[FAQ] How do I design a-stable timer, oscillator, circuits using LMC555, TLC555, LM555, NA555, NE555, SA555, or SE555?

This FAQ covers a-stable circuits. Figure 1 below shows an a-stable circuit using the LM555, NA555, NE555, SA555 and SE555 timers that are called LM, NA, NE, SA and SE respectively hereafter. As a group they are called bipolar timers due to their design. This FAQ also covers LMC555 and TLC555 timers that will be called LMC and TLC respectively hereafter. As a group they are called CMOS timers due to their design. This FAQ is applicable to the xx556, TLC551, and TLC552 timers as well. Supply voltage pins for the timers use different symbols, V_S, V_CC, V_DD, that have the same function. 

Figure 1. Basic a-stable circuit

 A-stable mode, as the name implies, has no stable state. The output alternates between low and high at a rate set by two resistors and a capacitor. The a-stable setup ties ‘threshold pin’, ‘trigger pin’, and timing capacitor together so that the timer retriggers itself as the capacitor voltage ramps between 1/3 V_S and 2/3 V_S.

Resistor (R_B) is connected from the timing capacitor to the ‘discharge pin’. This sets the discharge time, where the output is low. Low time is R_B * C * ln(2). Resistor (R_A) between ‘discharge pin’ and V_S sets the charge time, where the output is high. High time is (R_A + R_B) * C* ln(2). The charge current flows through both R_A and R_B. The cycle period is the sum of low and high times, (R_A + 2 * R_B) * C* ln(2). The frequency is equal to the inverse of the time period, 1 / [(R_A + 2 * R_B) * C * ln(2)].

Positive duty cycle will be high time divided by period, (R_A + R_B) / (R_A + 2 * R_B); note that, there is no “C” component in the formula. In other words, duty cycle does not change with capacitance. Positive duty cycle between 50% and 99% can be accomplished with this configuration. The relationship between R_A/R_B ratio and duty cycle is shown in Figure 2. A low resistance value for R_A will increase the current used by the timing resistors. Inverting the output with an external transistor or inverter is one way to get less than 50% duty cycle. 

Figure 2. Positive duty cycle versus R_A/R_B resistor ratio

The simple timing formula given by: 1.44 / [(R_A + 2 * R_B) * C], does not account for propagation delay. The error in the simple formula increases with higher a-stable frequency. The more complex formulas shown in Figure 3 can be used to calculate the expected low and high times more accurately. Propagation delays directly extend the end time of each output state. Propagation delay also creates a larger capacitor swing voltage, which increases the time needed for the next capacitor voltage to ramp. t_PLH is the output low to high time delay and t_PHL is the output high to low time delay. The graphic in figure 3 shows the effect of these output propagation delays.

Figure 3. A-stable frequency formulas including propagation delay where oscillation frequency will be 1 / [t_c(H) + t_c(L)]

Simple 50 percent duty cycle circuit

The output stage for the CMOS timers provide low and high output voltages, namely: V_OL and V_OH. These voltages approach the supply rails under light load. Therefore, a single resistor from output to timing capacitor can replace the two resistors that use the ‘discharge pin’. This provides nearly equal charge and discharge current flow. The V_OH on the bipolar timers is significantly lower than the positive supply rail. So output duty cycle won’t be 50 percent.

The timing capacitor charges and discharges via the same resistor that is powered by the output V_OH and V_OL levels. This leaves the ‘discharge pin’ available as an alternate output. It is acceptable to leave the ‘discharge pin’ floating. The V_OL drive is stronger than the V_OH drive so the duty cycle is slightly higher than 50 percent. However, the variance from 50 percent duty cycle is insignificant when R_C >> V_OH driver resistance. The load on the output can also affect the frequency or duty cycle.

The simple formula to calculate oscillation frequency is 0.721 / (R_C * C). The complete formula in Figure 3 can also be used by replacing (R_B + r_ON) with (R_C + R_OL) and replacing (R_A + R_B) with (R_C + R_OH). R_OL and R_OH are the output resistances of the output V_OL and V_OH drivers respectively; the oscillation frequency will be 1 / [t_c(H) + t_c(L)].

