Fundamentals

Fundamentals - Circuit Protection

Dan Koch

Issue 66, January 2023

An overview of the different things to protect your circuits, and some methods of doing so.

From simple reverse-polarity protection, so you don't roast your breadboarded project by connecting the battery the wrong way, to fully current-limited lab power supplies which can keep dumping the set amount of current into a short circuit without dropping the voltage, circuit protection is a multifaceted subject, and some facets have a bewildering array of options. In this month's Fundamentals, we will take a look at protecting against reverse polarity, short-circuiting, overcurrenting, voltage surges, and even under-voltaging.

This is more about filling gaps in both knowledge of newer makers, and the memory of more experienced makers. By memory, we mean that it is easy to get complacent and forget about some of these different categories of protection and when they might apply. We even learned some new things while researching for this article (we never, ever write Fundamentals or Classroom from memory, we always find sources). So, without further preamble, it is time to dig in.

REVERSE POLARITY PROTECTION - DIODES

Reverse polarity protection is about so much more than just accidentally reversing battery connections. It is not always even about connections at all. Reverse polarity can occur through voltage spikes from collapsing inductive loads, among other sources.

We consciously leave out reverse polarity protection on most of Kids' Basics circuits, because past feedback has shown it is generally easier for newer makers to go 'red to red, black to black (or blue)' on the breadboard rails directly rather than using a diode from a row up to the rail and connecting the battery into the row. However, it is otherwise not a good practice. The other reason we avoid it becomes very relevant to this discussion: Voltage drop.

The most common way of achieving reverse-polarity protection is with a diode. The diode is one of the first semiconductor devices many makers learn about, but we acknowledge that some of our makers come from completely non-electronic backgrounds and have discovered coding or some other pathway. So, to recap, a diode is a junction of P-type semiconductor material and N-type semiconductor material. Current will flow one way across this junction, but not the other. That's why the diode's symbol is an arrow with a bar. When current flows from anode to cathode, the diode is said to be 'forward biased' while if the cathode faces the positive voltage, the diode is 'reverse biased' and no current will flow.

The anode is the positive end where current can flow from, and the cathode is the negative end where current flows to but not into. The symbol 'K' is used for Cathode. It is said often that this is from the German 'Kathode' but our research suggests this is untrue and no more than hearsay or an assumption. K was chosen because C was taken. Very conveniently, most diode packages have a stripe at one end that corresponds to the bar on the diode symbol.

Diodes come in many packages and are based on many materials. The most common are silicon diodes but other mineral substrates or doping chemical combinations exist. Germanium diodes, for example, have properties loved by musicians building amps and also have the ability to detect certain radio signals. Most diodes come in cylindrical packages but the details vary, and some come in TO220 transistor packages as well. The size of the package is generally dictated by the current handling capability of the device, and its voltage rating.

There are several types of diode that belong in this general category of general purpose diodes. Small signal diodes is a name given to a group of light-duty diodes meant for in-circuit signal-level work rather than power rectifying. They generally have glass packages and a much lower current rating than others, but have other desirable properties.

Standard Silicon Rectifiers (SSR) are the general power diodes, and these usually come in a black plastic case which is bigger than the signal diode and with thicker leads. The cases get bigger as the current capacity gets bigger. In the image, we have a 1N4148 small signal diode at left, then a 1N4004, a 1N4007, 1N5404, 1N5819, 6A4, and FR302.

Besides the current limit of each device, there is also a maximum voltage limit. This is the voltage the device can withstand before being destroyed or arcing over. Many SSR diodes on the retail market are in the 400V to 1000V range but others are available. The voltage rating is not important to many of us because we will never even approach it, let alone exceed it. The voltage and current ratings of the diodes in the photo are included here, as well as the voltage drop that we measured with our Peak Atlas Semiconductor Analyser.

Diode values

Diode Model

Voltage Rating

Current Rating

Measured
Voltage Drop

Measurement Current

1N4148/1N914

100V

150mA

0.698V

5mA

1N4004

400V

1A

0.670V

5mA

1N4007

1000V

1A

0.665V

5mA

1N5404

400V

3A

0.617V

5mA

1N5819

40V

1A

0.239V

5mA

6A4

400V

6A

0.590V

5mA

FR302

100V

3A

0.511V

5mA

One of the biggest factors in using diodes is voltage drop, or forward voltage. There is a loss over the PN junction in diodes, and it can be significant. The often-quoted figure is 0.7V for basic SSRs. Our table shows lower values because of the 5mA test current of the Atlas test instrument.

