N Mosfet Choice

Discussion in 'General Electronics Chat' started by chuttney1, Nov 12, 2014.

  1. chuttney1

    Thread Starter New Member

    Nov 12, 2014
    So a similar question was posted about two years ago due to a vacuum tube amp build found on from this link.

    I am about to build this headphone amplifier. The IRF510 mosfet used in this build has clipping issues at higher volumes. I checked the support thread to see if there is any updates over any component changes since the first write up. Reading the comments on page 137-138, someone posted they used a Hitachi N Mosfet part of the 2SK213, 2SK214, 2SK215, 2SK216 instead of the IRF510 due to a Vgs of 2 V and solved the issue of the clipping and now can go to 100% volume.
    Link to 2SK213

    Obviously the part isn't made by Hitachi anymore and found on ebay. Searching for a replacement on Mouser similar to the 2SK213, I have a STM STP5N52K3 with a Vgs of 3 V. Close enough but seem like it will do.
    A few comparisons:
    IRF510: Vds breakdown voltage: 100 V, Vgs, breakdown: 20 V, Rds On: 540 mOhms
    STM STP5: Vds breakdown voltage: 525 V , Vgs, breakdown voltage: 30 V , Rds On: 1.5 Ohms
    2SK213: Vds breakdown voltage: 140 V, Vgs breakdown voltage: 15 V, Rds On: not stated.

    One thing to note is the internal schematic of the mosfet from Hitachi and STM nearly similar are the similar. I would like to know what the two arrow with the squiggly lines mean. The main question is how feasible is using the STM STP5N52K3? The goal is to have a low Vgs, but the 510 V Vds seems a little to high for me to know as a total beginner in Mosfet selection.
  2. alfacliff

    Well-Known Member

    Dec 13, 2013
    it is very difficult to bias mosfets like that tinto a linear regeon for good amplification. that mosfet is designed for switching high power. you will have to watch your parts layout to avoid rf oscilation as you vary the bias, that mosfet can operate into the low vhf regeon easily.
  3. Jony130

    AAC Fanatic!

    Feb 17, 2009
    Simply use IRL510 and you will be fine
  4. wayneh


    Sep 9, 2010
    The solution is not to change the MOSFET, it's to limit the input voltage range of the signal on its gate. Using an old MOSFET like the IRF510 may actually help you if it has a wider range of gate voltage between "on" and "off". A sharper transition would narrow the voltage range in which the signal has to fit.
  5. Jony130

    AAC Fanatic!

    Feb 17, 2009
    But if you set the DC bias voltage at IRF source into 0.5Vcc (6V) then the plate voltage will be Vgs higher than the IRF source voltage. So the DC plate voltage is around 9V or 10V. So the tube will start to clipping first.
  6. wayneh


    Sep 9, 2010
    Ah, that makes sense. So lowering the MOSFET's gate voltage gives more headroom at the tube?
    I was thinking the clipping was at the MOSFET. Thanks for clarifying.
  7. bertus


    Apr 5, 2008

    The circuits uses a constant current source of about 125 mA (the LM317 with the 10 Ohms resistor).
    You will have to adjust the 50K so that the output voltage is a bit above 1/2 the powersupply voltage.
    (ca. 8 Volts for maximum swing, as the LM317 + the 10 Ohms resistor will take 4.5 Volts from the bottom).
    In my opinion the mosfet should have a small heatsink, wich I do not see on that site.


  8. Jony130

    AAC Fanatic!

    Feb 17, 2009
    Yes, but not much. Also the circuit will be clipping do to very large headroom voltage for LM317.
    LM317 is not the best choice here.

    The power dissipation in the mosfet will be around 0.8W. So the the mosfet will survive without the heat sink. Of-course the metal tab will be very hot but for sure no magic smoke will appear.
    I tested this in practice.