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.

LM3409HV: Using two or more MOSFETs in parallel

Expert 1350 points
Part Number: LM3409HV
Other Parts Discussed in Thread: LM3409, TPS92640, TPS92691

Hello,

I had posted e2e.ti.com/.../2550493#mce_temp_url# about using the LM3409HV for driving upto 10A. I built that design and while testing it I noticed that the MOSFET gets really hot. So much so that it begins to melt the solder used to hold it to the board. Can I use two or more MOSFETs (of the same part #s) in parallel with this design? 

  • Hello,

    Yes but you may run into issues with the strength of the gate drive. You also want to use individual gate resistors to each MOSFET so it helps balance switching loss (Vgth varies with temperature and can shift all the switching loss to one part if the gates are tied together). If you are using a large Qgate MOSFET already it could be contributing to your thermal issues.

    Regards,
  • Hello,

    Thank you. The data sheet specifies a 'Vin to Vcc, Pgate' value as 7V and the Icc-lim as 30mA. Hence the gate resistors would be 7V/30mA = 233R. Is this calculation correct? 

  • Hello,

    Let me know the part number of the MOSFET you are using and your switching freqeuncy (highest switching freqeuncy since it varies with Vin and LED stack voltage). The gate resistor does not want to be that large, I would consider 10-20 ohms but it depends on the MOSFET being used.

    Regards,
  • Hello,

    The part number is FQP17P10 and the datasheet is available here: http://www.onsemi.com/pub/Collateral/FQP17P10-D.pdf. The highest switching frequency this would be designed for is 750 KHz and the highest VIN is 70V.

  • Hello,

    A few more questions;

    What is your output voltage?

    What is your minimum Vin?

    Switching at 750 KHz will be difficult with this design unless you add an external gate driver.  The switching loss will be very high due to the limited gate drive.

    Regards,

  • Hello,

    Output voltage is 5.2V. Minimum Vin is 7.4V (using a battery pack but this can be increased to 11.1V by using a 3S-xP pack). The switching frequency in the design mentioned in the earlier related question was about 500 KHz. What is an appropriate switching frequency? The data sheet mentions that anything above 1MHz is not viable. Adding an external gate driver should not be a problem as far as BOM count or PCB space is concerned. VIN range might reduce though. Can you suggest any? Perhaps something like the PMD3001D or the NUD3160?

  • Hello,

    Correct me if I'm using incorrect information:

    Vout = 5.2V

    Vinmax = 70V

    Vinmin = 7.4V

    Alternate Vinmin = 11.1V

    Iled = 10A

    Fsw = 500 KHz

    Highest Fsw = 750 KHz

    This will be a difficult design because at the low end Vin the duty cycle will be (5.4V + 0.8V)/7.4V = 84%, this doesn't include Rcs drop or other IR drops in the design so conduction loss will be the key design parameter at this input voltage.  If the RDSon was 0.14 ohms, it will double when hot (150C) the conduction loss at this operating point would be 10A * 10A * 0.84 * 0.14 ohms * 2 = 23.5W.  That is with one FQP17P10.  This is with VGS at -10V, it will be worse with the LM3409 gate drive voltage.  The other thing to note in the datasheet is the VGS threshold.  At -16.5A it's above -6.0V.  Since that is close to or at the gate drive voltage of the LM3409 the turn-on switching loss will be very high which will be a problem when Vin is 70V.

    Are you sold on using the LM3409 or would you consider looking at the TPS92640/641?  It is a synchronous part that uses N-channel MOSFETs which are much better for low RDSon at 100V.  There also is a high current design example using this part.  There are other parts that would not be synchronous that can use N-channel MOSFETs that would be better as well.

    If not and you want to use the LM3409 across the entire operating range above I would suggest lowering the switching frequency significantly (meaning larger inductor).  The tough part with the P-channel MOSFET is the wide input operating range, needs to be a -100 rated part but also can reach 90% duty cycle.

    Regards,

  • Hi,

    Yes those are the starting specs.

    Yes I can see that using the LM3409HV may not be very conducive. You suggested using an external gate drive. Would that help?

    I will take a closer look at TPS92640/1. Thank you.
  • Hello,

    Yes, an external gate driver would help but does that make sense? Three FQP17P10s would get the power dissipation down for conduction loss but the high Vgs threshold will still make the turn-on switching loss high. The gate driver would probably have to run from an external generated VCC to get the voltage up some but that gets difficult with your lowest Vin, or, find a different P-channel with a lower Vgs threshold. There are better parts than the FQP17P10 that may help with this.

    Regards,
  • Thank you. I was looking at the TPS92691. Will this part be suitable for the current specs?
  • Hello,

    Appears there are MOSFETs that would work for this since it's an N-channel driver. I would design it to operate at a lower switching frequency since switching loss will dominate when Vin is high in your design. There are 20-40 nC gate charge N-channel MOSFETs with RDSon in the 20-40 mohm range that should work depending on heatsink (thermal), the switching frequency and the diode choice. You would have to go through the calculations to see what the conduction losses and switching losses would be for your design however, compared to the FQP17N10 at 140 mohm typical and 30 nC total gate charge the MOSFET is much better for this application. High current high duty cycle at low Vin, high current low duty cycle but high switching loss due to 70V input.

    Regards,