I've searched but I couldn't find answers to a few questions. First of all, how are MOSFETs rated in gain? I can understand how transistors are rated because a certain current on the base will pass another current between collector and emitter, so gain is just a ratio of the two. How do you describe the gain of MOSFETs? They pass a current depending on voltage, also with an almost infinate input impedence, isn't it kind of hard to derive gain?
My second question is about second breakdown on regular transistors. I'm not 100% sure, but I think it has to do with maximum current and voltage ratings reducing each other when one rating is near the peak. Correct me if I'm wrong.
My last question is about MOSFETs again. I've looked at specs and the base voltage vs drain current is extremely exponential (or logarithmic?) on most devices. How do you get a linear response from such a device? NFB? I know they're better at driving low impedence loads, but how do they compare with bipolar transistors when it comes to noise and distortion?
My second question is about second breakdown on regular transistors. I'm not 100% sure, but I think it has to do with maximum current and voltage ratings reducing each other when one rating is near the peak. Correct me if I'm wrong.
My last question is about MOSFETs again. I've looked at specs and the base voltage vs drain current is extremely exponential (or logarithmic?) on most devices. How do you get a linear response from such a device? NFB? I know they're better at driving low impedence loads, but how do they compare with bipolar transistors when it comes to noise and distortion?
Check for the parameter tranconductance, gm. Gain = gm*drain resistanceSolid Snake said:
Solid Snake said:
Check a datasheet and the graph of Current-Voltage. This is enough really without knowing what's behind. BJT can't handle very well high voltage and high current at the same time. MOSFET's are much better in this respect.
The answer is feedback and a simple emitter resistor is feedback! If we talk emitter follower (output stage) the unlinear characteristics aren't so bad after all. You will notice it more in gain stages.Solid Snake said:
In a BJT, as you say, emitter current is controlled by the base current. In a FET, the gate voltage controls the channel resistance between the drain and source. Thus, a voltage is used to control a resistance, rather than using a current to control a current. Current through the drain-source path is essentially modulated by the gate voltage, so we can say that the source current is proportional to the gate voltage. While the equations for FETs do not yeild results as precise as for BJTs, they are generally close enough. You may wish to consult a text book on FET operation.
Secondary breakdown is related to power dissipation. You can have either high current or high voltage and still have low power, but if both are high, then you have a lot of power in the device and if the device cannot shed this heat, then thermal runaway will set in (for BJTs) and the device will fail. FETs don't have the thermal runaway, but they will certainly fail if overstressed. The transistor curves will set maximum limits for both voltage and current, then show how you need to limit the in-between voltage and current to avoid problems. Always remember that even if a transistor is rated for 100 W, if you don't have adequate heatsinking, the transistor could burn up iwth just a few watts.
If you can operate over a small enough portion of the gain curve, a good approximation of a linear function will result. As said above, the wonderful magic of feedback will further linearize the response, at the expense of reduced overall gain.
Different people will say different things about the "sound" of BJTs and MOSFETs and even JFETs. I think there are highly rated amplifiers of each type, so it may come down to personal choice.
Secondary breakdown is related to power dissipation. You can have either high current or high voltage and still have low power, but if both are high, then you have a lot of power in the device and if the device cannot shed this heat, then thermal runaway will set in (for BJTs) and the device will fail. FETs don't have the thermal runaway, but they will certainly fail if overstressed. The transistor curves will set maximum limits for both voltage and current, then show how you need to limit the in-between voltage and current to avoid problems. Always remember that even if a transistor is rated for 100 W, if you don't have adequate heatsinking, the transistor could burn up iwth just a few watts.
If you can operate over a small enough portion of the gain curve, a good approximation of a linear function will result. As said above, the wonderful magic of feedback will further linearize the response, at the expense of reduced overall gain.
Different people will say different things about the "sound" of BJTs and MOSFETs and even JFETs. I think there are highly rated amplifiers of each type, so it may come down to personal choice.
You say that you can make MOSFETs linear by using a small portion of their gain curve. On most that I've seen, the first few hundred mA slope is much less steep than once you start getting into a few Amps. Would you just bias them so they counduct maybe the first few hundred mA all the time? Also, for using several devices in parellel, I know that you need a very small value resistor on the emitter of the final output transistors so that the transistors share current. Is this necessary for MOSFETs since they're voltage controlled?
1) You make any gain device linear by operating it over a small portion of its bias.Solid Snake said:
2) You are right, in a push pull Class AB output stage, you start getting good linearity around 100 mA with BJT's and about 500 mA with Mosfets. (your mileage may vary)
3) A Source resistor is not essential for Mosfets, but I recommend something to add additional stability and to allow measurement of current of each device.
4) It is not correct to think of a Mosfet's gain as the transconductance times the Drain resistance. Transconductance, in Siemens, is the inverse of the equivalent of the Source resistance, and is the gain figure itself. A transconducance of 5, typical for a power device, means five amps for 1 volt variation.
5) The gain of a power Mosfet is quite linear after you get above a couple volts Vds and an amp or so Ids. Above an amp, a Mosfet tends to be roughly equivalent in linearity to a BJT, and at higher currents it tends to get better, where a BJT will fall off.
6) I find that Mosfet are best for Class A due to the effect of (5), and BJT's are more useful in low bias Class AB, and personally I also like Mosfets because you can build more amplifier with fewer parts and simpler circuits.
- Status
- Not open for further replies.