How they work, how to use them, and how not to use them.
The vast majority of makers have at least heard of Zener diodes, though not as many actually know what they are or how to use them. There’s a general idea that they’re some sort of voltage controller, but often that’s where the information stops. They are considered old-fashioned and redundant in some circles, and it’s understandable with the prevalence of highly-efficient switchmode regulator modules, some that are smaller than a TO220 package. Add to this the ease of access to more traditional linear regulators in small packages, such as the LM78LXX series in TO92 packages, and it’s easy to overlook the humble Zener.
We don’t think that’s fair. Though a contender for the oldest form of voltage regulator around, they are as relevant today as they were when they were invented. Some of the reluctance to use them comes from the fact that they do have different considerations to those other forms mentioned, so we’re going to treat them in detail, starting with a recap of regular diodes, then launching into the world of Zeners.
A long, long time ago, in the earliest issues of DIYODE, we looked at regular diodes, how they work, and what they’re used for. ‘Regular’ is an interesting term here, because even within the category of general diodes, there are big enough differences that ‘regular’ is a bit misleading. We really mean ‘other common types besides Zeners’.
Common diodes can generally be grouped into three categories: Power, Signal, and Regulating. Power diodes are the ones most makers are familiar with. These are used for making sure current only flows one way in a circuit. The two main reasons to do this are to rectify alternating current (AC) into direct current (DC). The other main reason is to protect a DC circuit from having power connected the wrong way around. Most DC circuits and many components don’t like this to occur, and it will result in damage in many cases.
Signal diodes are much the same except that they are made to handle a much smaller current. They are used to control the flow of signals within a low-power circuit, rather than to control bulk power. They are usually in smaller packages than their rectifier counterparts, often made of glass. Most signal diodes have a current limit of 100mA or less.
WHAT DO THEY DO?
Ultimately, most electronic components, and indeed electricity itself, can be likened to water in a pipe. Diodes play the part of one-way valves. Electricity can flow one way through them, but not the other. While there are a variety of ways to make a diode, the most common material is silicon. Regardless of the substrate material, doping chemicals are added to the material, which is then joined. Alternatively, one piece of substrate can be treated with different doping chemicals to achieve the same result, depending on manufacturing processes and a bunch of characteristics needed in the end result. Doping, by the way, is adding certain chemicals in trace amounts which form compounds with the atoms in the surface of the main material, and alter its properties.
Whatever the manufacturing technique, doping results in ‘P’ type materials and ‘N’ type materials. One of each is bonded together to form PN junction. Current will flow from the ‘P’ type material to the ‘N’ type material, but not the other way around. Well, not the other way around, most of the time. More on that later. The diagram here shows the PN junction made of its different doped layers, and the diode symbol aligned the same way. Notice that the triangle forms an arrow in the direction of current flow, which you can think of as a funnel. Current can go into that and out the other side. Current going the other way hits a brick wall and stops.
The terms for this are ‘forward biased’ when current is flowing through the diode from P to N, and ‘reverse biased’ when current is presented from N to P. We say ‘presented’ because until things go wrong, it won’t flow. It remains a potential force, not a kinetic one. In most diodes, the P material is called the ‘Anode’, represented by the letter ‘A’. The N material is called the Cathode, and is denoted by the letter ‘K’. As an aside, the names Anode and Cathode date from Michael Faraday’s work on electrolytic cells, and the story of why each was chosen is somewhat convoluted, but cathode is said to have been inspired by the Greek word ‘kathodos’, which means ‘the way down’ or ‘descent’. In these early experiments, current wasn’t fully understood and electrons had not been identified. This is why conventional current flow and electron flow are opposite each other. The Cathode was ‘the way down’ for the current into the cell, which we now know to be the opposite. While the Greek word is sometimes cited as the reason for the letter ‘K’ representing, it seems to have been assigned some time after the word ‘Cathode’ was in use and likely comes from the German spelling ‘Kathode’ because ‘C’ was associated with capacitors then being explored.
