Various kinds of diodes have a property called "storage charge." The effect of storage charge is that when a diode is carrying a current in its forward conduction mode, getting that current to stop flowing is a process that doesn't just happen right away. There are a variety of turn-off behaviors which merit some consideration.

The fundamental result of having storage charge is that the diode will not turn off immediately when a reverse voltage appears across the diode junction but will continue to flow through that junction in the reverse direction for some finite amount of time.

Just to make visualization easy, let's look at a half-wave rectifier circuit. For our first case, we have an ideal diode with zero storage charge and as you would imagine, there is no reverse current flow at all. The idealized waveshapes are seen in Figure 1.

Figure 1
Ideal diode half wave rectification

So far, so good, but look at what can happen when there is storage charge in the diode:

Figure 2 Slow recovery diode half wave rectification

In this case, the diode turn-off doesn't happen right away when the input sine wave crosses zero volts and, as a result, there is a brief but significant reverse conduction time. Also, the steps-to-zero of the output waveform are extremely rapid and are therefore fertile ground for generation of EMI at harmonics of the excitation frequency which can be very difficult to contain and suppress.

Also, if we digress for a moment to a full-wave rectifier, we can still get these reverse diode current pulses and those pulses could possibly lead to pulsed short circuits occurring across the source of excitation as in the following sketch.

Figure 3
Momentary short circuit

Line frequency, short circuit current pulses can really make for some severe EMI and ripple troubles. Consider, for example, a crude estimate where a pretty slow diode, the 1N4007, has a nominal recovery time specification of 30 µSec.

Figure 4
A slow diode's recovery time

If the excitation frequency is 60 Hz so that the period of one half cycle is 1/120 second or 8.3333 mSec, then the reverse current conduction angle would be 180° times 30 µSec divided by 8.3333 mSec which comes to 0.648°. If an excitation level of 120 volts RMS were applied, then the excitation voltage at 0.648° would be 120 × sqrt(2) × sin (0.648°) = 1.92 volts which is plenty high enough to drive an unwanted pulse of short circuit current through that red arrow path shown above in Figure 3.

You would be very well advised to use fast recovery diodes for this application, even though the line frequency is low. These pulsing events happening twice for each cycle of the AC input can be quite problematic. By using fast recovery diodes, things will look more like the following sketch with the unwanted pulsing effects being far less energetic.

Figure 5
Fast recovery diode half wave rectification

So far, storage charge has been seen as the enemy, but sometimes it can be a friend if we want to make use of it. If we take our excitation frequency from being 60 Hz or 400 Hz power line numbers and go up to radio frequencies as in HF/VHF/UHF and we use a device called a step recovery diode, we can have a rectification situation that looks rather like this next sketch:

Figure 6
Step recovery diode half wave rectification

Here, the step recovery diode's storage charge sustains a reverse current for a deliberately long part of the excitation waveform cycle, ideally to the 270° point as shown. Just as the slow recovery diode we looked at before yielded unwanted harmonics of the excitation frequency, the step recovery diode used here yields wanted harmonics of the excitation frequency which allows us to make a frequency multiplier along the lines of the following sketch.

Figure 7
Step recovery frequency multiplier

Picture 100 MHz going in and having 300 MHz coming out. That's pretty nifty!

If we next look at still another device, a PIN diode (p-region - intrinsic region - n-region), that device has storage charge that intentionally never does get depleted if the excitation freqeuncy is high enough. We can exploit the fact that the PIN diode's dynamic impedance can be made to vary as a function of the current level being carried. This trait makes the PIN diode useful as (among several other purposes) a variable RF/microwave signal attenuator element.

Figure 8
PIN diode signal attenuator element.

John Dunn is an electronics consultant, and a graduate of The Polytechnic Institute of Brooklyn (BSEE) and of New York University (MSEE).

Also see: