We take a deep look at this iconic 50-year-old timer IC, how it works, and what makes it tick*. Then, we build a huge one! - by Daniel Koch.

The venerable 555 timer integrated circuit (IC) turns 50 this year, but is still as relevant today as when it was new. Designed toward the end of 1971, the device was released to market and first sold in 1972. We still haven't found an exact date. It did not take long for the IC to become a success, and it is now one of the most-produced ICs in history. Thanks largely to the combination of its simplicity and its versatility, the 555 Timer IC has become genuinely ubiquitous. We will explore the device itself and its variants, its operation and different functional modes, and then finish by making a giant one from discrete components.

*Yes, that terrible pun was intentional.

A BRIEF HISTORY

The 555 came about when Hans Camenzind was exploring Phase Locked Loops (PLL). It was during this process, developing oscillators to run PLLs, that he had the idea for an integrated circuit timer based on the charging of an external capacitor. Initially, a constant current source was part of the design, but this was dropped after he realised that an exponential RC curve would still work reliably. While Camenzind was originally an employee of Signetics, he left to write a book after the PLL work was done. The 1970 economic recession had hit, and he proposed returning as a contractor for Signetics instead of working under the available employment arrangements. His contracting rate was still lower than his old pre-recession pay rate, but he was allowed to borrow equipment that was not being used after mass redundancies.

This contract was specifically to develop a timer IC from the work done on oscillators for PLLs. While some of Signetics' engineers did not like this idea, marketing manager Art Fury did, and so the go-ahead was given. In the second half of 1971, Camenzind submitted a report on his work, with a design using a constant current source. While this report was being reviewed by the company, an employee who had been party to the process left and took the information to a rival company.

During the time it took for that company to develop a production design, Camenzind realised that a regular RC curve with external fixed resistor would work fine, and thus, the constant current source was no longer necessary. This reduced the number of pins needed from nine to eight, meaning that it would fit in an 8-pin DIP package and not a 14-pin DIP.

This design is the one Signetics pursued. While the rival company beat Signetics to market, their product lasted three months. Art Fury was the one, according to Camenzind himself, who chose the number '555' arbitrarily. Popular belief holds that the IC was named for the three 5kΩ resistors in the voltage divider. However, Camenzind relates that it was named because Fury, very much the marketing man, expected the IC to be a significant success and wanted a part number easily recognisable and iconic. 555 was that number, and NE is Signetics' prefix.

In 1972, the NE555 was released to market and success was nearly instant. It is one of the most popular ICs ever produced, although some sources which estimate that 'billions are produced every year' are probably overstated, considering that the world's population is a little under eight billion and no sources are given for the estimate. By the end of 1972, twelve other companies were producing the NE555 (there was a trend at the time not to patent designs for ICs, for a variety of reasons), and it is still in production today. Signetics became part of Philips Semiconductor, which was itself spun off from the Philips group to create NXP, who in turn divested their discrete semiconductor business into the independent Nexperia, which still produces the NE555.

AVAILABILITY AND DATASHEETS

Along with Nexperia, current manufacturers include: Texas Instruments, Harris Corporation, ON Semiconductor, Analog Devices, Renesas, Advanced Linear Devices, Diodes Incorporated, Microchip Technology, National Semiconductor (bought by Texas Instruments but still run as a brand), ROHM Semiconductor, and STMicroelectronics.

These are companies which we found listings for on the major wholesalers' websites. There is likely to be even more manufacturers than these. In addition, there were other major manufacturers in the past who have since been absorbed by some of the current manufacturers, or folded entirely in some cases.

Notable examples were National Semiconductor as above, Fairchild (bought by National Semiconductor, then divested as independent again, then bought by ON Semi), Intersil (Acquired by Renesas), Motorola (spun off as ON Semi but also related to Freescale Semiconductor which was bought by NXP), and Maxim Integrated (now part of Analog Devices).

The end result of all this is that the 555 Timer can be bought in a bewildering range of brands. Some of the merged or bought names are still used as brands despite being owned by other names who also produce the IC, and in other cases new old stock^ of retired brands or companies still exists. Datasheets are even more of a challenge, as there are copies floating around the internet for most of the brands that have ever produced the 555, even if these brands are long since defunct.

Because of the differences between manufacturers in any device, it is always best practice to find the datasheet that matches the brand of device you have. In the case of the standard NE555, there is little variation, but you may still find maximum voltage, quiescent current draw, and current handling vary between manufacturers. When it comes to other variants, the differences can be more pronounced.

^New old stock is a term meaning stock that has been stored for a long time but never used, and is in new condition. It generally applies to products that are no longer made, or stock from companies which no longer exist. The term is often encountered when searching for items that are becoming hard to find, but not always. In the case of the NE555, it may be used to inform why stock may be branded as, say, Signetics, when signetics have not made any for so long. This is important in the face of Chinese counterfeit copies of many components available today, which are not made to the same standards.

THE DEVICE

Parameter	Value	Unit
Supply Voltage, Typical	4.5 - 16	V
Output Current	+/- 200	mA
Operating Temperature	0 - 70	°C
No-Load Supply Current	15	mA
Output Rise Time	100 - 300	ns
Output Fall Time	100 - 300	ns
Minimum Trigger Pulse	100	ns
Maximum Frequency	100	kHz

The functional block diagram is the simplest representation of the inside of an NE555. It's certainly easier to follow than the internal schematic! Both diagrams are colour-coded, with the sections highlighted in one colour on the functional block diagram match the components drawn in that colour in the internal schematic. It is hard to pin down the 'heart' of the device. The flip flop is useless without the comparators and the comparators are useless without the voltage divider, but the comparators are pointless if not driving the flip flop. As such, we'll move left to right.

The dark green section is the voltage divider formed from three identical resistors. It can be seen connected to the supply rail and ground at the outer ends, and the comparators at the junctions. Note that the divider is connected to the inverting (-) input of the blue threshold comparator, but the non-inverting (+) input of the trigger comparator. This section is where the values of one third and two thirds come from for the comparators: The supply voltage is divided across the resistors evenly, into thirds, so the junctions are at one third and two thirds.

The threshold comparator is depicted in blue. With its inverting input connected to the upper junction of the voltage divider, when the voltage on its non-inverting input increases above the inverting input, the output changes. The output is high when the voltage on the non-inverting input is greater than the voltage on the inverting input. The output is connected to the flip flop's Reset input.

The trigger comparator shown in purple is wired the opposite way, with the voltage divider reference connected to the non-inverting input, and the trigger input variable fed to the inverting input. When the voltage on the trigger pin is below the reference set by the voltage divider, the output is high. This is connected to the Set input of the flip flop.

The flip flop is shown in brown, and is an SR (Set/Reset) type. However, there are differences between this and the flip flops we covered recently in Classroom. Only output Q-BAR is used, but it is shown on diagrams as a single Q output with an inverter immediately after it, denoted by the small circle. There is also an external reset input, which is active low. It is generally tied high until needed, which is why many circuits have pin 4 connected to the supply rail.

The lime green section is the NPN discharge transistor. This is activated by the output from the flip flop, and is connected internally to ground on the emitter. The collector is open and taken to pin 7. This transistor also handles the discharge current from the capacitor, and has a 50mA rating for most standard NE555 versions. Therefore, the total current from the resistor between pin 7 and Vcc (RA in most generalised drawings), and the capacitor discharge current through the relevant resistor (RB in most generalised drawings), cannot exceed 50mA. The datasheet will give exact details for a specific brand and model.

The orange section is the output. Not shown in the functional block but visible in the internal schematic is the fact that this section has connections to both supply and ground. This is how the output can both source or sink current. When high, it supplies (sources) current from the supply rail. When low, it sinks current to the ground rail. It also has an inverter on its input, so the final output from pin 3 is opposite whatever is going on with the discharge transistor.

So far unmentioned is the control voltage input on pin 5. This is connected to the voltage divider at the junction of the first and second resistors, which is the two-thirds value point. It is used to modify the voltage at which the threshold comparator changes. Adding voltage here will increase the point above the two thirds value, lengthening the timing signal. Note that because the current fed to it also flows through the other two resistors in the divider, it is divided across those two as well and will alter the value at which the trigger threshold is reached as well. This alters both frequency and duty cycle at once. Control voltage can be varied between 45% and 90% of the supply voltage, though some sources have other figures.

In the majority of situations, however, the control voltage pin is tied to ground with a 10nF to 100nF decoupling capacitor to prevent noise and instability. Many makers assert that the capacitor is unnecessary and that they have never had a problem. However, it would be fairer to say they have never noticed a problem. Noise and spikes on this pin can cause very fast transient changes that may never be noticed by the user, but may still produce varied results. Fitting the capacitor is still best practice, and is a fairly cheap exercise in caution.

PART NUMBER VARIATIONS AND OTHER VERSIONS

So far, we have discussed the NE555. That is the original Signetics part number, and the one by which the IC is very commonly, but not universally, known today. However, other part numbers exist as well and sometimes, this is important. Some part number, such as the MC1455 from On Semiconductor, are just manufacturer's names for their copy of the NE555. Other versions, however, have a part number that means something. The SE555 is the original Signetics part number for a higher-rated version of the NE555. It has a slightly higher maximum voltage, tighter tolerances, and a much greater temperature range. In addition to this, the suffixes can carry meaning as well. Sometimes it is the package, while at other times, the guarantee. 'M' part numbers from Texas Instruments, such as the SE555M, are military rated. Note there is no NE555M, because part of the military spec is the higher temperature rating of the SE part. The datasheets always have this information in them, often toward the end where a breakdown of part numbers is located. Unfortunately, every manufacturer presents this information differently, so we cannot guide you much.

In addition to variations on the basic NE555, there are more recent versions which are drop-in replacements but different internally. The differences may be minor or major. Most manufacturers who list these still incorporate 555 in the device name, but with different pre- and suffixes. Harris Semiconductor makes a 'CA555', which is listed as a 'highly stable replacement' for the regular NE555 varieties. Unfortunately, the datasheet is shared with the company's NE555 options, so it is hard to tell what changes have been made because the schematic is the regular NE555 layout. If ever you are unsure, check the datasheet for each product. The 'Description' paragraph on the first or second page generally makes clear what the device is.

The NE556 and its related part numbers are also worth a mention. This is a 14-pin DIP package with two NE555s in it, with common Vcc and GND rails. All other pins are independent. It is used in circuits where two NE555s are used side-by-side, like our recent servo driver project. However, in that case we deliberately avoided it because we found the breadboard-format layout easier with separate ICs. If we were designing a PCB, that would not have been the case and the NE556 would have been useful.

Finally, there are CMOS versions of the 555 series. Some are called a '7555', while some others are listed with prefixes to denote the CMOS, like Texas Instruments' TLC551CP and TLC555CP. While the majority of the 555 range is made with bipolar transistor logic, these are versions made with Complementary Metal Oxide Semiconductor transistors. They are much faster than the bipolar equivalents, have far less current draw, have greater stability and accuracy, and have high-impedance inputs rendering them suitable for longer timing durations and smaller capacitors. They have a lower minimum supply voltage down to 1.5V in some cases, better temperature stability, and reduce the prevalence of current spikes in the supply line. In return for this, they have a much lower current capability of around 10mA sink and 50mA source depending on brand, are more expensive, far less rugged, and are static sensitive.

We have direct experience with this, having destroyed several accidentally in one session while regular 4000-series chips, which are also considered static sensitive, survived. That does not mean that all brands are that sensitive or that you should not try a 7555. The 7555 may have been a better way to solve some of the challenges we encountered recently when turning the servo driver into a PWM motor driver, but we avoided it for the above reasons. It would have been a good way to gain faster frequencies and a cleaner waveform, however.

MODES OF OPERATION

We will discuss four mainstream modes of operation: Astable, Monostable, Bistable and Schmitt Trigger.

ASTABLE

Astable mode is arguably the most common mode that many makers will use the NE555 in. Also known as a multivibrator, astable mode has no stable state and instead oscillates between high and low. This oscillation is predictable and controlled, and was the main impetus for designing the NE555 in the first place, after Camenzind's search for a suitable clock generator for PLLs.

One of the most common electronics learning example circuits using the NE555 is to flash two LEDs. This is a great example of how the output can both source and sink current, and we have used it in the diagrams explaining the charging and discharging phases of the timer.

