HQ Notch Filter Without Close Tolerance Components

A notch for a narrow frequency band of a few per cent or even less normally requires close-tolerance components. At least, that’s what we thought until we came across a special opamp IC from Maxim. In filters with steep slopes, the component tolerances will interact in the complex frequency response. This effect rules out the use of standard tolerance components if any useful result is to be achieved. The circuit shown here relocates the issue of the value-sensitive resistors that determine the filter response from ‘visible’ resistors to ready available integrated circuits which also make the PCB layout for the filter much simpler. The operational amplifiers we’ve in mind contain laser-trimmed resistors that maintain their nominal value within 1‰ or less. For the same accuracy, the effort that goes into matching individual precision resistors would be far more costly and time consuming. The desired notch (rejection) frequency is easily calculated for both R-C sections shown in Figure 1.

Figure 1. Special opamps incorporating laser-trimmed resistors.

Dividing the workload:

The circuit separates the amplitude and frequency domains using two frequency-determining R-C networks and two level-determining feedback networks of summing amplifier IC2, which suppresses the frequency component to be eliminated from the input signal by simple phase shifting. IC1 contains two operational amplifiers complete with a feedback network. The MAX4075 is available in no fewer than 54 different gain specifications ranging from 0.25 V/V to 100 V/V, or +1.25 V/V to 101 V/V when non-inverting. The suffix AD indicates that we are employing the inverting version here (G = –1). These ICs operate as all-pass filters producing a phase shift of exactly 180 degrees at the roll-off frequency f0. The integrated amplifier resistors can be trusted to introduce a gain variation of less than 0.1 %.

They are responsible for the signal level (at the notch frequency) which is added to the input signal by IC2 by a summing operation. However, they do not affect the notch frequency proper — that is the domain of the two external R-C sections which, in turn, do not affect the degree of signal suppression. In general, SMDs (surface mount devices) have smaller production tolerance than their leaded counter-parts. Because the two ICs in this circuit are only available in an 8-pin SOIC enclosure anyway, it seems logical to employ SMDs in the rest of the circuit as well. Preset P1 allows the filter to be adjusted for maximum rejection of the unwanted frequency component.

Figure 2. This deep notch is within reach using just 5%-tolerance resistors and 20%-tolerance capacitors.

R-C notch filter:

Using standard-tolerance resistors for R1 and R2 (i.e., 1%, 0806 style) and 10%-tolerance capacitors for C1 and C2 (X7R ceramic) an amount of rejection better than that shown in Figure 2 may be achieved. The notch frequency proper may be defined more accurately by the use of selected R-C sections. Pin 3 of IC2 receives a signal that’s been 90-degrees phase shifted twice at the notch frequency, while pin 1 is fed with the input signal. These two signals are added by way of the two on-chip resistors. IC2 is a differential precision operational amplifier containing precision resistor networks trimmed to an error not exceeding ±0.2‰. Here, it is configured as a modified summing amplifier with its inverting input, pin 2, left open.

For frequencies considerably lower than the resonance frequency f0 = 1 / (2 π R C) the capacitors present a high impedance, preventing the inverting voltage followers from phase-shifting the signal. At higher frequencies than f0, each inverting voltage follower shifts its input signal by 180 degrees, producing a total shift of 360 degrees which (electrically) equals 0 degrees. The phases of each all-pass filter behave like a simple R-C pole, hence shift the signal at the resonance frequency by 90 degrees each. The three precision amplifier ICs can handle signals up to 100 kHz at remarkably low distortion. The supply voltage may be anything between 2.7 V and 5.5V. Current consumption will be of the order of 250µA.

Source :www.ecircuitslab.com

Router UPS

It can be handy to have your phone and Internet router continue working for a while after  a power failure for example, if they provide  access to a security system. This requires a  backup power supply for the router. The version described here consists of a 12-V lead-acid battery and a voltage converter capable  of supplying an output voltage in the range of  15 to 30 V. It has built-in protection to prevent  excessive battery discharge.  This DIY uninterruptible power supply (UPS)  operates in standby mode as long as the  mains voltage is present. 

Circuit diagram :

Router UPS Circuit Diagram

The UPS consists of four parts: a backup  detector circuit that monitors the supply  voltage from the AC mains adapter, a battery  circuit that monitors the battery voltage to  prevent it from dropping below 11.8 V, a FET  switch between the battery and the voltage  converter, and a voltage doubler (inside the  dashed outline). To understand how it works, first consider  the situation with a router supply voltage  above 20 V, for which the voltage doubler is  not required. In this case the outputs of com-parators IC1a and IC1b (pins 1 and 7) are connected directly to the gate of the FET (G1 is  connected to G2). 