Figure 4. 50 percent duty cycle schematic for CMOS timers

Three resistor, a-stable circuit for 25 percent to 60 percent duty cycle

Adding a ‘discharge pin’ resistor (R_D) will help reduce the output duty cycle because it increases the discharge resistance. R_D lowers the discharging voltage magnitude due to the voltage divider of R_A and R_D. The discharge resistance is R_B + (R_A||R_D) and the discharge voltage to supply voltage ratio is equal to 2/3 – R_D/(R_A+R_D). The discharge ramp needed is 1/3 of the supply voltage. So the discharge time is C*(R_B + R_A||R_D)*-ln(1-1/3/(2/3 – R_D/(R_A+R_D)). This simplifies to C*(R_B+R_A*R_D/(R_A+R_D))*ln((2*R_A-R_D)/(R_A-2*R_D)), which is still complicated. The charge time is the same as the standard setup, (R_A + R_B) * C* ln(2). This method works well for all types of timers.

Figure 5. Three resistors a-stable circuit

All three resistors (R_A, R_B and R_D) control the output frequency and the duty cycle of this circuit, as shown in Figure 5. First choose (R_A+R_B) total and C to get the desired output high time. Then choose one of the five R_B/R_A ratios [0, ½, 1, 2 and 10] plotted in Figure 6 and Figure 7. Finally, choose the R_D/R_A ratio needed to get the desired duty cycle.

Figure 6. Resistor ratios [0, ½, 1, 2 and 10] vs output duty cycle

Figure 7 shows the same data as Figure 6 except that the duty cycle range is limited from 25% to 60%. Duty cycles below 25% are omitted from the chart because the R_D/R_A ratio asymptotically approaches 0.5, as shown in figure 6. Duty cycles above 60% are omitted from this chart because the two resistor (RA and RB) circuit in Figure 1 is preferable for this output. Three common choices for scalars [R_A, R_B, R_D] that produce a 50% duty cycle are [9, 9, 2], [1, 2, 0.15] and [2, 1, 0.59]. The last two sets fall nicely into standard 1% resistor choices.

Figure 7. Resistor ratios [0, ½, 1, 2 and 10] vs output duty cycle for recommended duty cycle range

Less than 50 percent duty cycle using a diode

In the standard circuit, the timer’s high time charges the capacitor through R_A and R_B (in series) and the low time discharges the capacitor through just R_B. Therefore, the positive duty cycle can be between 50% and 100% exclusive. An additional steering diode can achieve lower duty cycle by bypassing R_B. However, the voltage loss due to the diode is temperature sensitive; therefore the output duty cycle and frequency will have a greater temperature coefficient. Figure 8 shows an example of a circuit that includes a steering diode.

Figure 8. Diode helps with output duty cycle below 50 percent

The output high time is set by R_A and the diode voltage drop (V_D), whereas the low time is set by R_B. The capacitor, C, scales low and high time equally. Just like the standard a-stable circuit, the low time is R_B * C * ln(2) and the high time is (R_A + R_B) * C* ln((2V_S-3V_D)/(V_S-3V_D)), where V_S is supply voltage and V_D is diode voltage. As supply voltage increases, the V_D term becomes less significant. As supply voltage decreases, the V_D term becomes more significant. For very low supply voltages, where V_S < 3 * V_D, the timer will never oscillate.

Less than 50 percent duty cycle or to make discharge a dedicated output using two diodes

To lessen the effect of a single diode variance on duty cycle or to make the discharge pin free, two diodes can be used at the expense additional frequency shift.  The discharge pin is not needed for timing, so it is available as an open drain output. Figure 9 shows an example of a circuit that includes independent high and low times using two steering diodes. Additional output load can have a small impact on output duty cycle and frequency.

Figure 9. Two diodes for second output and more accurate duty cycle than using one diode

The output high time is set by R_A , diode voltage drop (V_D1), and capacitor C. The low time is set by R_B, a similar diode voltage drop (V_D2), and capacitor (C).  The ambient difference in voltage drops (Vd1-Vd2) is approximately 60mV * log (RB/RA) and scales directly with absolute temperature. The capacitor, C, scales low and high time equally. The low time is R_B  * C* ln((2V_S-3V_D2)/(V_S-3V_D2)) and the high time is R_A  * C* ln((2V_S-3V_D1)/(V_S-3V_D1)), where V_S is supply voltage and V_D1 , V_D2 are diode voltages. Both equations ignore propagation delays. As supply voltage increases, the V_D term becomes less significant. As supply voltage decreases, the V_D term becomes more significant. For very low supply voltages, where V_S < 3 * V_D, the timer will never oscillate.

