Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

When to use a BJT or FET?

  1. Oct 1, 2012 #1
    Fresh EE here that hasn't seemed to cover the main application of either a BJT or mosfet.

    When should you use one in place of the other? There doesn't seem to be a clear distinction between the two, however I would imagine there are substantional practical differences.

    Situation in example, ive got a BJT (BD771) handling about 30v at 2A through it. It generates an enormous amount of heat when its fully on. This has me wondering whether a power mosfet is more applicable, since I read they have lower on resistances.

    If anyone has some literature they could point out to me, that would be appreciated.
     
  2. jcsd
  3. Oct 1, 2012 #2

    uart

    User Avatar
    Science Advisor

    Yes there are many differences that can be used to different advantages in different applications.

    - High input impedance and consequent very high current gain are particular advantages of the mosfet that's sometimes very useful.

    - High speed due their minority carrier operation (they don't require carrier injection for their operation) is another mosfet advantage, particularly in switching circuits.

    - Approximating a constant (low) resistance in part of their operating region is another feature of the mosfet that is exploited in certain application.

    - Traditional the BJT was the easiest enhancement mode (normally off) transistor to fabricate, making it cheap and readily available. I'm not sure if this has any relevance today though (I don't know for sure but I suspect not).

    - The BJT inherently has a fairly linear current gain, which is very useful in many applications.

    - The base emitter diode relation of the BJT is sometimes a nuisance, but sometimes can play a useful part in the circuit. I'm sure there's tons of other differences that I could mention. :)
     
  4. Oct 1, 2012 #3

    uart

    User Avatar
    Science Advisor

    It's not necessarily that simple. The power is the product of voltage times current, so the power will depend upon what exact conditions you are measuring it under.

    If that transistor was fully saturated the power dissipation could be less than one Watt at 2A (assuming it's not a darlington). If however you're operating it in its linear region then the power would be very much higher. It is true however that for a low voltage like that (30V) you could quite easily get a similar sized mosfet with a low enough on resistance to beat out that BJT. Essentially anything with an on resistance under about 0.2 ohms is going to outperform the BJT here.
     
    Last edited: Oct 1, 2012
  5. Oct 1, 2012 #4
    There are difference between the two.....obviously. I can only point out a few comes into my mind so chances are I am going to miss some:

    1) First and most important is the input current. BJT is a current device that amplifier the base current by β. So it needs base current. MOSFET on the other hand as the name indicates.."Insulated gate". The gate is really a capacitor and the charge on the gate cause inversion at the channel right under the oxide insulator when turn on. So if you have very high impedance drive that cannot provide current, but only voltage, MOSFET will be a much better device to use.
    2) BJT can go into saturation where both the BE and BC diode both turn on at the same time. When that happens, it will take some time ( into u sec.) to get out of saturation and turn off even if you remove the base drive. This is commonly called recovery time. MOSFET do not have that problem at all, as soon as the Vgs goes below the turn on voltage, the FET turns off. That's one important reason why digital IC today mostly are CMOS logic because the internal speed is so so much faster than TTL which is BJT.
    3) When BJT goes into saturation, it still has about 0.3V across the collector emitter, MOSFET behave like a resistor when fully turned on. The bigger the MOSFET, the lower the Rds on resistance.


    That all sounded good, but there are few catches:
    1) The capacitance of the input of MOSFET can be high if the transistor is large. A power MOSFET can have input capacitance in excess of 3nF. Driving MOSFET require strong drive to get speed. That's the reason why the processor can run GHz internally, but when the signals and bus interface to external logics, the speed is not very fast. Transistor for internal logic is very small, don't have much gate capacitance, so very little drive requirement is required. But when it has to drive external bus, much stronger drive is necessary to sink and source a few mA. The transistor has to be much bigger than those for internal logics and it need much strong pre drive. Sometimes, the internal logics has to be buffered a few stages before gaining enough drive for external bus. That really kill the speed. It is common for power amp to have BJT as a pre driver stage just to drive the big power MOSFET.
    2) Small signal MOSFET has much higher 1/f noise that dominates the lower audio frequency range. The noise can extend all the way into MHz and beyond. Noise performance is almost inferior for regular speed op-amp, much worst for audio circuits.
    3) the Gate source turn on voltage variation between MOSFET even on the same die is much greater than Vbe of BJT, temperature drift is much higher. If you want precision op-amp, BJT op-amp are mostly superior.
    4) Because the gate source turn on voltage vary with temperature, propagation delay of logic change with temperature is much worst than TTL or ECL. For precision timing control, you cannot use CMOS logics....at least up to a few years ago.

    The major problem of MOSFET is the input capacitance that limit the usefulness. But in RF world, people make use of this. LDMOS is very popular in RF power amps. We always use matching network to get best power transfer. The matching network usually have a capacitor right before the input of the power transistor. A lot of RF power BJT are actually a hybrid where they put chip capacitors inside before the base. They have to put it inside because the lead inductance of the transistor become inductive at that frequency and create an inductor after the capacitor if it is outside. The input capacitance of the LDMOS is actually being used as that stage. This will reduce the cost tremendously as you don't need to put chip capacitors anymore.

    MOSFET and new group III and V transistor ( function either like MOSFET or JFET) really taking over BJT in microwave electronics. At microwave frequency, the noise of these devices are lower than BJT as it is beyond the 1/f noise range. All you have to do is using high pass filter to block the lower 1/f noise. At this frequency, BJT has a big disadvantage that it draws base current and thereby has current shot noise that can be bad. FETs has only thermal noise to deal with at microwave frequency. It is common to find FETs with noise figure below 2db or even 1db.
     
    Last edited: Oct 1, 2012
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook