Diodes usually come in glass or plastic cylindrical packages, marked with a stripe on one side to indicate polarity. In a perfectly ideal diode, current flows in one direction only, from the anode (positive side) to the cathode (negative side) which is marked with the stripe. The schematic symbol is a triangle pointing towards a bar, where the current flows in the same direction, towards the barred (striped) end. Surface mount versions of diodes tend to follow the same labeling convention, where the cathode end is marked with a broad stripe.
If we hook up a diode in a simple circuit with a variable voltage source and a current-limiting resistor, we can measure the current I through the diode when a given voltage V is applied across it. For an ideal diode, no current at all passes when the voltage is less than zero: the diode completely prevents reverse current flow. For small positive voltage ("forward bias," or sometimes "foward voltage"), a tiny amount of current may flow, and a very large amount of current will flow above a given threshold. The amount of current that flows is actually exponential with increasing voltage.
The threshold where an appreciable amount of current flows is typically around 0.7 V for simple semiconductor diodes, but may be as low as 0.15 V for Schottky diodes, or as high as 4 V for certain types of LEDs.
Of course, no diode is truly ideal. In real diodes, when the voltage is reversed, a very small amount of current (leakage) may flow. And, more significantly, each diode is rated for a certain maximum amount of reverse voltage. If you apply voltage more negative than that limit, the diode will undergo "reverse breakdown"
and begin to conduct a significant amount of current, but backwards from the normal direction of diode current flow. For a regular diode, we would say that the diode has failed if it begins to conduct current in that direction.
Aside: The actual physics of what happens at breakdown is quite interesting; two separate effects, the Zener effect and Avalanche breakdown both contribute to this behavior.
A typical application might be as above: A 10 V Zener diode (type 1N4740) is placed in series with a resistor and a fixed 12 V power supply. The resistor value is chosen such that several mA flow through it and through the Zener, keeping it in its breakdown region. In the circuit above, there is 10 V across the Zener diode, and 2 V across the resistor. With 2 V across a 400 ohm resistor, the current through that resistor (and the diode, in series) is 5 mA.
There are a couple of requirements to consider. First, the input voltage has to be higher than the Zener voltage. Second, the resistor value must be chosen such that there is always current flowing through the Zener.
Some caveats: This is not necessarily a good power supply for all purposes-- the resistor limits how much current can be drawn. It is also not necessarily a precision voltage reference; the voltage will depend on the amount of current drawn. (That is to say, for the voltage to be steady, the load driven by that reference voltage must be consistent.) The voltage also depends upon the temperature. Zeners in the range 5-6 V have the best temperature stability, and there are high-precision Zener diodes (like the LM399) that include their own temperature-stabilized oven to further keep the diode temperature as steady as possible.
Taking this idea a bit further, you can actually build a full multi-rail power supply using nothing more exotic than a set of Zener diodes to generate all the voltages that are needed, provided that the current requirements are modest on the different supply voltages. The circuit above is part of a working laboratory instrument.
A varying analog signal can be constrained to a fairly narrow range of voltages with a single Zener diode. If you have a voltage that swings between + 7 V and - 7 V, you could use a single 4 V Zener, connected to ground, to ensure that the signal does not exceed 4 V or go below -0.7 V (where the diode conducts forward to ground).
If you wanted to constrain the signal to never go negative-- e.g., for input to an analog-to-digital converter that accepts signals in the 0 - 5 V range, you could connect the anode of the Zener diode to a power rail at 1 V, instead of ground. Then, the output signal range would be constrained to the range of 0.3 V - 5 V.
Another neat trick is to use two Zener diodes, oppositely oriented, in series. This can provide a symmetric limit on the excursion of a signal from ground, for example. This is also a common configuration for using Zener diodes as transient supressors.
Here's something that doesn't work. We have a TL750L05, which is a type of 5 V output linear regulator, which can source up to 150 mA output, and its load will be variable. We need to drive it from a 36 V source. Unfortunately, the maximum input voltage of the TL750L05 is 26 V.
Our output load can be as high as 125 mA and as low as 10 mA. So, what value resistor will work for us?
Suppose that we assume 125 mA load. Then to take up (say) 20 V on the resistor, 20 V / .125 A = 160 Ohms. If we use 160 ohms, that will drop only 160 Ohm * 0.01 A = 1.6 V at 10 mA load, and 36 V - 1.6 V is still larger than 26 V. In order to be safe for the 10 mA load, we should pick a resistor that gives us at least an 11 V drop, for 25 V input to the regulator. So, 11 V / .01 A = 1100 Ohms would be safe for the 10 mA load. But if the load increases to 125 mA, the drop across 1100 Ohms would be V = 0.125 A * 1100 Ohms = 137 V, which means that the input to the regulator would be below 5 V, and it would cease to operate.
Clearly, there is no resistor value that you can pick that actually will work for both the low and high current cases.
Finally, let's try using one fat 20 V Zener diode (type 1N5357BRLG), to drop some of the load. Then, the output on the anode of the Zener is just 16 V, well within the safe input range of the regulator. The 1N5357BRLG is rated for 5 W maximum. At 20 V and .125 A, the power dissipated by the Zener would be 2.5 W, so it will get warm, but we are well in the safe operating conditions of the Zener, and now the circuit will work.
Evil Mad Scientist Laboratories
http://www.evilmadscientist.com/article.php/zeners