Downloadable TLC555 timer tool

 Texas Instruments provides a special timer tool to aid customers in their TLC555 timer design. The tool includes propagation delay (T_phl and T_plh) and discharge transistor on-resistance (Ron) to select components that will more closely match actual device performance. Default values for internal parameters reflect the TLC performance. For the other timers, set the “internal parameters” cells with values for the desired timer. This tool also supports adding a diode for low duty cycle output. 

The first a-stable pulse may be longer

 Often overlooked is that the initial pulse high time will be longer than the successive high pulses. The first high pulse will be 59% longer, ln(3)/ln(2). Usually, the timing capacitor voltage starts at 0V at power up or after a reset release (similar to a mono-stable circuit) whereas all remaining cycles will start at 1/3 supply voltage. Figure 10 is basic a-stable timer circuit used for waveforms in figure 11.

Figure 10. Power up schematic for simulation

Figure 11. Power up waveforms with standard R_A, R_B, and C setup.

The first high output will be the same width as the successive pulses if the control and timing capacitors are replaced with pairs of properly scaled capacitors (as seen in Figure12). The power up should be quick compared to the output high time, but not instantaneous (as seen in Figure 13). The two capacitor pairs create a capacitive voltage divider to bias the timing node to 1/3 V_DD and control the node to 2/3 V_DD at power up. The total timing capacitance is C = C₂ + C₃.

The original single capacitor helps isolate the timer comparators from noise that may be on the supply voltage pin. However, the capacitor pairs will couple supply noise to the timer comparators. 2/3 of V_DD noise is present on the control pin and the threshold comparator; while 1/3 of V_DD noise is present at the timing capacitor node and the trigger comparator. Note that the coupled noise can increase output pulse jitter.

Figure 12. Equal pulse width on power up schematic

Figure 13. Power up simulation for C₂ = C₄ = 2nF and C₁ = C₃ = 1nF

Swapping capacitors C₂ and C₃ in Figure 14 will force the output to start at low level. The first pulse will have the same period and duty cycle as the successive pulses in Figure 15. The power up should be quick compared to output low time, but not instantaneous. The capacitor pairs create capacitive voltage dividers to bias the timing node and the control pin node to 2/3 V_DD. The total timing capacitance is C = C₂ + C₃.

The two capacitor pairs will couple V_DD noise to the timer comparators. 2/3 of V_DD noise is present on the control pin, the threshold comparator and the timing capacitor node. 1/3 of the V_DD noise is present on the trigger comparator. The coupled noise can increase output pulse jitter.

Figure 14. Starts low with equal pulse width on power up schematic

Figure 15. Power up simulation for C₃ = C₄ = 2nF and C₁ = C₂ = 1nF

A more complicated setup shown in Figure 16 eliminates the V_DD noise coupling paths and supports reset pin toggling. At power up after a reset release, all output pulses will be the same.    

R₁ || R_B and R₃ generate a 1/3 V_DD reference voltage that passes through T₁, then to the timing capacitor, C, when the reset voltage is low. The reset input can be switched between 0V and V_DD to control the output stream. The reset V_IH, voltage input high, needs to be greater than 2/3 V_DD to allow the timing capacitor to reach the threshold input level.

For a delayed power up start, add a capacitor to ground on the ‘Reset pin’. In order for T1 to have enough headroom to turn on, the VDD must be 3V at minimum. D₂ and R₄ will pull-up ‘reset pin’ voltage more quickly once the output goes high. D₂ and R₄ are recommended when a reset capacitor is used; these components are not needed for a clocked reset input signal.

As usual, R_A, R_B, and C set the frequency and duty cycle of the output stream. The timing capacitor pre-charge time is selectable. Choose a pre-charge time,‘t’, that is less than the minimum reset low time. Set R₁ = t / (C * 2) where ‘C’ is the timing capacitance. Set R₂ = R₁ * 7 for a base current that ensures T₁ to saturate. If V_S > 8V, then use the base emitter protection diode, D₁. An accurate value for R₃ matters to get a 1/3 V_DD reference while under the load currents of R₁, R₂, and R_B. Calculate the factor ‘k’ = (V_DD – 1.8V) / V_DD, when D₁ is not used and ‘k’ = (V_DD – 3.6V) / V_DD when the D₁ diode is implemented. Then set R₃ = 2 / (1/R₁ + 1/R_B + k/R₂ ). T₁ could be replaced with a PMOS, which makes the ‘k’ factor zero. With a PMOS transistor, R₂ is optional and D₁ diode cannot be used.

