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.

LMG1210: HO voltage drop

Part Number: LMG1210
Other Parts Discussed in Thread: UCC27200

I am using LMG1210 to drive EPC2045 in a converter similar to buck. The internal LDO was bypassed and 5V is provided by a bench power supply. BAS170W was used as the bootstrap diode. The switching frequency is 3 MHz.

I ran the prototype and noticed the HO voltage is smaller than LO voltage by ~ 1V. Was it caused by the forward voltage drop of bootsrap diode? Is there anything else I could change to reduce this voltage drop?

Thanks

  • Cong,

    Thanks for reaching out!

    -------

    If your measurements are referenced to GND, the HO voltage should always be larger than LO (assuming you are switching with enough speed and duty cycle), to allow for the bootstrap capacitor to charge up enough.

    -------

    One question, do you also have a bootstrap capacitor or one that is appropriately sized?  Typically 0.1uF is a good start.

    If this is not your problem, please follow up with a schematic so I can better analyze your circuit.

    Thanks!

    Aaron Grgurich

  • Hi Aaron,

    Thanks for your quick response. I was measuring the gate-source voltage of the high-side switch which was labeled as HO voltage in the prior post. Sorry for the confusion. I measured the voltage across the bootstrap cap and did not see any significant voltage drop across the bootstrap cap. The voltage ranged from roughly 4 to 4.1V. The theoretical cap value is less than 10 nF following the guidance from the 1210 datasheet. The bootstrap cap was then sized appropriately as 0.1 uF. I also tried 1 uF and I got the same test results. 

    1. Could you give guidance for estimating the voltage drop caused by the boostrap diode?

    2. I am happy to share the schematic with you if you don't mind take a bit more time to digest the design. Could you provide an email address?

    Thanks!

    Cong

  • Cong,

    Okay, so it sounds like things are set-up correctly.

    -------

    1)  Regarding the bootstrap diode, voltage drop; this will depend on the current being drawn through it.  Below is a screenshot from the BAS170W diode datasheet.  The current draw may be pushing your voltage drop higher; you could consider using a "better" diode (with less drop) for your application; but know this often comes with tradeoffs.

    To figure out how much current is being used, please reference Section "8.2.2.5 Computing Power Dissipation" of the LM1210 datasheet.

    You will need to reference the datasheet of the FETs you are driving to get the gate charge spec.

    I expect the current draw is the reason you are seeing that drop in voltage.

    -------

    I hope this helps and answers your question.  Please press the green button if this answered your question, otherwise feel free to follow up.

    Thanks,

    Aaron Grgurich

  • Hi Aaron,

    This helps a lot! My calculation shows Ifet,g = 15 mA and results in a voltage drop of 0.9V. I am planning to switch to a different diode BAT46WJ which has a voltage drop of 0.4V with 15 mA. 

    If I prefer to have a 5 V on-signal for smaller Rdson, is there an alternative solution to offset the diode voltage drop?

    Thanks,

    Cong

  • Cong,

    Glad I could help!

    -------

    Regarding getting a higher HO gate voltage with the LMG1210, the only way is getting the lowest voltage drop Schottky diode you can get.  (which you have just done)  Schottky diodes are typically the way to go for this application due to their inherently low forward voltage.

    The only other way is to use a different half-bridge driver that does not restrict you to the internal 5V LDO of the LMG1210.  Depending on your application, you could look at other half-bridge drivers such as UCC27200.  The gate drive voltage is based on the supplied VDD to the IC.

    -------- 

    I hope this helps!

    Aaron Grgurich

  • Hi Aaron,

    (1) I am looking at Fig 20 in the LMG1210 datasheet. I see both HO and LO are 5V. The bootstrap diode caused little voltage drop. May I know what circuit it is and which diode is used?

    (2) In the same datasheet, could you define the rising/falling threshold in the section of Undervoltage lock on page 5?

    (3) I am a bit surprised that the Schottky diode also hurts the HO (gate-source voltage). The best Schottky diode, to my knowledge, causes a 0.3V voltage drop for driving a GaN in a practical application. This means that high-side GaN can never see a 5V Vgs which is the optimum voltage level for minimizing Rdson. Do I miss anything here?

    I appreciate your help!

    Cong

  • Cong,

    1)  The circuit used is likely our EVM (evaluation module) for the board, part of the circuit is shown below.  The user guide and ordering details for the EVM can be found on the LMG1210 product page.

    The diode used is the BAT46WJ,115.  It does not seem to be significantly better than the one you have.

    2)  We have a technical article on Undervoltage Lockout here.  The rising threshold is what voltage is needed at the specified pin to keep the device "happy" and operate as it should.  The falling threshold is what voltage at the specified pin will cause the device to go into "lockout" and essentially shutdown, to protect the power transistors down stream.  The delta or difference between these two voltages is known as "hysteresis", a topic worth looking into.  It is basically a "buffer" between the two thresholds so that if the voltage is near one of the thresholds, the device does oscillate in between the two states.

    3)  Yes, that is an inherent characteristic of the bootstrap topology.  The approximate 0.3V voltage drop is accounted for in the datasheet here:

    -------

    I hope this clarifies things for you!

    Thanks,

    Aaron Grgurich