If you push against a brick wall with enough force, it will break apart or fall straight over. You’ll probably not be able to do this with your body strength alone (unless it’s a very weak wall), but you get the point. Diodes are the same: They have a limit to how much electricity can push against them before they fail. Like the brick wall, it’s not just about how much you push, it’s how hard. Current in a wire behaves like the rate of flow when talking about water in a pipe. Amps are the equivalent to litres per minute, or some such. Voltage, on the other hand, is the same as pressure in a pipe. So, diodes have a certain current they can pass through safely, but the voltage matters too. If the ‘pressure’ is exceeded, the diode is damaged.
However, if the pressure limit is exceeded in a one-way valve when the pressure is against it, it will break and allow water to flow the wrong way. So does a diode, and this quoted in datasheets as Peak Repetitive Reverse Voltage (V PRR), Working Peak Reverse Voltage (V WPR) or DC Blocking Voltage (V R). There is also a small leakage current that flows when a diode is reverse biased, even below its reverse limit. This is quoted as the Peak Reverse Current, usually written as I RM. While voltage is always quoted in the reverse bias, current is quoted for forward bias, aside from the leakage current mentioned. This is the current discussed earlier, and is the maximum a diode can handle before it is damaged. The datasheets give this as Maximum Average Forward Rectified Current I F(AV). These are the main factors to consider when choosing a diode, but there is one more.
All materials dissipate some energy, and therefore create a voltage drop. The diode is no different. While Silicon is a common material to use for diodes, it does have a bigger voltage drop than others. Most power diodes have a voltage drop of 0.6V to 1.2V depending on the diode type and the current going through. This is listed in datasheets as V F. This is usually fine for power diodes, but needs to be thought about. For example, if you have an IC with a minimum reliable operating voltage of 4.5V, and you power it from four AA batteries totalling 6V, the voltage drop can become a problem.
We measured with both a multimeter and a semiconductor analyser and found the voltage drop of a randomly-selected 1N4004 was 0.7V, but this was measured at 5mA. As the datasheets show, when the forward current increases, so does the voltage drop. This will quickly leave you little room for the batteries to discharge when using a 1N4004 as reverse-polarity protection. The supply voltage must be high enough to cope with the voltage drop of protection diodes.
SIGNAL DIODES AND SPECIALISED DIODES
Signal diodes behave in much the same way, and most silicon versions have similar voltage drops to their power counterparts. For this reason, some signal diodes can be found made from exotic materials. One of the most popular of these (and that’s a relative term in this case) is Germanium. This semiconductor material is a metalloid (not actually a metal) like silicon, and predates silicon in the semiconductor device field. While silicon has properties that make it desirable in many cases, Germanium diodes have a forward voltage drop as low as 0.15V. That makes it very useful for weak signals, such as those from AM radio waves. It is a germanium diode that is found at the heart of ‘crystal’ radio receivers.
Another term makers may come across is the Schottky diode. This is a power diode but is designed with a PN junction made with one piece of semiconductor, and one piece of metal. The upshot of this is a diode that has appreciably lower forward voltage drop that silicon diodes, and switches between conducting and non-conducting states far more quickly. Specialised materials like Germanium are not listed on circuit diagrams, instead relying on parts list numbers to ensure the right device. Schottky diodes, however, do have their own symbol.
Now for what you came here for. Zener diodes occupy a unique space in the world of electronics. There are two ways to make them, and both give different characteristics. Both use the same symbol in circuit diagrams, and it looks uncomfortably close to a Schottky. However, the arms have only one angle, not two as in the Schottky, and the direction of these is reversed. There are two factors at play in Zener diodes. One is called the ‘Zener Effect’ and it, like the device, is named after Clarence Zener. Without getting deep into electronics, valances, holes, pairs, and regions, the Zener effect occurs when an electrical field enables a ‘tunnel’ of electrons to cross the depleted region of the PN junction.
The other factor at play is avalanche breakdown, which happens more gradually when compared to the Zener effect but is still sharp enough. Avalanche breakdown, again overly simplified, involves accelerating free electrons and allowing a current flow that way. Both occur at once in any Zener diode, but the Zener effect is mostly responsible at lower Zener voltages, while avalanche breakdown is most relevant above that point. The threshold varies between sources but most agree on a transition between the two effects of 5 volts. This means that diodes the maker uses could easily fall into either category.