In astable mode, two external resistors and one capacitor set the frequency, and high and low times. When power is first applied, current flows via RA and RB to charge the capacitor CT. This occurs because the discharge transistor on pin 7 is off. The flip flop is in its set state, which, because of the inverter, means its output is off.

The inverter on the output driver stage means that it is on during the charging phase. The capacitor is connected to both pins 2 and 6, which are the trigger and threshold comparator inputs respectively. The first timing cycle is longer because the capacitor charges from 0V, all the way up to two thirds Vcc (supply voltage). The diagram shows current flowing into the comparator inputs because it does, but only at a small rate. The bulk of the current follows the resistor/capacitor path.

Marginally past the two thirds point, the voltage on the non-inverting input of the threshold comparator exceeds the voltage on the inverting input, and so the output goes high. This resets the flip flop, the inverted output of which goes high. Remember, because of the inverter, the reset input turns the output on, not off as would normally be in an SR flip flop.

With the output of the flip flop high, the discharge transistor is active. The output buffer is also active, but it is inverted too. Therefore, the output is low, sinking current, whenever the internal flip flop is high. The discharge transistor now discharges the capacitor, but only via RB. RA is still pumping current through the discharge transistor, which is one of the reasons to limit the lowest value ever used for RA.

As CT discharges, the voltage across it falls, until it dips just below one third of Vcc. This activates the trigger comparator, the inverting input of which is used this time as the variable input, with the non-inverting input being tied to the voltage divider as the reference. Therefore, when the variable voltage falls below the reference voltage, the comparator's output goes high, setting the flip flop.

The flip flop now sets, which would normally mean a high output, but the inverter means the output is now low. This turns off the discharge transistor, allowing CT to begin charging again. It also turns on the output stage at pin 3, because of the inverter present on that section's input. Thus, the cycle repeats. This inverted output may seem silly at face value, but it has its uses. The first timing period mentioned above is longer because the first cycle charges the capacitor from 0V, while every charging cycle thereafter has the capacitor starting at one third the supply voltage.

To calculate the times involved, there are some formulas. You can calculate for known values to find time, or decide a time and find values. When doing this, we find it easier to choose a capacitor value and then find resistors to suit, as there are more values available. Capacitors tend to jump further between values. Of course, guess and check is not the most engineer-like method (or is it?) but it is the one within reach of the most people.

Remember that all calculations must be done in base units. Don't forget that Farads are the base unit. We get a lot of enquiries about maths in articles where it turns out people forget that microFarads are actually a millionth of a Farad, because we very rarely use anything bigger than µF. So, convert nano- and picoFarads into Farads, not microFarads! We have included a look-up table at the end to help. Also, see pages 45 and 46 of the Servo Controller project in Issue 57 for more. You can access this free online.

In the formulae above, T is the total period of the oscillation, including the high and low. T1 is the high, while T2 is the low. ƒ is the frequency in Hertz, and of course RA, RB, and CT are the components shown in the diagram. Sometimes, however, you want to construct the circuit for a certain time period or frequency, and find component values to suit. All mathematical equations can be transposed, which is the maths term for rearranging without altering the facts of the equation.

The general rule is to do to one side what you do to the other, but things must stay equal. For example, if something is multiplied by RA on one side, then moving it to the other side means dividing that side by RA. To save those unpractised at maths the trouble, we show the formulae here for finding resistor values. Choose an arbitrary capacitor value, because there are far fewer of them, and use guess-and-check until you get resistor and capacitor combinations to give the desired frequency. With the total resistor value of RA and two lots of RB calculated, you can substitute values into the regular formula above until you have numbers that work.

There are some points to note with the astable circuit. T1, the high time, cannot be less than the T2 low time. This can be a problem in some circuits when a short clock pulse with a long interval is needed, but luckily, the output both sources and sinks, so an inverter is possible. For a short high time, construct a circuit with a long T1 high time, a short T2 low time, and use the sink capability of the output to control a PNP transistor. Also of note is that changes to the resistor values affect both timing and overall frequency. That is not a problem in very basic circuits like Choppy from Issue 9, but was an issue for servo driving in Issue 57. There is a minimum trigger pulse, which in part determines the maximum frequency of the astable circuit, 100kHz. RA should not be less than 1kΩ, although less is permissible if managed carefully in some circuits. If RB is significantly larger than RA, then an almost square wave can be produced. It is not possible to have an exactly 50% duty cycle, but it can be close. Frequency and duty cycle are linked in the basic astable circuit, so any change to RA, RB, or CT affect both.

