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.

LM5122: Output lockup and stand by lowest power consumption

Part Number: LM5122

Hi all, I have make a circuit need to use LM5122 for DCDC boost from Vin=15~21V to Vout=30V with 7A output, I just put those information to web bench design and generate the schematic to follow the component value as shown below:

Unfortunately, I have face problem:

1. At the beginning, I had found that more spike and noise at the output voltage especially when large load, make a fault trigger to the current sense so very early off the LM5122, after I add some ceramic cap. it's seems the noise is much smaller, but I don't know the problem is actually solved or not.

2. Since I am making a product which need to very lower power consumption in idle mode, so I will switch off the LM5122 when no use (by UVLO, or cut the Vin by external circuit), and wake up when some current drawing in the output. I have MCU to turn the LM5122 on if the current sense circuit have signal then trigger the MCU. If no current draw at start and wake up LM5122, everything is work fine. However, I observed when the current consume is larger when sleep, the LM5122 seems wouldn't wake up and the input is over current, it seems Cout need more power to charge up the capacitor but not success, then LM5122 is shut down, and wake up again and try to charge the capacitor....looping. Beside, M1 temperature is rising up. Where is the problem occur?

3. In previous talk, I need low power consumption in idle mode, I find that most current consume in sleep are Rfbt and Rfbb, leakage from the output, In addition to the LM5122 current is leakage even UVLO is off(so I cut the LM5122 Vin when the product sleep), I just attached my modification as below, is this reasonable? or any other issue need to pay attention?

Thanks so much for answer my question. 

  • Hi Simon,

    Thank you for reaching out to TI for help. Please see my comments below.

    1. At the beginning, I had found that more spike and noise at the output voltage especially when large load, make a fault trigger to the current sense so very early off the LM5122, after I add some ceramic cap. it's seems the noise is much smaller, but I don't know the problem is actually solved or not.

    Please make sure that you are probing the output voltage correctly. A short ground lead on the probe should be used to minimize high frequency noise. Adding ceramic capacitors will help to reduce this high frequency noise as well. Here is a good blog about how to measure output ripple accurately.  

    2. Since I am making a product which need to very lower power consumption in idle mode, so I will switch off the LM5122 when no use (by UVLO, or cut the Vin by external circuit), and wake up when some current drawing in the output. I have MCU to turn the LM5122 on if the current sense circuit have signal then trigger the MCU. If no current draw at start and wake up LM5122, everything is work fine. However, I observed when the current consume is larger when sleep, the LM5122 seems wouldn't wake up and the input is over current, it seems Cout need more power to charge up the capacitor but not success, then LM5122 is shut down, and wake up again and try to charge the capacitor....looping. Beside, M1 temperature is rising up. Where is the problem occur?

    If the UVLO pin is pulled low the VIN pin should only sink about 17uA (worst case). I think that it would be simpler to pull the UVLO signal low then disconnecting the VIN pin. It should like when you start-up under heavy load that you are entering hiccup protection mode. This will cause the LM5122 to loop like you are describing. To disable this mode please pull the RES pin to GND. If M1 is heating up this can be due to MOSFET selection. Please ensure that the selected MOSFET is correct for this current level. It might also be a good idea to have a schottky diode in parallel to help reduce dead time losses. 

    3. In previous talk, I need low power consumption in idle mode, I find that most current consume in sleep are Rfbt and Rfbb, leakage from the output, In addition to the LM5122 current is leakage even UVLO is off(so I cut the LM5122 Vin when the product sleep), I just attached my modification as below, is this reasonable? or any other issue need to pay attention?

    The circuit looks reasonable. I would suggest adding a zener diode to clamp the FB pin as it's abs max voltage is only 15V. If the current consumption of the zener is too large please try moving Q1 between R141 and the FB node. This will allow for the FB pin to be protected and reduce the current consumption when the circuit is in it's low power state.

    Please let me know if you have any questions.

    Thanks,

    Garrett

  • Dear Garrett,

         Thanks your speedy reply, I will try to modify based on your idea, for parallel the schottky diode, you mean both FET (Q6 and Q7) need to parallel right?, below picture is correct position of parallel the schottky diode?

    Beside, I check the Mode pin description:

    "When MODE pin voltage is greater than 1.2 V diode emulation mode threshold,
    forced PWM mode is enabled, allowing current to flow in either direction through the high-side Nchannel
    MOSFET switch."

          Does it mean that it will force LM5122 generate PWM without any warning/ hiccup protection? can I submit this modification in my case?

    Thanks!

  • Simon,

    If Q7 is the device getting hot, I suggest only adding the diode across Q7. The position of the diode is correct.

    The MODE pin doesn't control the hiccup mode operation. This is the RES pin as I mentioned above. To disable this mode the the RES pin should be pulled to GND. Please test with the RES pin pulled to GND. See section 7.4.3 of the datasheet for all the details.

    The MODE pin is used to select at light load if the converter will operate in continuous conduction mode (CCM,) or discontinuous conduction mode (DCM, inductor current cannot be less than 0A).

    Please let me know if you have any more questions.

    Thanks,

    Garrett