Zener diodes are designed to be reverse biased above their breakdown voltage and not be destroyed. They can pass current quite happily in the reverse direction, if used within their limits. This selected voltage is known as the Zener Voltage, because Zeners still have an absolute reverse breakdown voltage after which they will be destroyed. This means that Zener diodes can act as a basic voltage regulator. Any potential difference (voltage) above the Zener voltage is passed to ground, so a voltmeter connected between the anode and cathode will show the zener voltage rather than the supply voltage. The water analogy is water flowing into a bucket with a tube in the side.
Water fills the bucket up to the tube, but after that it does come out. No matter how much water flows in, the water level in the bucket stays the same. That is, unless the water flows in faster than it can flow out, but all analogies fall down at some point. This would be the equivalent of exceeding the current limit. In the accompanying diagram, the power supply is 12V. The Zener voltage for the Zener diode is 9V, while the breakdown voltage for the power diode is 400V. What will happen when power is applied?
A potential difference exists between the anodes and cathodes of both diodes. Because the power diode has a much higher breakdown voltage than the Zener, no current flows through it, except for the tiny, ignorable (in most cases) leakage current. In the Zener, however, things get very hot, and smoke comes out. If you’re unlucky, glass comes out too, probably quite fast. So, cover your eyes or wear safety glasses. Sadly, using a Zener is not as simple as connecting it in reverse bias to a power supply and enjoying lots and lots of lower-voltage amps.
Like most electronic devices, Zeners are made to a power rating. Like the regular diodes in forward bias, the Zener can only cope with so much current in the reverse mode. On top of this, there is quite a bit of heat generated in the reverse mode, and that is why Zeners are still made with glass cases. The plastic version tends to melt too easily, although ceramic and Bakelite versions are around. The most common ratings for Zeners on the retail market are 1 Watt and 5 Watts. Plenty of others are around from trade and commercial suppliers. This means a 1 Watt Zener can dissipate 1W of electrical energy. In other words, if you have a 9V Zener and feed it from a 10V supply, it could pass 1A. That may not be exactly the case, however, as the devices also have limits in the datasheets just like any other. The datasheet from On Semiconductor, for example, lists the Non-Repetitive Peak Reverse Current for the 1N4739 9.1V Zener as 500mA.
The catch is that the diodes have no way in their internal structure to current-limit themselves. Neither does a power diode, for that matter. If you connected a 1N4004 with its 1A limit in series with a 120W light bulb on 12V, it would fail, probably explosively. The same goes for the Zener but while a regular diode uses the load of the rest of the circuit for current limiting, the Zener is, by its nature, connected across power rails and can therefore pass as much current as the power supply can provide. This is why, in the diagram above, the smoke comes out.
By the way, for the uninitiated, there is an old saying in the electronics industry that electronic components are made from smoke. Rather than a reference to ‘smoke and mirrors’ meaning magic, it refers to the fact that often, components emit smoke when they fail because it usually involves significant heat and the partial combustion of materials in or on the component. Once you ‘let the smoke out’ by overloading or otherwise destroying the component, it no longer works. This led to the in-joke that components are made from smoke because once it’s out, they don’t do what they should.
THE ZENER EFFECT UP CLOSE
While for practical purposes, the Zener diode breaks down at its rated voltage and allows current through at any voltage over that, the reality is a little different. There is a defined graph that represents the Zener diode’s behaviour, and the same graph doesn’t vary much across devices. By leaving units off and only labelling axes with the name of the value, we can give a pretty universal representation.
Notice that there are slight curves on both sides of the graph. The point in the Zener breakdown region where the graph starts to curve is where the current (the x or vertical axis) has fallen to a minimum value. The y or horizontal axis shows voltage, so that part of the graph may change between devices. However, note the fact that there is a voltage above which the diode isn’t conducting much of anything. It is, however, a tiny voltage and you won’t create this situation very often.