ASTABLE VARIATIONS

Several variations on the basic astable exist. Earlier, we discussed the inverted nature of the flip flop output and the input to the output stage. This becomes useful here, because pin 3 can be used instead of the discharge transistor to create a 50% duty cycle square wave. If connected as in the diagram, then the capacitor is charged via the source mode of the output, then discharged via the sink mode, both through the same resistor and therefore with the same charge and discharge times.

There is also a way to gain independent control over charge and discharge times. The use of diodes can give two different circuits. One gives separately controlled charge and discharge times, while the other provides a duty cycle alteration while maintaining the same frequency because the total resistance has not changed, only the relationship. However, these circuits are not perfect. The presence of diodes means the circuits are now more susceptible to variations in supply voltage. Diodes have a fixed voltage drop, and this affects calculations made based only on resistance. While a purely resistive charge and discharge path is independent of voltage supply, one with diodes in it introduces an invariable element.

MONOSTABLE

The other mainstream mode of operation for an NE555 is monostable, meaning only one stable state. That state is low, and the trigger pulse causes a transition to high on the output, for the duration set by the timing components, before the output reverts to low and stays there until the next trigger.

The timing period here is the charging of the transistor from 0V to two thirds Vcc, because the trigger comparator is not connected to the capacitor, and it discharges to ground via the discharge transistor on pin 7 after the timing period. This dictates a minimum low after the timed period has expired, for the capacitor to discharge.

It also means care must be taken that the capacitor current never exceeds the 50mA discharge transistor limit. This will not be a problem on short durations with small MKT, ceramic, or greencap capacitors, but long delays with large electrolytic capacitors may elicit a problem.

It also means that the timing resistor RT must be big enough not to overload the discharge transistor with too much current while the transistor is held low, and the combination of capacitor discharge current and RT current must also stay below 50mA.

In basic monostable mode, the input pin 2 is held high, at Vcc or at least above one third Vcc. Generally, a simple pull-up resistor to Vcc is the easiest thing to implement but if the input signal is stable and of high enough voltage, the pull-up resistor is not needed. At first, the flip flop is in its reset state, but its inverted output is high, holding the discharge transistor on. Current flows through RT straight to ground via the discharge transistor, and CT is also kept completely discharged.

When the trigger signal falls below one third Vcc, the trigger comparator output goes high, setting the flip flop. Its inverted output now goes low, removing the short circuit provided by the discharge transistor, and allowing current from RT to charge CT. During this time, the output stage is also high, giving the output pulse that is used beyond the circuit. Because the junction of RT and CT is connected to pin 6, the threshold pin, as well as the discharge pin 7, the voltage rising across the capacitor is monitored until it reaches two thirds of Vcc. At this point, the flip flop resets, turning off the output and turning on the discharge transistor which allows CT to rapidly discharge. Because there is no resistor in the way, the discharge is limited only by the capacitor's internal equivalent series resistance, and the resistance of the current path through the transistor.

In monostable circuits, the reset pin 4 is more likely to be used. It can terminate the timing cycle immediately, regardless of the status of the trigger or threshold comparators. Some circuits use this for external control, particularly in very long delays, while others do not need it and tie the reset to Vcc as for an astable circuit. After the first trigger input, the timing period must expire before any other trigger pulse can activate the circuit. This means the monostable circuit can be configured as a divide-by-n counter if the output pulse is carefully timed.

The trigger comparator switches on the falling edge of the trigger pulse, when the voltage goes from high to low. However, the reset terminal is low-going as well, but timing starts on the rising edge, as the reset pulse is removed, if the trigger is still low. Because of the way the NE555 is configured internally, the output stays high if the trigger pulse is longer than the timed period, until the trigger pulse goes back high. This must be factored into designs needing very short output pulses: The trigger pulse cannot be too long.

The minimum possible pulse is about 5µs, while the maximum is in the order of hours. In practical terms, however, a maximum of five minutes is often quoted. This is because the high-value components needed introduce tolerance and stability issues that make periods longer than five minutes irregular and unpredictable, or at least not predictable with acceptable accuracy. The minimum value for RT should be around 1kΩ, while the upper limit is generally best kept below 1MΩ despite the fact that theoretically, 20MΩ is possible.

