDO) to provide the 3 A load? Alternatively, separating the load and using two separate LDOs is a good option.

    Best,
    Eric Preiss | Validation Manager | Linear and Low Dropout Regulators | Texas Instruments

  • In reply to Eric.Preiss:

    Hi Toshihiro-san,

    I suggest using a different parallel connection for 2 LDOs as described in this TI app note:
    www.ti.com/.../tidu421.pdf

    This method is less sensitive to differences between the LDOs, and smaller series resistors < 0.030 ohms are used.

    Best,
    Eric Preiss | Validation Manager | Linear and Low Dropout Regulators | Texas Instruments

  • In reply to Eric.Preiss:

    Hello Eric-san,
    Thank you for your suggestion. I will suggest the manner which TPS7A85 can use -> Load sprit -> TIDU421.
    Best regards,
    Toshihiro Watanabe

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.