ZENER DIODES AND TEMPERATURE
Temperature affects Zeners in other ways. Not only is there a maximum current limit because of the heat generated, but the heat may affect the power rating. For the On Semi datasheet already quoted, the 1W Zeners have a temperature limit of 50°C. Above this, the power rating must be reduced (derated) by 6.67mW/°C. So, if operating at 70°C, the difference is 20 x 6.67 = 133.4mW. That isn’t a lot but it can be if temperature climbs, it is also the junction temperature, not the case temperature.
The Zener Effect and Avalanche Breakdown are both affected by temperature. The Zener Effect has a negative temperature coefficient, which means that as the temperature climbs, the Zener voltage falls slightly compared to the nominal Zener voltage. By contrast, avalanche breakdown has a positive temperature coefficient, so as temperature increases, so does the breakdown voltage. The graph we have here is highly simplified and only representative of the concept, as not only is temperature coefficient quite complex, and curved, but the data is often not in the datasheets.
Temperature Coefficients are likely to be something most makers will not need to worry about, especially given the fact that no graphs or numbers are given in the datasheet for the On Semiconductor devices, nor any other we could find. However, we did find several sources which state that the two curves overlap at around 5.3V, and for all intents and purposes, cancel each other out. Coincidentally, the same voltage is where the Zener and Avalanche breakdown effects cross over, making the 5.6V Zener (the nearest preferred value) a good choice when a stable reference voltage is needed and you have the flexibility to choose that voltage rather than having it dictated by the circuit.
Like many other components, it’s not practical or economical to make Zener diodes in every value possible, nor to an exact value. Generally, they are available as a selection of common values, and tolerances. On the retail market, 5% tolerance is the most common. The On Semiconductor datasheet for the 1N47XX series states this at the top, then gives minimum, nominal, and maximum values for each voltage in the range. Common retail voltages are from 3.3V to around 40V, with some outliers. Below 10V, there are generally one or two decimal values available between each volt, and the range changes in whole volts thereafter.
In order to use the Zener as a regulator, there must be a limit to the current through it. In the simplest case, and probably the majority of examples, this is done with a resistor. That’s it, just one plain old boring resistor, although it must be a big enough one to handle the current. It’s no use trying to drag 500mA at 12V through a 1/4W carbon film resistor. The smoke will come out of that, too.
CHOOSING A VALUE FOR R
Choosing a value for the series resistor RS isn’t as straightforward as calculating the maximum current through the Zener. There is voltage drop across the series resistor too, which has to be factored in. The diode needs to dissipate the minimum current possible, which sometimes means choosing a bigger series resistor. This might seem counterintuitive seeing as we’re used to getting resistors to be as small as possible to avoid losses through dissipation, but in this case, it helps to dissipate some power in the resistor and rely on the Zener for regulation.
With no load connected, all the available current flows through the Zener. This situation must be factored in as the worst-case scenario. Added to this, there is a minimum current value through the diode in order for it to maintain regulation. This varies depending on Zener voltage but the datasheet for the series quotes between 0.25 and 1mA. Some of the test conditions were conducted at 4.5mA for this data, so 5mA is a safe bet for a minimum load.
To calculate the maximum current that passes through the Zener, we divide its power rating in Watts by the Zener voltage it will dissipate. We use this equation:
With a 1 Watt 9V Zener, that’s 0.1111A, or 111mA. That’s not a lot!
The series resistor RS is calculated by:
Where V S is the supply voltage, V Z is the Zener voltage, and I Z is the current through the Zener. With a nominal supply voltage of 12V, that gives us 27Ω.
Now, Rearranging equation 1 again, we can figure out the power rating this resistor needs by multiplying the current through it by the voltage that will be across it, which is 12V supply minus the 9V Zener, leaving 3V. We end up with 0.333W, which means a 0.5W resistor should do.
However, Zener voltage regulators wired this way depend heavily on the load current being constant, as any load change affects the voltage drop across the resistor. If the voltage drop changes, and the resistance is of course a fixed value, Ohm’s law tells us that the current through the resistor has to change. The healthiest thing to do is use the above to calculate the maximum current a Zener diode can handle, and then figure out how much current it draws. Choose the resistor to give that current plus a little margin, and no more. If the load current is going to fluctuate, Zeners are so far out the window. The second equation works fine if, instead of the maximum current of the Zener diode, you use the load current.