If the circuit is configured to trigger on the edge of the low-going pulse, the problem of the circuit remaining high for longer than one trigger period goes away. This is done with a capacitor, a diode, and a resistor. The capacitor discharges into the low of the trigger pulse, but then is recharged by the resistor, giving a minimum length to the trigger pulse. When the trigger pulse goes high again, altering the conditions of the capacitor, the voltage pulse is dumped to Vcc by the diode.

The maths for the timing of a monostable is a little easier than for the astable because there is only one resistor RT and capacitor CT involved, no frequency, and only one timed period, T. The time of the high period is T = 1.1 x RT x CT. Remember to work in base units.

BISTABLE

It is possible to utilise the flip flop inside an NE555 to create a high-current flip flop, as the 200mA current capability of the output far exceeds most dedicated bistable ICs. There is no timing capacitor involved, and the threshold and trigger inputs are tied to the rails with 10kΩ resistors to stop them floating.

However, in its basic form, the circuit is triggered with a negative or low-going pulse for pin 2, but a positive or high-going pulse for the reset input on the threshold pin 6. Low-going means the input is normally high (held there in this case by the 10kΩ resistor) and is taken low to trigger. High-going means the input is normally low, and the input is a high pulse.

This is good for some circuits, but not for others. An inverter circuit can be created in either case, to give two low-going inputs, or two high-going inputs. Additionally, the 10kΩ pull-up and pull-down resistors may not be needed if the input signals are consistent.

In other words, if the trigger input has a constant voltage that stops at the right moment, this will keep the trigger pin high. However, if the external input is not a supplied voltage but rather a connection to ground at the right time, the pull-up resistor is needed.

SCHMITT TRIGGER

The NE555 can function as a Schmitt trigger by virtue of the two comparators. Used for converting AC signals to DC signals, this circuit uses capacitive coupling at the input and a voltage divider to bias the input to half Vcc. Because the comparators already have their references set at one third and two thirds Vcc, the other inputs (pins 2 and 6) are tied together to form the new input. Because the trip points are one and two thirds Vcc, the output waveform will not be in phase with the input waveform.

LIMITATIONS

It is important to keep in mind some of the limitations of the NE555 when designing with it. Some we have covered, such as the 5µs trigger pulse minimum and its effects on maximum frequency; the 50mA limit of the pin 7 discharge transistor; and the 1kΩ minimum value for RA.

This last point is somewhat variable, as explained in the astable section. However, if there is not a pressing need to reduce the value below 1kΩ, don't do it. We have used values as low as 100Ω in the past, but always at a low enough voltage and with other factors in play that meant the discharge transistor was not overloaded.

One limitation we have not covered comes from electrolytic capacitors. Two facets of these components need to be kept in mind. The first is leakage. For long time delays, large electrolytic capacitors are needed. However, they do leak considerable current across them, enough to lengthen the time delay well beyond that calculated.

The situation can be likened to a bucket with a hole in it. A hose will fill the bucket as long as the water flowing out through the hole is flowing slower than the water coming in through the hose, but it will take longer than a bucket with no hole. There is a firm correlation between the size of the hole and the amount of extra time needed to fill the bucket. With electrolytic capacitors, the relationship is still there, but non-linear. Dedicated low-leakage capacitors are needed for long time delays.

The other property of electrolytic capacitors concerns their behaviour related to voltage. An electrolytic capacitor does not start behaving as such until the voltage across it reaches approximately 10% of the value of its rating. So, for a 16V capacitor on a timing circuit, that's not a problem.

However, if using a 63V capacitor on a 5V circuit, the calculations will not be accurate because the normal RC curve has been altered by this property. The capacitors will still work, but the timing calculations will not be correct. Sometimes, this alters the choice of capacitor. For example, most retail ranges of electrolytics use 63V as the only voltage available for values like 1µF or 2.2µF.

THE 555 AS YOUR BEST FRIEND:
CRITICISMS AND VALIDATIONS

The 555 is not without its problems and detractors. While we love it as makers, many engineers do not love it at all. Bob Pease, one of the most famous and accomplished analog engineers with a long career designing ICs at National Semiconductor, once said: "My favourite circuit to use a 555, is: A blank piece of paper".

