Shop OBEX P1 Docs P2 Docs Learn Events
in which cases would you prefer a bipolar transistor over a MOSFET like IRF3708 — Parallax Forums

in which cases would you prefer a bipolar transistor over a MOSFET like IRF3708

StefanL38StefanL38 Posts: 2,292
edited 2015-05-05 00:47 in General Discussion
Hi everybody,

lately I used the MOSFET IRF3708 to switch 12V-Relais. Worked out just fine.
The IRF 3708 needs only about 2V to switch on. So a 3.3VµC can drive it directly.

So I will use this MOSFET more often in future. Anyway I would like to know in which cases
would you prefer a bipolar NPN-Transistor instead of a MOSFET?

best regards

Stefan

Comments

  • tonyp12tonyp12 Posts: 1,951
    edited 2015-05-04 14:44
    I never used a transistor, transistors are more analog in nature, so could be better for voltage or current control. (a lot of heat though)

    always look at the mosfets Vgs chart, I like it to be at least 80% fully on at 3.3V
    the IRF3708 @ 3.3v gate: ~80Amp and ~0.011ohm resistance (this is a power mosfet and not small signal mosfet obviously, but need 4.5Vgs to be fully utilized)

    here are good mosfet relay drivers:
    http://www.mouser.com/Semiconductors/Discrete-Semiconductors/Transistors/MOSFET/_/N-ax1sfZscv7?Keyword=relay
  • kwinnkwinn Posts: 8,697
    edited 2015-05-04 14:47
    I use bipolars (mostly 2n3904 or 2n2222) to drive leds with a constant current. That way I can power the led from the unregulated voltage and reduce the current through the power supply regulator. Good for SSR's, indicators, displays, and optoisolated current loop comms.
  • evanhevanh Posts: 15,920
    edited 2015-05-04 15:06
    BJTs are fine for low current digital switches, they don't have static concerns like MOSFETs do. For analogue power, as Tony indicates, BJTs have better linear response. Both are used in both situations, no biggie. Sometimes it's nice to have the diodic base of a BJT, other times it's nicer to be controlling the floating gate of a FET.

    But then there is IGBTs, they shine at higher voltage switching, ie: PWM on mains. I don't know their exact construction but functionally they are supposedly a MOSFET driving a BJT, arranged a bit like the darlingtons of old but much much faster because they don't saturate.
  • evanhevanh Posts: 15,920
    edited 2015-05-04 15:36
    BTW: With regard to your low volt gate control, when it comes to making a MOSFET switch higher currents you really do need to drive the gate a lot further than 2 volts, ie: The higher the current being switched the higher you'll need to lift the gate voltage.
  • StefanL38StefanL38 Posts: 2,292
    edited 2015-05-04 21:19
    thank you all for answering and commenting.

    I took the IRF3708 because a longer time ago this type was mention as switches on at low voltage.
    I studied the Vgs-Diagram. I think as long as I stay in a range of 1-5A 3,3V will do.
    It would be much easier to decide if the diagramm drain-current over On-resistance would have more curves for lower voltages.

    best regards

    Stefan
  • jmgjmg Posts: 15,173
    edited 2015-05-04 22:33
    StefanL38 wrote: »
    Anyway I would like to know in which cases
    would you prefer a bipolar Transistor instead of a MOSFET?

    Price.
    We use a BCP53 as a PNP current source, at a price P-FETS cannot touch, not to mention a BCP53 can bias from a BAS31 dual diode, can't see any PFET doing that...
    So it depends on the application....
  • evanhevanh Posts: 15,920
    edited 2015-05-05 00:05
    Another example is if you examine the schematics of various op-amps you'll find the fast ones are pure BJT while the extreme sensitivity ones have JFET fronts to a BJT drive and for very low power designs pure MOSFET is the norm.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2015-05-05 00:47
    MOSfets are wonderful for switching a lot of power without much waste -- ideal for brushless motor control and H-bridge designs.

    The primary drawback is procurement of the right devices. And just about anything you desire is going to be surface mount.

    For merely driving relays, BJT transistors can do fine. Most relay coils require something less than 100ma and the 2N2222 will easily handle that. And these devices remain available in TO-92 packaging, procurement is easy and cheap.

    But the primary drawback is much more wasted energy due to a 0.7 volt drop from voltage input to voltage output, plus the heat generated in consuming that 0.7 volts. (this can in some cases go to 1.0 volts for full saturation under high current). That just isn't good for battery operated robotics.

    Darlington transistors are just about the worst choice one can make -- the voltage drop is doubled to between 1.4v to 2.0 volts, and the heat is even more problematic as paired transistors are combined in one package. There are some super-beta transistors that will provide similar high gain, but a bit harder to acquire.

    ++++++++++
    Conclusion, I personally will use transistors to drive relay coils and prefer 12VDC relays over lower voltage relays to minimize the issue of voltage drop. Meanwhile, I tend to hoard my MOSfets fro H-bridges and other power hungry applications where getting the most out of battery life is important. MOSfets can be used in parallel to get even bigger power outputs; whereas, BJTs will self-destruct.
Sign In or Register to comment.