At higher currents, the voltage drop increases. This voltage drop can be a significant loss. In a battery powered Kids' Basics circuit, for example, a 0.7V drop means the 5V from our USB-powered options would only give 4.3V to the circuit. In applications where a 3.7V battery is powering 3.3V devices, the drop is even more serious. SSR diodes can only be used in situations where that voltage drop will not have a problematic impact.

The other issue with voltage drop is power dissipation. At 3A, 0.7V of drop is 2.1 Watts of power wasted as heat! It does not take long for such losses to become a problem.

There are other types of diode on that list, some with lower forward voltages (voltage drop). Schottky diodes are a special diode in which a semiconductor bonds with a metal as the PN junction. The voltage drop across a Schottky diode is generally 0.2V to 0.45V. In addition, semiconductor PN junctions have some parasitic capacitance which slows down the speed to switch from non-conducting to conducting when reverse-biasing has been removed. An SSR often takes a few microseconds for this change, while a Schottky diode can do it in as little as a few tens of nanoseconds for the power version and even picoseconds for the small signal version.

This matters in high-frequency applications but is rarely an issue for reverse-polarity protection. While switching speed may not be important, the lower voltage drop is attractive. However, Schottky diodes tend to have a lower reverse-bias voltage rating than SSR diodes, and are not always viable depending on your situation. It is always worth checking the datasheets for each component you want to use, even though they are daunting for the inexperienced.

It is also worth mentioning the 'fast' and 'ultrafast' diodes you will see in product listings. These are different constructions of SSR but designed to be between ten and fifty times faster at recovering after reverse-bias is removed. Again, this is not an issue for regular DC reverse-polarity protection, but rather important in high-frequency rectification or signal processing. There is no need to spend the extra money on them for a DC situation.

THE BRIDGE RECTIFIER

The bridge rectifier is an arrangement of four diodes, placed so that current can always flow to the outputs no matter which way around it is connected to the inputs. It is generally used to rectify alternating current (AC) into direct current (DC). However, it has uses in DC applications in that the circuit will work no matter which way around the power is connected.

This can be very useful when designing a circuit where it would be easy for power polarity to be mixed up. Such examples are when non-technical people are connecting batteries or when a certain power plug does not have a polarity control like offset pins. The significant disadvantages here are literally double the SSR disadvantages - double the voltage drop and double the power dissipation losses.

REVERSE POLARITY PROTECTION - OTHER OPTIONS

The other most common method of implementing reverse-polarity protection is with a MOSFET. For full details on MOSFET types and usage, see Classroom 32. A summary of the relevant bits follows here.

Metal Oxide Semiconductor Field Effect Transistors are a type of transistor which is voltage- rather than current-controlled. The three terminals are Gate, Drain, and Source. The gate controls things, the source is connected to the power rail (supply or ground as appropriate), and the drain is connected to the load. Just like transistors with NPN and PNP, there are N-channel and P-channel MOSFETS, often abbreviated to NMOS and PMOS respectively. Also like transistors, whether the source or drain are connected to the power supply rails or the load depends on the type of MOSFET.

There are other properties of MOSFETs, however, that make them useful for reverse-polarity protection. They too are PN junction devices but they have very significant current capabilities with very small voltage drops. Some MOSFETs have forward voltage drops as low as 0.02V! They actually have an 'on resistance' rather than an absolute voltage drop and so the actual voltage drop will vary with forward voltage and current. However, it will generally be lower than any PN junction diode.

The gate of a MOSFET is actually a capacitive element, and is not directly connected as a current path. Whether the capacitor is grounded or fed a voltage determines whether or not the MOSFET is on. For an N-channel MOSFET, the gate must be supplied with a voltage to charge the gate capacitance and keep it charged for the MOSFET to be on.

Leaving the gate floating can lead to erratic or unpredictable behaviour, so the gate is generally grounded through a resistor so that the capacitance discharges and stays that way when the gate is not being deliberately fed a voltage.

