JFET as an amplifier
My answer to SE EE question JFET as an amplifier
Answer
Philosophy. In my answer I will try to expose the philosophy of this kind of biasing called "sorce (emitter, cathode) degeneration". Saying "philosophy", I mean not only to describe what is made but to explain why it is made this way...
Problem. We have to ensure some initial input voltage across the gate-source junction so that the transistor starts working even before the input voltage is applied. This "bias voltage" should be stable so that the drain current is stable... and the DC output (drain) voltage is stable as well.
Self-biasing. The clever trick of such self-biasing is that the transistor itself creates the necessary bias voltage by passing its current through the resistor Rs in the source; then it keeps up this voltage drop almost constant by the mechanism of the negative feedback.
Implementation. The bias voltage drop across Rs should be applied to the transistor input - the gate-source junction. How do we do it?
Fortunately, the N channel JFET requires a negative gate-source bias voltage. The source is already connected to the positive (upper) end of Rs; all that remains is to connect the gate to the negative (lower) end of Rs. How do we connect it?
The simplest way is to connect the gate directly (by a "piece of wire") to the lower end of Rs. And this is actually done in stages that do not require additional input voltage, such as JFET 2-terminal current sources (aka constant-current diodes).
When there is an input voltage (this case), we have to somehow add that voltage to the bias voltage. The simplest way is by connecting it in series. If the input source is "galvanic" (with a DC internal resistance), we can insert it directly between the gate and ground. In our case it is not possible because of the input coupling capacitor C1. The solution is to make the connection through a resistor (R2) with high resistance. Then the gate is galvanically connected through R2 to ground and the AC input voltage is applied through C1 to R2. As a result, a total voltage VIN + VRs is applied across the gate-source junction as needed.
Mixed biasing. But what do we do when the transistor is NMOS or NPN BJT... so it requires a positive bias voltage... and we do not want to give up the clever "source degeneration"?
Then we recall the simple classic biasing through an input voltage divider that would work here. But why not combine the two techniques?
Implementation. For this purpose, we add another gate bias voltage Vg in series to the source bias voltage Vs... and adjust Vg so that Vg > Vs. The total bias voltage between the gate and source will be positive as needed (Vbias = Vg - Vs > 0). So, in this case, the bias voltage is a difference between two partial bias voltages (they are grounded but the difference is "floating").
Operation. This technique is visualized in a geometrical way for the case of a BJT in the picture below. The voltages and voltage drops are represented by voltage bars (in red and warm colors) with proportional height and relevant position. The currents are represented by current loops (in green) with proportional thickness and relevant trajectory (only the base current is not shown... maybe it us forgotten).
As you can see, VBE = VB - VE is always positive; you can also see how the input voltage VIN (EIN) is "lifted" by VC1 (see also another relevant answer). To be more precise, we have to say that three voltages (VIN, VC1 and Ve) form the total base-emitter voltage.
You can get a good idea of the operation of the input bias circuit from the animation below:
I believe you will appreciate the power of this geometric approach of visualization...
My comments
- I made this picture in 2000. But the idea for this way of visualization came to me in the early 90's. Here are some ideas sketched in my archive on sheets of paper yellowed by time...
- Only to add that all these things are just tools that can't undo the need to think with our brains. As you can see from my answers, I always reveal what the idea behind the circuit is and how it is implemented in the case. Everything else just helps. We should not oppose the means to the goal and the form to the content...
- Here is the further development of the voltage bars in voltage diagrams. I invented it also in the early 90s. And here is an attractive real experiment based on Apple II and the AD periphery MICROLAB - living voltage diagram... which can give another idea to the designers of simulation programs... to make them more attractive to buyers..
- I am doing the same thing as I understand more and more circuit ideas. It would be very nice to discuss principles, concepts, smart circuit ideas... Last month I did something I should have done a long time ago - I started a blog Circuit Stories where I began collecting materials and links to everything I have created on the web, including what I am doing right now.
Comments
Post a Comment