The frequency is limited compared to more modern chips, and the consistency is not there either. The output characteristics can vary with voltage fluctuations, and there are limitations on what an RC timing circuit can achieve compared to a crystal or ceramic oscillator. However, many of the challenges that concern an engineer working with the NE555 do not concern the maker, especially considering there are no production and commercial considerations.

The 555 iterations are not great with long timing delays, and the comparators do have a small region where they behave as linear amplifiers that is noticeable at long time periods. Output switching is not as crisp as more modern designs, and again, problems show up at long time periods.

While there are many better options available to the engineer, there are still plenty of reasons to use an NE555 or variant in less critical roles. For the maker, it does the jobs generally asked of it very well. Within its modes, it is versatile and adaptable.

It doesn't take any particularly special knowledge to use it, like advanced maths or coding languages, or the ability to see tiny but important details obscured in datasheets. It has a wide voltage range and is ruggedly dependable, with a reputation of being hard to kill.

There are few situations where it will lock up or hang, so it tends to keep doing its job indefinitely unless these rare conditions are met. There are still disadvantages, such as the 4.5V minimum effective working voltage.

However, versions exist with lower voltages, and few Ics or MCUs have a 200mA current capability. Few can take the abuse that an NE555 can take, either. The biggest advantage of the NE555, however, comes from the sheer volume of information available.

Entire books have been written about it. You can see in the image that these two are very well used. The datasheets from some suppliers also have more than the usual amount of information and application example circuits. You can even find look-up graphs online which show approximate timings for different combinations of capacitors and resistors, and online calculators, too.

If it can be done with an NE555, someone has, and you can find it online. It is important to note, though, that just because it can be done with an NE555, does not mean that it should be. Sometimes it is a poor choice, while at other times it is just not the best choice in pure engineering terms. However, what constitutes the 'best choice' may be varied by other factors. For example, a device which functions at higher accuracy may be not available over the counter, or the knowledge required to use a particular alternative effectively may not be within reach.

The fact is, every device has advantages and disadvantages, and the NE555 is still in mass production for a good reason. It may rarely be found in new commercial designs, but it will remain a hobby staple and experimental tool for many years to come. It was ground-breaking at the time it went to market, and it remains a sound choice within those caveats above. While Hans Camenzind passed away in August 2012, we still salute him for this innovative creation!

The Build:

The build to go with our celebration of the NE555 is a bit different to the usual Classroom fare. This month, we're assembling a kit we bought from Evil Mad Scientist in the United States. This great product is an NE555 built with discrete components: A whole bunch of resistors and transistors individually applied. The whole lot is soldered to a very well-cut, thick, plated-through PCB, with screw terminals for the pin connections and a stand made of aluminium to represent the legs of a DIP IC package. We first discovered this while background researching some of the more obscure points raised in the article.

The kit has a lower output rating than the real NE555: 100mA versus 200mA. This is due to the transistors used. The kit is made with 2N3904 and 2N3906 transistors, which are not ones we commonly use here at DIYODE. They are

extremely popular in the US where the kit is designed, and many an online circuit uses them too. In Australia, there has long been a preference for the BC series, like the BC548 and BC557. There is little reason to fight about it one way or another, it's more a preference thing in most experimental circuits.

The only thing we changed about the kit is to replace the three 4.7kΩ 5% tolerance resistors that came with the kit as the voltage divider components, with 5kΩ 0.1% tolerance metal film resistors. These are hard to get and much more expensive than standard resistors, so we had to order them from Element14.

There is no reason to use them except for the sake of purism. If you buy one of these kits yourself, there is nothing wrong with using the included resistors. In fact, the 0.1% resistors may well be a better tolerance than the real ones etched onto an NE555! We're not sure. However, while the tolerance is unknown, the value of the resistors in the real NE555 is definitely 5kΩ, which is not a standard value. 4.7kΩ is the closest from the most commonly-available resistor series, while more comprehensive ranges include 5.1kΩ.