Of course, Zeners are most valuable to the maker when things don’t need to be big. Some people would recall the early days of LED lights in cars. Many of the aftermarket plug-in versions that were meant to be replacements for incandescent globes were made quite cheaply. They were designed overseas for a nominal value of 12V if they were very cheap, and maybe 14.4V if they were good. However, the power from a car’s alternator is rarely stable, and spikes of up to 18V can occur. Some of the more expensive lights, when broken open, had a Zener inside to cope with these spikes. However, because of the current restraints, these were usually for interior or dashboard lights with small current requirements. For break and indicator lights with up to thirty LEDs, this wasn’t so viable.
Having said that, this basic regulator circuit is still perfectly viable for low-current applications. Any situation that demands less than a hundred milliamps or so may benefit from a Zener regulator, especially if there is a design reason to power a subcircuit at a lower voltage than the main circuit. This may be especially true of situations where 3.3V sensors are used in 5V circuits, or something similar. In this case, the current requirements are small, and so the resistor does not need to be of such a high power rating.
REGULATOR AS A REFERENCE
While a limited load current can make the Zener a less than ideal load current, it makes an ideal reference voltage device. The left-hand arrangement in the diagram below shows an op-amp connected with a voltage divider formed from two resistors as its reference voltage. On the right, is an op-amp connected with its reference taken from a Zener diode, with a current limiting resistor included.
For the case of the voltage divider, which is quite common, consider what happens within. The supply voltage is shared between the two resistors, and the voltage at their junction is a percentage of the supply voltage. This is dedicated by the ratio of the two resistors. If they are the same, the reference voltage is half the supply voltage. If the first resistor is twice the size of the second, the reference will be one third the supply voltage, and if the first is a third the size of the second, the reference voltage will be three quarters of the supply voltage. The concept goes on.
The challenge here is that if the supply voltage fluctuates, as can happen when a high-current load kicks in or the supply is not stable, then the reference voltage changes. Some circuits don’t mind this but in other situations, that can be a problem. The op-amp in most cases would have a wide enough operating voltage that it will carry on quite happily but if the supply voltage dipped from, say 12V to 10V, then the reference voltage would change by 1V. That’s for the even resistor example, of course. This can cause false triggering in a comparator situation or drastically alter the amplification in other situations.
With the Zener as a reference, nothing changes regardless of supply. The resistor involved is there to limit current through the diode and has no bearing on the reference voltage. If a Zener diode of 5.6V or 6.2V was used (the two standard values closest to 6V), the voltage could fall even to 7 volts without affecting the reference. This would likely affect the IC or the circuit supplying the input signal before it affected the reference. If you need absolute stability in a reference signal, then a Zener might be what you need. Additionally, there are many uses for stable reference voltages besides op-amps.
If you do need an odd voltage, it is possible to ‘stack’ Zeners by connecting them in series. When this happens, the sum of the Zener voltages adds up. This can be a very stable voltage divider or a custom reference or regulator. If you wanted a stable, accurate 13.8V reference, common in 12V battery and automotive situations for testing and design (although operating voltages are designed to be 14.4V), then your choice in standard values is between a 13V and a 15V Zener.
However, there is no 0.8V Zener diode available, or at least we could not find one. Common values start from 3.3V. To make 13.8V, we’ll have to be fairly creative about our selections. Instead of starting at 13V or even 10V, we ended up with a 9.1V and 4.7V Zener diode in series. Because the current stays the same in a series circuit, we can choose one resistor and it won’t need to change, but the dissipation is shared across the devices, so we now have 2W of Zener diode and not 1W. The series resistor is still needed, and can be calculated in the same way as for the basic regulator.
It is also plausible to use Zener diodes in series for voltage divider purposes. If, like in the reference example, you need reference voltages that do not change, then a series of Zener diodes may be of more use to you than a resistor-based voltage divider. This will be immune from supply voltage fluctuations up to a reasonable point. Naturally, of the supply drops below the largest Zener value, there will be problems. However, when we talk about supply voltage fluctuations, we usually mean fairly minor amounts. For anything bigger, there is a problem somewhere that needs to be solved.