N-channel MOSFETs are generally used as low-side switches. It would be more accurate (but less clear to less electronically-experienced makers) to say that the gate to source voltage must be positive. That generally just means applying a voltage, though, because in most uses, the source of an N-channel MOSFET is connected to ground.

For a P-channel MOSFET, the gate capacitance must be discharged, and held that way for the MOSFET to be on. Therefore, grounding the gate turns on the MOSFET and the gate is tied to the positive supply rail by a resistor so that the capacitance stays charged when the gate is not being deliberately grounded.

P-channel MOSFETs are generally used as high-side switches. It is more accurate to say that the voltage

between the source and the gate must be negative to turn the MOSFET on. Normally this is done by having the supply voltage connected to the source, and the gate connected to ground. That way, the gate is at a lower potential voltage than the source.

Happily, it is not hard to configure a MOSFET to be used as a reverse-polarity protection device. It is best to use a P-channel device on the high side, because if there are any other current paths to ground, an N-channel device may not provide protection: Current can still flow through the load via these other paths.

It is far less common to have multiple routes from the power rail to the load than from the load to ground. The MOSFET is reversed compared to its normal usage. The diode inside the MOSFET (which most modern devices feature) allows current through in the forward direction as well as the MOSFET itself. This is important because, before the MOSFET conducts, there still needs to be a negative voltage between source and drain. With the MOSFET connected this way around, that cannot happen without the diode.

If polarity is reversed, there is no longer a negative voltage difference between gate and source and the diode is now reverse-biased and out of play, so the MOSFET switches off. The Zener diode and resistor are included to limit the gate voltage to safe levels. The gate voltage of a MOSFET often has a much lower limit than the Drain-Source voltage. Choose the values based on the datasheet for the MOSFET you are using. The resistor value should be between around 100Ω and 500Ω to make sure the zener current is high enough to fully bias the Zener.

The P-channel MOSFET circuit has some advantages over a diode, namely the very low RDS, or Drain-Source Resistance, and therefore the voltage drop that results and the power dissipation that is associated with it. The disadvantages are the increased cost and component count, plus connections, and the current draw from the Zener diode and resistor. If you are using the MOSFET with a supply voltage that is within the gate voltage limit, then you can skip the resistor and Zener, but be careful that the gate voltage is high enough.

While the threshold voltage is that at which a MOSFET starts to turn on, it is not fully on until quite a bit more than the threshold voltage. There is a graph in the datasheet that shows this, but for many MOSFETS which state a threshold voltage of around 4V, the MOSFET is not fully turned on until around 10V.

So, this is fine for a 12V circuit, but not a 5V circuit. Below this value, where the MOSFET is not fully on, RDS increases markedly, and power losses increase too. Heat will be far greater and the current capacity reduced. Note that the voltages can be negative or positive depending on whether it's a P-channel or N-channel MOSFET.

A MOSFET-based reverse-polarity protection circuit can also be made from an N-channel MOSFET. The circuit is virtually the same, but upside down. There are some advantages and disadvantages to this over a P-channel. The wider variety of N-channel devices that are retail-accessible as opposed to commercial is one of the more significant reasons.

The disadvantage is the aforementioned fact that if there is more than one ground path, as there is in situations like cars with chassis grounding, then the N-channel option may not provide protection. It will also raise the ground potential ever so slightly, which will be ok for most circuits but problematic for very sensitive ones.

Both of these circuits feature some current draw through the Zener diode and are not suitable for applications which need a very small current draw on battery, for example. There is also a possibility that some backflow can occur momentarily when polarity is reversed, due to the short but real time taken for the MOSFET gate to change state.

This is not a problem for most basic circuits but could be for complex or fragile circuits. Additionally, it is not a problem for protecting against incorrect power connection, as there is no state change as such in that case. The problem arises if the applied voltage is suddenly reversed, like it is with sudden voltage spikes.

As a final note on MOSFET protection, it is also possible to produce an N-channel protection circuit on the high side. This gets around the standby current drawn by the Zener but does require a charge pump or step-up converter to make sure the gate voltage is higher than the source voltage. This may seem counterintuitive, and it is for most workbench makers. Industrially, however, it is possible to produce some charge pump designs which consume microamps, much less than the Zener.