Figure 16. Equal pulse after power up and reset release schematic.

Maximum Frequency

 The CMOS data sheets have a typical value for the ‘maximum frequency’ specification.

LMC555 Data sheet (revision M)

TLC555 Data sheet (revision I)

In both cases, the test conditions use R_A = 470Ω, R_B = 200Ω, C = 200pf. These values would produce 8.3MHz on an infinitely fast timer. Therefore the actual frequency reflects the maximum frequency of timer. Only the TLC timer has a minimum value. The bipolar timers do not have a maximum frequency specification. Guidance is given by a chart in the ‘Typical Characteristics’ section of the data sheet. 100 kHz is the highest frequency on the chart.

Figure 17. Output frequency vs. timing capacitance, applies to bipolar and CMOS timers

Minimum Frequency

 The minimum output frequency achievable is based on leakages in the timing capacitor, bias current and leakage in the timer’s ‘threshold pin’, ‘trigger pin’, ‘discharge pin’, and the PCB. The circuit board can be the biggest source of leakage if PCB condensation is possible. Leakage can also affect the accuracy of the frequency because leakage changes the effective capacitor charging and discharging currents.

The ‘threshold pin’ and ‘trigger pin’ in the bipolar timers will start to draw a bias current near both the threshold and trigger voltages as shown in the timing window lines of figure 18. Therefore, the bias current does not affect the charging current until the end of the capacitor voltage ramp. The charging resistor current needs to be greater than the threshold pin current. The CMOS timers do not have a bias current. 

Figure 18. Typical Bipolar ‘Threshold pin’ and ‘Trigger pin’ input bias current vs. pin voltage where V_s is 5V

 Since a large capacitance is required to achieve a very low frequency, a nano-timer can be used as an alternative for low-frequency applications. They only require a single resistor to set the time delay up to 2 hours and use little power. Another alternative is to pass the timer output to a logic multi-stage ripple counter listed in table 1 to get a final very low frequency output with a small timer capacitance value. CD4060B has its own oscillator. So the 555 timer may not be needed. 

Table 1. Ripple counter options to extended minimum frequency or reduce timing capacitance. Output is 50% duty cycle by default.

Accuracy

 In most cases, the tolerance of the timing capacitor will be the largest source of timing variation. Ceramic timing capacitors should be C0G/NP0 type for the best temperature coefficient, voltage modulation, and aging characteristics. The timing resistors’ tolerance and temperature coefficient is also important. The higher the application frequency, the more propagation delays will affect output frequency.

The initial a-stable timing accuracies are typically 1 percent for TLC, 1.5 percent for SE, 2.25 percent for LM, NA, NE, SA, and unspecified for LMC. The maximum frequency error is 3 percent for TLC and 17 percent for LMC, which is very pessimistic. All accuracies are based off 25°C values and temperature coefficients in the data sheets are only typical values. For full temperature range final accuracy better than 5 percent, the cost of improving the timing capacitor and resistor may be higher than the cost of a crystal oscillator and a logic multi-stage ripple counter.

LMC555, calculations [4.0/4.8-1 = -17% ; 5.6/4.8-1 = +17%] (datasheet revision M)

TLC555 (datasheet revision I)

LM (data sheet revision D)

NA/NE/SA/SE (data sheet revision I)

Control voltage

The ‘control’ pin voltage is internally generated and is normally 2/3 of the V_S. The control voltage can be forced to a different voltage. Both the frequency and duty cycle will change with control voltage. Adding a capacitor to the control pin helps reduce jitter when the supply voltage is noisy. On power up the control capacitor will lag the supply voltage based on the impedance of the control pin and the control capacitor. The data sheets do not specify control pin impedance; refer to table 2 for sample data ‘control pin’ impedance, temperature coefficients, and voltage coefficients at 25˚C and V_S = 5V

Table 2. Sample data ‘control pin’ impedance at 25˚C and V_S = 5V

Reset input

When the ‘reset pin’ voltage goes low, the output will go low as well. The ‘discharge pin’ will also go low and the timing capacitor will discharge through the resistor R_B. The output will remain low until ‘reset pin’ is returned to high. When reset goes high, the output will go high if the timing capacitor voltage is below 1/3 V_S. Otherwise, the output will remain low until the timing capacitor voltage drops to 1/3 V_S.