BUILDING THE KIT

The kit comes with a printed sheet which details the parts list and assembly steps. It also covers some mechanical tips, like making sure to bend component legs under the board before soldering, and some soldering method tips. On the reverse is a set of assembly steps. Online, you can access a PDF with information on the 555, a schematic, specifications, and example circuits, all oriented around the kit.

We found the suggested assembly steps were inline with our normal practice anyway: low-profile components first, then working upwards in height. Accordingly, we did just that. However, before starting, we cut off all the transistors from the tape, and straightened the legs. The PCB has close spacing but this particular batch of transistors had bent legs. This was a tedious but worthwhile and relatively pain-free task that did not take as long as we imagined once we had the right pair of pliers.

We first determined which resistors on the board are the voltage divider, from the schematic on the online instructions and product datasheet.

Because there are more 4.7kΩ resistors than those used for the voltage divider, it was not immediately obvious which ones on the board were the divider. We ended up replacing R7, R8, and R9 with 5kΩ 0.1% resistors we bought from Element14. We did this for the sake of it, and imagine that most people won't bother and will use the perfectly acceptable included 4.7kΩ units. It makes no realistic difference, except in our heads.

If you do the same as us, be really careful of the colours! The tolerance band for 0.1% is purple, and combined with the fact that 5kΩ is a value with a colour combination (green, black, black, brown, space, purple) we are not used to, it's really unsettling in the back of the mind! We soldered these down on their own first.

Next, we added the rest of the resistors. Being very careful to bend the legs to a crisp 90° elbow at the exact length of the holes makes for a much neater job. The same can be said for the colour codes. An old manufacturing and service industry standard was to have the colour codes read left to right for horizontal, and top to bottom for vertical.

There were variations on the theme, too. The important point is consistency. It rarely gets mentioned or taught in maker circles now, but the practice is particularly helpful when colour codes are not spaced well, are not well coloured, or become faded or damaged. Knowing the order the colours are going often helps determine what the value is when things are ambiguous. It is also much easier and faster when fault finding.

Next was the transistors. These are a mix of NPN and PNP, all of two types: 2N3904 and 2N3906, both 100mA rated. We were tempted to change one of these, Q14, the discharge transistor, with a higher-rated BC337. However, we elected to keep to the kit, with the 5kΩ resistors our only change. Not only did we want to stay with the kit, the gain range of the BC337 has a greater spread.

We worked in groups for the transistors. Even doing this, it could at times be hard to get the soldering iron into all the spaces needed. With too many on the board, it is very easy to miss one or more connections, too. The board is broken into four sections in the screen printing, so we used these as a guide. We took care to align them all as healthy as we could, and make sure they were level. It's not perfect, but we were pretty happy with the result.

There was not much left now. Next was the aluminium legs. These are attached with three self-tapping screws each, with a washer on each screw between the board and metal. Installation involved putting the three screws of one side through their PCB holes, then turning it over with the fingers of one hand holding the screws. Then, the aluminium is lined up with the screws.

We gently let go of one screw, and turned it until it just started to take up and bite, then a little more. Then the next screw, then the last. Then, back to the beginning to tighten each a few turns at a time. This practice makes sure that the whole arrangement stays in alignment and tensions evenly. You can just go ahead and tighten the screws one at a time, but the results are not always sound.

Finally, the only task left was to add the thumbscrews. These fit into threaded receptacles that are factory-attached to the PCB, so they should just screw in. Ours went in very easily, even though it looked by eye that the tolerance on one was going to be close. There are six grey ones, for the function pins, and a red and black one for the Vcc and ground connections respectively. And that's it! A giant NE555.

The real NE555 in DIP8 form is 9mm long. The Evil Mad Scientist (EMS) Three Fives kit is 133mm long. So, it's close to a 15:1 scale enlargement! The kits can be bought from EMS in the United States of America, but postage is expensive.

EMS are looking at solutions and we have been discussing it with them. So, if you're interested in the Three Fives kit, its UA741 cousin, or any of their other awesome products, don't rush. Instead, watch this space, because we hope to bring you news on this topic soon!

Want One?

If you do want to buy one straight away, head to https://shop.evilmadscientist.com/directory

All prices are in US dollars, and shipping can be quite expensive, especially if you want tracking. EMS only charge what they get charged by the United States Postal Service or United Parcel Service, but use the quote feature first!