As an aside, you can stack regular diodes, too. The diagram depicts a very basic voltmeter built with a series of silicon diodes. Each one has a voltage drop of around 0.7V, so each LED lights when the supply voltage is 0.7V higher than the one before it. The circuit has a limited use as-is but might still serve as a battery meter for smaller voltages. The load it places on the battery would not be high enough to give a true indication, so it would only be useful if it were connected in parallel to the load the battery is normally powering. The circuit can be modified with a few extra components to make a basic VU (audio level) meter.
MAKING BIGGER REGULATORS
You might have guessed that Zener regulators are not going to power your next electric car or high-torque motor. The current limits involved are small and the losses somewhat high. However, adding ancillary components can boost the capability of the Zener diode. The example shown here is called a ‘series voltage regulator’, because the component that provides the regulation of the power rail is in series with the load. Most regulator circuits we see online are series regulators in some form. However, you might already be questioning that the regulator is a transistor? Yes, and this is how we get usable current from a Zener diode.
While this circuit is an electronics staple, there are many forms and variations. They share a lot in common, however. The voltage from the supply is connected to the collector of the transistor Q1, which can be any of a large selection of NPN transistors. The supply is also connected to a resistor, which limits the current to the base of the transistor. However, because the Zener is connected between the base and ground, the voltage at the base is held at the Zener voltage, current-limited by the resistor. This makes it immune to variations in the supply voltage, and is easier to fix at a value than a resistor voltage divider. If the resistor were connected to the base without the Zener present, the base would be at almost supply voltage, as the current from base to emitter is tiny.
In this way, the transistor passes only a voltage that is at its base, minus its base-emitter voltage drop, which is around 0.6V for most common transistors but may be more for the higher-power ones. Consult the datasheets for that. While we tend to think of transistors as current amplifiers, this arrangement is called an Emitter Follower, and the voltage at the emitter reflects the voltage at the base, minus the voltage drop discussed.
Other components of note are C1, which is a standard filtering capacitor used in most power supply circuits, and C2, which helps keep fluctuations to a minimum at the Zener. There are two more filter capacitors on the output. The current limit is the limit of the chosen transistor, derated a bit for safety. We chose a BD139 for this circuit, though we’re not building it, just showing values. To regulate 5V from 12V, we use a 5.6V Zener to cope with the base-emitter voltage drop. This circuit, with heatsink fitted, should comfortably supply 1 amp of current.
We calculate R S the same way as we have so far, with the supply voltage minus the Zener voltage, but this time we can calculate a nominal current because there is no load to supply besides enough current to keep the Zener in regulation.
To minimise losses, we can make this current around 10mA. That results in 640Ω, so the nearest value of 620Ω will do just fine. Just remember the order of operations rule. Brackets and powers first, then division and multiplication, then addition and subtraction.
If you enter 12 - 5.6 ÷ 0.01, the calculator will do 5.6 ÷ 0.01 first, then subtract that from 12. To make it work on a calculator, either subtract 5.6 from 12 first, then divide the answer (either with the NAS key or by re-entering it) by 0.01, or just add brackets, like this: (12 - 5.6) ÷ 0.01
Interestingly, the regulation from the venerable LM78XX series regulators also involves Zener diodes. Here is the internal schematic for the LM7805. While there are lots of transistors and cross-connected components to give very stable regulation, look closely at the left of the diagram and see if you recognise something.
While there isn’t a very big call anymore for Zener diodes as regulators. When used in the right role, they’re as useful now as when there were no other options.
One of the reasons we didn’t present the voltage regulator as a build was that it offers no advantage over just inserting an LM7805, or better still, the tiny switchmode regulators available now with the same pinout and footprint of the LM78XX series. It’s useful to know how it works, and one day you may well have a reason to build one. Perhaps for a custom voltage, or to drive a really big transistor and provide lots of current.
As references, the Zener will find more use for the maker. They’re stable and quite precise, offering the advantages we covered over resistor voltage dividers linked to the supply. Perhaps, after reading this, you may never use a Zener, or they may appear in your projects when they weren’t going to before. It’s hard to say, but hopefully, you’ve taken something away from this and have an appreciation for these underappreciated components.