Under normal conditions the router, which is  connected to K3, is powered from the voltage  on connector K1. In this situation the voltage  on pin 2 of comparator IC1a is higher than 5.6  V. The output on pin 1 is therefore low, and  the FET is switched off. If the external volt-age on K1 drops out, the voltage on pin 2 of  IC1 drops and the output on pin 1 goes high,  switching on the FET. In this state the battery and the voltage converter supply power  to the router. The battery will gradually discharge, and to prevent the battery voltage  from dropping below 11.8 V the output of  the second comparator (on pin 7) goes low  when the voltage reaches this threshold level,  switching off the FET. The battery voltage  may rise quickly after the FET is switched off, so capacitor C3 is included to ensure that this  does not cause the FET be switched on again. 

Switch S1 allows the UPS to start up without an external supply voltage on K1, and capacitor C4 enables the comparators to continue operating in the event of a brief dropout of  the two supply voltages on K1 and K2. The emergency stop switch S2 and fuse F1 are included for safety reasons. The voltage converter has a high inrush current, so F1 must be generously dimensioned. 

If the router supply voltage is below 19 V, the  comparator output level in the high state is  too low to achieve a gate–source voltage of  4.5 to 5 V, since the source voltage is always  the same as the battery voltage under continuous charging, which is 13.8 V. This means  that the gate voltage must be at least 18.3 to  18.8 V, which is difficult or impossible with a  router supply voltage under 19 V. This can be  remedied by including the voltage doubler,  which is built around the well-known 555  timer IC (CMOS version). The frequency of the  oscillator (IC2) is approximately 40 kHz. Components C6, D5 and D6 add the AC voltage to  the switched supply voltage delivered by T2,  which is driven by the comparators in parallel  with the timer reset. An 18-V Zener diode (C7)  protects the FET gate–source junction against  overvoltage. 

Be careful to select a 555 with a maximum  rated supply voltage sufficient for this application; they are available in 16-V and 18-V  versions. The voltage converter of this UPS is a note-book power converter designed for in-car  use, with an input voltage of 12 V, selectable output voltage, and a minimum current  capacity of 0.5 A. Most voltage converters can  handle this easily. The battery must be connected to a good charger capable of maintaining a lead-acid battery in good condition  under prolonged no-load operation. Various  designs for this have been described in Elektor in the past. 

Adjust P1 for a voltage of approximately 7 V. With a lab power supply connected in place of the battery, adjust P2 for a threshold volt-age of 11.8 V.

www.ecircuitslab.com

Baud Rate Generator

In this article, an RC oscillator is used as a baud rate generator. If you can calibrate the frequency of such a circuit sufficiently accurately (within a few percent) using a frequency meter, it will work very well. However, it may well drift a bit after some time, and then…. Consequently, here we present a small crystal-controlled oscillator. If you start with a crystal frequency of 2.45765 MHz and divide it by multiples of 2, you can very nicely obtain the well-known baud rates of 9600, 4800, 2400, 600, 300, 150 and 75. If you look closely at this series, you will see that 1200 baud is missing, since divider in the 4060 has no Q10 output!

If you do not need 1200 baud, this is not a problem. However, seeing that 1200 baud is used in practice more often than 600 baud, we have put a divide-by-two stage in the circuit after the 4060, in the form of a 74HC74 flip-flop. This yields a similar series of baud rates, in which 600 baud is missing. The trimmer is for the calibration purists; a 33 pF capacitor will usually provide sufficient accuracy. The current consumption of this circuit is very low (around 1mA), thanks to the use of CMOS components. 

www.ecircuitslab.com

Short Wave Superregenerative Receiver

Superregenerative receivers are characterised by their high sensitivity. The purpose of this experiment is to deter-mine whether they are also suitable for short-wave radio. Superregenerative receivers are relatively easy to build. You start by building a RF oscillator for the desired frequency. The only difference between a superregenerative receiver and an oscillator is in the base circuit. Instead of using a voltage divider, here we use a single, relatively high-resistance base resistor (100 kΩ to 1MΩ).