MOSFET CHOICES

There are several things to consider when choosing a MOSFET for any of these roles. The most important is RDS, which is the resistance between drain and source when the MOSFET is fully on. It can be used to calculate the actual voltage drop depending on how much current is being drawn. The VDS is the drain to source voltage, and is effectively the voltage rating of the device. The gate voltage (VGmax) and threshold voltage (VThr) are also important.

The resistor selection should be balanced so that the gate charge drains as fast as possible, but without wasting current through the Zener or even overheating it. We quoted 100Ω to 500Ω earlier, but if there is no expectation of a reversal in polarity during operation (in other words, when using to protect against connecting power in reverse rather than reversal during operation) then a higher value up to around 20kΩ is fine.

MOSFETs are not the only transistors which can be used to protect against reverse polarity, even though they are the most common. It can also be done with a PNP bipolar transistor. This is useful for small circuits where a TO220 P-channel MOSFET will not fit or is overkill.

There are TO92 N-channel MOSFETs on the retail market but we had to go to a trade supplier and order in commercial quantities to get a TO92 P-channel. It still has a much lower forward voltage drop than a diode, as long as the transistor is saturated. The heat generated and power dissipated are also much reduced compared to a diode.

The current-driven bipolar transistor also has the same voltage limits for its base terminal as for its other terminals, eliminating the need for a Zener diode. The main disadvantage is that the base current required is constant, and for circuits that are not on constantly or have small current draws, that is an issue. If the circuit is required to have a low stand-by current, then this may not be a great option.

OVERCURRENT PROTECTION

Overcurrent protection refers to devices or circuits which shut down the circuit in the case of the circuit current exceeding a certain value. This includes fuses, circuit breakers, and polyswitches. However, we're not including clamping circuits here. We've called those 'short circuit protection' and they and others are described next. The items covered in this section are usually meant to be 'just in case', and not used regularly. Clamping circuits, on the other hand, are for situations where overcurrent is expected regularly, like bench power supplies and some tools where stall could occur, among other applications.

Fuses are among the most common overcurrent protection. While many of us are familiar with them, it is helpful to describe what makes them work for the discussion ahead. A fuse in its simplest form is just a piece of wire. That wire can carry a certain amount of current before becoming so hot that it melts. In real-world terms a fuse is some kind of housing to hold a very precisely-calibrated piece of wire, often made of a different material than regular signal or power wiring, and facilitate electrical connection to it.

If the rated current is exceeded, the fuse fails. However, this is rarely precise. A 10A fuse, for example, is unlikely to blow at 10.1A. Time is also a factor. Metals become softer and weaker as they are heated, so a fuse that is running overcurrent enough to heat up is more likely to actually blow the longer it is running overcurrent. Most fuse types come in fast blow and slow blow. Fast blow will fail quickly when the current is exceeded, while slow blow types are designed to withstand some surge current for a certain amount of time before blowing.

There are sub-types of slow blow fuse, such as slow, time-delay, and others. They all represent a different tolerance for overcurrent before they blow. On the retail market, however, they are usually considered the same. Thus, you might buy a 5A glass fuse which is marked S5A for slow blow, then end up at your next purchase with a T5A for time delay.

Fuses come in a variety of physical constructions. Some are glass, some ceramic, some plastic. Some use text while others use colours to denote the ratings. All have some things in common. Once a fuse blows, it needs to be replaced. They are single-use devices. Older homes had fuses for mains electrical connections that were rewirable.

These have largely given way to circuit breakers, partly because circuit breakers are more convenient, partly because they are more consistent, and partly because it was easy for a householder to accidentally or deliberately use the wrong wire to rewire a fuse, removing the protection they are supposed to provide.

When designing your own circuits, the main consideration with a fuse is the voltage drop across it. By nature, this has to exist in order for the heat to be generated and the wire to fail. Without voltage drop there is no power dissipated and therefore no heat.

The voltage drop can matter in some circumstances. For a 5A M205 glass fuse, we measured the voltage drop at 1.4mV with a 100mA load, 5.0mV with a 420mA load, 12.8mV with a load of 840mA, 32.9mV with a 2.15A load, and 64.6mV with a 3.92A load, all at 12V. Also consider whether you need fast or slow blow. Slow blow fuses are necessary in anything which has a start-up or surge current, like motors or incandescent globes.