Superregenerative oscillation occurs when the amplitude of the oscillation is sufficient to cause a strong negative charge to be applied repeatedly to the base. If the regeneration frequency is audible, adjust the values of the resistors and capacitors until it lies somewhere above 20 kHz. The optimum setting is when you hear a strong hissing sound. The subsequent audio amplifier should have a low upper cutoff frequency to strongly attenuate the regeneration signal at its output while allowing signals in the audio band to pass through. This experimental circuit uses two transistors. A Walkman headphone with two 32-Ω earphones forms a suitable output device. 

Circuit diagram :

Short-Wave Superregenerative Receiver Circuit Diagram

The component values shown in the schematic diagram have proven to be suitable for the 10–20 MHz region. The coil consists of 27 turns wound on an AA battery serving as a winding form. The circuit produces a strong hissing sound, which diminishes when a station is received. The radio is so sensitive that it does not require any antenna to be connected. The tuned circuit by itself is enough to receive a large number of European stations. The circuit is usable with a supply voltage of 3 V or more, although the audio volume is greater at 9 V. 

One of the major advantages of a superregenerative receiver is that weak and strong stations generate the same audio level, with the only difference being in the signal to noise ratio. That makes a volume control entirely unnecessary. However, there is also a specific drawback in the short-wave bands: interference occurs fairly often if there is an adjacent station separated from the desired station by some-thing close to the regeneration frequency. The sound quality is often worse than with a simple regenerative receiver. However, this is offset by the absence of the need for manual feedback adjustment, which can be difficult. 

Author :Burkhard Kainka  - Copyright : Elektor

Low distortion Class-B circuitry 6V Battery Supply

Some lovers of High Fidelity headphone listening prefer the use of battery powered headphone amplifiers, not only for portable units but also for home "table" applications. This design is intended to fulfil their needs and its topology is derived from the Portable Headphone Amplifier featuring an NPN/PNP compound pair emitter follower output stage. An improved output driving capability is gained by making this a push-pull Class-B arrangement. Output power can reach 100mW RMS into a 16 Ohm load at 6V supply with low standing and mean current consumption, allowing long battery duration. The single voltage gain stage allows the easy implementation of a shunt-feedback circuitry giving excellent frequency stability.

Circuit diagram :

Battery-powered Headphone Amplifier Circuit diagram

Notes:

• For a Stereo version of this circuit, all parts must be doubled except P1, SW1, J2 and B1.
• Before setting quiescent current rotate the volume control P1 to the minimum, Trimmer R6 to maximum resistance and Trimmer R3 to about the middle of its travel.
• Connect a suitable headphone set or, better, a 33 Ohm 1/2W resistor to the amplifier output.
• Switch on the supply and measure the battery voltage with a Multimeter set to about 10Vdc fsd.
• Connect the Multimeter across the positive end of C4 and the negative ground.
• Rotate R3 in order to read on the Multimeter display exactly half of the battery voltage previously measured.
• Switch off the supply, disconnect the Multimeter and reconnect it, set to measure about 10mA fsd, in series to the positive supply of the amplifier.
• Switch on the supply and rotate R6 slowly until a reading of about 3mA is displayed.
• Check again the voltage at the positive end of C4 and readjust R3 if necessary.
• Wait about 15 minutes, watch if the current is varying and readjust if necessary.
• Those lucky enough to reach an oscilloscope and a 1KHz sine wave generator, can drive the amplifier to the maximum output power and adjust R3 in order to obtain a symmetrical clipping of the sine wave displayed.

Technical data:

Output power (1KHz sinewave):
    16 Ohm: 100mW RMS
    32 Ohm: 60mW RMS
    64 Ohm: 35mW RMS
    100 Ohm: 22.5mW RMS
    300 Ohm: 8.5mW RMS
Sensitivity:
    160mV input for 1V RMS output into 32 Ohm load (31mW)
    200mV input for 1.27V RMS output into 32 Ohm load (50mW)
Frequency response @ 1V RMS:
    flat from 45Hz to 20KHz, -1dB @ 35Hz, -2dB @ 24Hz
Total harmonic distortion into 16 Ohm load @ 1KHz:
    1V RMS (62mW) 0.015% 1.27V RMS (onset of clipping, 100mW) 0.04%
Total harmonic distortion into 16 Ohm load @ 10KHz:
    1V RMS (62mW) 0.05% 1.27V RMS (onset of clipping, 100mW) 0.1%
Unconditionally stable on capacitive loads

Source : red circuits

Comparator Based Crystal Oscillator

Although a simple crystal oscillator may be built from one comparator of an LT1720/LT1721, this will suffer from a number of inherent shortcomings and design problems. Although the LT1720/LT1721 will give the correct logic output when one input is outside the common mode range, additional delays may occur when it is so operated, opening the possibility of spurious operating modes. Therefore, the DC bias voltages at the inputs have to be set near the center of the LT1720/LT1721’s common mode range and a resistor is required to attenuate the feedback to the non-inverting input. Unfortunately, although the output duty cycle for this circuit is roughly 50%, it is affected by resistor tolerances and, to a lesser extent, by comparator offsets and timings.

 

If a 50% duty cycle is required, the circuit shown here creates a pair of complementary outputs with a forced 50% duty cycle. Crystals are narrow-band elements, so the feedback to the non-inverting input is a filtered analogue version of the square-wave output. The crystal’s path provides resonant positive feedback and stable oscillation occurs. Changing the non-inverting reference level can vary the duty cycle. The 2k-680Ω resistor pair sets a bias point at the comparator + (Comparator IC1a) and – (Comparator IC1b) input. At the complementary input of each comparator, the 2k-1.8k-0.1µF path sets up an appropriate DC average level based on the output.

 

IC1b creates a complementary output to IC1a by comparing the same two nodes with the opposite input. IC2 compares band-limited versions of the outputs and biases IC1a’s negative input. IC1a’s only degree of freedom to respond is variation of pulse width; hence the outputs are forced to 50% duty cycle. The circuit operates from 2.7V to 6V. When ‘scoping the oscillator output signal, a slight dependence on comparator loading, will be noted, so equal and resistive loading should be used in critical applications. The circuit works well because of the two matched delays and rail-to-rail outputs of the LT1720.

 

 

www.ecircuitslab.com

Pump Controller For Solar Hot Water System

This circuit optimises the operation of a solar hot water system. When the water in the solar collector is hotter than the storage tank, the pump runs. The circuit comprises two LM335Z temperature sensors, a comparator and Mosfet. Sensor 1 connects to the solar collector panel while Sensor 2 connects to the hot water panel. Each sensor includes a trimpot to allow adjustment of the output level. In practice, VR1 and VR2 are adjusted so that both Sensor 1 and Sensor 2 have the same output voltage when they are at the same temperature. The Sensor outputs are monitored using comparator IC1.

When Sensor 1 produces a higher voltage than Sensor 2, which means that sensor 1 is at a higher temperature, pin 1 of IC1 goes high and drives the gate of Mosfet Q1. This in turn drives the pump motor. IC1 includes hysteresis so that the output does not oscillate when both sensors are producing a similar voltage. Hysteresis comprises the 1MO feedback resistor between output pin 1 and non-inverting input pin 3 and the input 1kO resistor. This provides a nominal 12mV hysteresis so that voltage at Sensor 1 or Sensor 2 must differ by 12mV for changes in the comparator output to occur.

Circuit diagram:

Pump Controller For Solar Hot Water System

Since the outputs of Sensor 1 and Sensor 2 change by about 10mV/°C, we could say that there is a degree of hysteresis in the comparator. Note that IC1 is a dual comparator with the second unit unused. Its inputs are tied to ground and pin 2 of IC1 respectively. This sets the pin 7 output high. Since the output is an open collector, it will be at a high impedance. Mosfet Q1 is rated at 60A and 60V and is suitable for driving inductive loads due to its avalanche suppression capability. This clamps any inductively induced voltages exceeding the voltage rating of the Mosfet.

The sensors are adjusted initially with both measuring the same temperature. This can be done at room temperature; adjust the trimpots so that the voltage between ground and the positive terminal reads the same for both sensors. If you wish, the sensors can be set to 10mV/°C change with the output referred to the Kelvin scale which is 273K at 0°C. So at 25°C, the sensor output should be set to (273 + 25 = 298) x 10mV or 2.98V.

Note:

The sensors will produce incorrect outputs if their leads are exposed to moisture and they should be protected with some neutral cure silicone sealant. The sensors can be mounted by clamping them directly to the outside surface of the solar collector and on an uninsulated section of the storage tank. The thermostat housing is usually a good position on the storage tank.

 http://www.ecircuitslab.com

Railway Points Sequencer

Dedicated model rail enthusiasts using sophisticated train and points controllers often have the problem that as their layouts get bigger and more complex, the transformer supplying power to the points does not have enough current to switch several points at the same time. The actuators in the points are designed for ac operation so it doesn’t help by rectifying the supply and adding reservoir capacitors, the coils can overheat and burn out if they get jammed during their travel (ac operation actually helps to overcome friction in the mechanism). The circuit shown here solves this problem by using a sequencer to ensure than only one points actuator can be active at any point in time. During operation the controller will switch all the points on one line at the same time as usual, but the other connection to each coil is connected to the sequencer unit. This circuit will only allow current to flow through one coil at a time. 