Other things to consider with a fuse is the mounting style. Some are easier to replace than others, and some holders are available in PCB only, some line only, and some both. Make sure the voltage rating is high enough for your intended use, too. Fuses may not be the best idea when vibration is an issue, either, as the active element, the wire, is fragile and prone to damage from vibration even in its casing, if the current through it is near the limit. Sometimes, even without the current limit being approached, fuses fail due to vibration. Automotive blade fuses are designed to be rugged and cope with vibration, but they also have a tendency to require a greater overcurrent to blow. This is rarely a problem in a car where the greatest danger is chaffed or damaged wiring, leading to a direct short circuit.

Some devices are only called a fuse for convenience. Polyfuses or polyswitches, the common name for a Positive Temperature Coefficient device, also called a resettable fuse, are one such device. These are made with a non-conductive polymer full of conductive carbon. It is made so that the polymer, when cold, is in a regular crystal lattice and the carbon is forced into the spaces between the polymer, forming a grid of conductive material.

As the device heats due to current, however, the lattice starts to change shape, expanding and reaching a point where the carbon conductive paths are broken. This cuts the circuit completely and they remain this way until cooled enough to make contact again, resetting the circuit.

Polyswitches have two ratings, rather than one. They have a hold current, which is the current the device is guaranteed to pass without cutting out. Then, there is the trip current, which is the current at which the device is guaranteed to cut out. The value at which the device actually does cut out could be anywhere in between. This is similar to the tolerance value on a resistor or capacitor, except it is generally expressed as two values rather than a percentage. Like resistors and capacitors, different tolerance ranges can be found, with the tighter the tolerance, the higher the price. If a very significant overcurrent occurs suddenly, it is still possible to permanently blow a polyswitch.

When buying polyswitches, you may find only one tolerance available on the retail or domestic market. In addition, they are usually labelled with a device number, not a rating, meaning that you will have to look up the datasheet to find the hold and trip currents for a particular device.

That is not necessarily a bad thing, as a case stamping or printing would only show one or the other, and you need to know both most of the time. While you're there, check the on resistance as well, because polyswitches have a voltage drop just like fuses do, and it may or may not be big enough to matter for your circuit.

Circuit breakers are another resettable option. In general, they are a manually reset device, where the polyswitch is an automatically resetting device. Most take the form of some sort of switch with a bigger body casing and some goings-on inside of that. There are many types but two dominate the market that a maker is likely to encounter with PCB or car level work. Many other types are industrial or large-scale, or specialised.

Some circuit breakers are thermal, and feature a bimetallic strip. All materials, and metals in particular, expand when heated. All metals expand at different rates. If a flat strip of metal is made by laminating two different metals side by side, then the strip will flex as one side expands more

than the other when heated. A thermal circuit breaker can be made by adding a contact to each end of this strip. One end is fixed, while the other end has a pad type contact that rests against another, like a switch. As current passes through the strip, it heats up, eventually pulling the contacts apart. When the strip cools, the contacts are connected again and current can flow.

Many of us are familiar with these, as they are part of the temperature control device in most fan heaters. Many bimetal strip circuit breakers have a spring with a screw adjustment. This applies a varying amount of pressure to the strip, pretensioning it so that less heat is required to bend the metal enough to open the circuit. In addition, the shape is often a complex curve with two semi-stable states, bent one way, or the other. This helps add hysteresis, or an upper and lower value for tripping and resetting, so that it does not vibrate or oscillate when close to the threshold.

Bimetal strips are cheap and easy to manufacture, and can be quite rugged in addition to ease of adjustment. Their main disadvantage is that they can only be made to cope with a certain amount of current, and can rarely be made with real precision except by spending a lot of money. As a result, the tolerance is poor. The other main type of circuit breaker contains a magnetic coil and a solenoid which pulls or pushes against a set of sprung contacts. The more current that passes through the coil, the stronger the magnetic field, and therefore the stronger the pull on the contacts. At a critical point, the contacts spring open, and the circuit is broken.

Generally, circuit breakers of this design are manually reset by the pushing of a button or moving of a lever. Most household circuit breakers work this way, as do the increasingly-common automotive relays. Some can also be made to be automatically resettable, but these are less common. Also note that this information applies to DC and low-voltage circuit breakers. Mains household ones work differently.

Some circuit breakers are small enough for maker projects. The image earlier shows some small chassis-mount varieties suitable for power supplies or the like, and even some which replace the blade fuses in cars and maker projects which utilise blade fuses. All circuit breakers involve a voltage drop, as there is some resistance in the coil. However, the magnetic reactance of the solenoid also comes into play, but this is variable. The more current flowing through the coil, the stronger the magnetic relationship, and therefore the more back-EMF there is. Circuit breakers may be a good choice for a circuit which normally operates at low current but sometimes has high demands, because losses will be low in the low-current state.

However, circuit breakers can be slower to act than other forms of protection. This is not as true now as it once was, however, as technology develops, but be aware of it. The data for most devices, whether a polyswitch or circuit breaker, should contain the response time. It should also contain the hold and trip currents, which are a thing for circuit breakers as well as polyswitches.

There is also the thermal fuse to consider. These are not quite polyswitches, but more a temperature-controlled fuse. Once blown, they need to be replaced. They have a temperature rating and a tolerance, but the current limit is the same as for any other component - higher than this will damage the device. They are not current-tripped devices and are not used for overcurrent protection.

We have included them here because they are often encountered when someone is trying to fix a failure in anything designed to heat. They are commonly found in ovens, sandwich makers, waffle makers, and so on. They permanently interrupt the current to the device when they have exceeded their temperature rating. Therefore, their value is usually chosen to be above the expected operating temperature of a device but below the danger point where the case or internals of the device would melt.

The idea is that if a fault occurs and the heating element does not turn off, or in some other way becomes too hot, the thermal fuse acts as a backup before catastrophic damage occurs. While only a licensed electrician is legally allowed to work on mains, even if it's an appliance that plugs in, we know that people do anyway and feel that in this case, it's worth discussing.

We have heard of people just replacing the thermal fuse and thinking they have fixed their device. This is actually quite dangerous. Rarely does a thermal fuse blow unless something else has gone faulty in order to cause too much heat. If you are fault-finding such an appliance and find a thermal fuse blown, keep looking! You have likely not found the true fault! The exception is designs where the thermal fuse is close to the expected operating temperature. With time, because the fuse is so close to its trip temperature for so long, they can fail without another fault causing it.

Many people have also encountered the Residual Current Device, or RCD. Many people think these are a current protection device for circuits but in fact, they really are not. Although they could be used as such in certain circumstances, they are really a safety device. Part of the confusion comes from the fact they are often built into current-activated circuit breakers for household use.

The diagram above is heavily simplified but shows the principle. These are generally AC devices for use on mains but they can be found in DC applications, too. We'll describe the simpler and far more common AC version. The active is at the top, labelled 'In'. Note that current changes fifty times per second (50Hz) in AC (except the United States or countries which were industrialised with US support, in which case it's 60Hz), and so current flows both ways in the circuit. The current flows to the load and back to neutral but both the in and out wires are wrapped around a metal core. We drew this as a rectangle for simplicity but it is usually a ferrite toroid. The turns have the same number and direction for in and out.

There is one more coil in the middle, connected to a solenoid switch. If all is well, the current goes in and out, and the turns on either side of the coil generate a magnetic field which cancels the other out. At the middle coil, there is a net zero influence. If any current leaks out, however, such as through a wiring fault to ground or through a person, then the coils are unbalanced between top and bottom as more current goes in than out. There is now a magnetic field at the middle coil, and because it's AC and changing, a voltage is generated which activates the solenoid switch, effectively a relay. This cuts out the power and disconnects the load from the supply.

RCDs can commonly be bought in 10mA, 20mA, and 30mA versions. That is the amount of current that can leak before the RCD trips and cuts power. The amount of current flowing through the device is irrelevant as long as it is balanced. Most RCDs are built into a circuit breaker that is designed as an overcurrent device, leading some people to think that RCDs are a form of overload protection. They are not, they are a safety device to prevent leakage. You will rarely need one in a DC situation or really, any maker situation.

Whether a fuse, polyswitch, or circuit breaker is the option for your overcurrent protection needs depends on the specifics of your circuit and its intended purpose. However, hopefully, now you have the information to make informed decisions and do some further research if needed, knowing what to look for and read up on. Next month, we will cover some other circuit protection topics like over- and under-voltage protection, short-circuit clamping, and noise reduction.