Circuit diagram :

Railway Points Sequencer Circuit Diagram

The sequencer circuit consists of a 555 timer configured as an astable multivibrator clocking a 4017 Johnson counter where the ten outputs are used to switch ten triacs in sequence, enough for ten sets of points. P1 alters the oscillator frequency of the 555 timer and can be adjusted so that each time interval of the sequencer is long enough to allow the points to switch. 

The switching time varies depending on the type of points but is typically between 1 s and 1.5 s. Any points that jam during switching give out a characteristic humming noise in time to the switching frequency so it makes them easier to find. The eleventh output of the 4017 can be connected to an LED (together with a series resistor). This will flash to give a visual indication of the sequencers operation. Power for the circuit is provided by 15 V ac from the points transformer. The B80C1500 bridge rectifier (80 Vpiv, 1.5 A) and regulator IC1 produce a stabilised 12 V for the circuit. Current consumption is only a few milliamps.

http://www.ecircuitslab.com/2011/07/railway-points-sequencer.html

Very Low Power 32kHz Oscillator

The 32-kHz low-power clock oscillator offers numerous advantages over conventional oscillator circuits based on a CMOS inverter. Such inverter circuits present problems, for example, supply currents fluctuate widely over a 3V to 6V supply range, while current consumption below 250 µA is difficult to attain. Also, operation can be unreliable with wide variations in the supply voltage and the inverter’s input characteristics are subject to wide tolerances and differences among manufacturers. The circuit shown here solves the above problems. Drawing just 13 µA from a 3V supply, it consists of a one-transistor amplifier/oscillator (T1) and a low-power comparator/reference device (IC1).

Circuit diagram:

Very Low Power 32kHz Oscillator Circuit Diagram

The base of T1 is biased at 1.25 V using R5/R4 and the reference in IC1. T1 may be any small-signal transistor with a decent beta of 100 or so at 5 µA (defined here by R3, fixing the collector voltage at about 1 V below Vcc). The amplifier’s nominal gain is approximately 2 V/V. The quartz crystal combined with load capacitors C1 and C3 forms a feedback path around T1, whose 180 degrees of phase shift causes the oscillation. The bias voltage of 1.25 V for the comparator inside the MAX931 is defined by the reference via R2. The comparator’s input swing is thus accurately centred around the reference voltage.

Operating at 3 V and 32 kHz, IC1 draws just 7 µA. The comparator output can source and sink 40 mA and 5 mA respectively, which is ample for most low-power loads. However, the moderate rise/fall times of 500 ns and 100 ns respectively can cause standard, high-speed CMOS logic to draw higher than usual switching currents. The optional 74HC14 Schmitt trigger shown at the circuit output can handle the comparator’s rise/fall times with only a small penalty in supply current.

Author: http://www.ecircuitslab.com/2011/06/very-low-power-32khz-oscillator.html

Snail Mail Detector

Since his letter-box is outdoors  and quite some way from the  house, the author was looking  for a simple means of knowing if  the postman had been without  having to go outside (contrary  to popular belief, the weather  isn’t always fine in the South of  France). Circuits for this kind of  ‘remote detection’ come up regularly, but always involve running cables between the letter- box and the detection circuit in  the house. Seeking to avoid running any extra cables, the author  had the idea of using the existing cables going to the doorbell,  conveniently located adjacent to  his letter-box.

 

The letter-box has two doors:  one  on  the  street  side  for  the  postman, and one on the gar-den side for collecting the post.  A  micro switch  is  fitted  to  the  street-side door, to light an indicator in the house showing that  the postman has been. A second  micro switch is fitted to the door  on the garden side, to turn off  the indicator once the post has  been collected. The only difficulty then remains to connect  these detectors to a remote circuit in the house that remembers  whether  the  postman’s  been or not.

 

The idea was to use the alternating half-cycles of the AC signal  on the cable going to the door-bell  to  transmit  the  information, according to the following logic:

• Both  half-cycles  present: no change in the status of the mail detector.
• An interruption (even brief) of one half-cycle: indicator lights permanently.
• An interruption (even brief) of the other half-cycle: the indicator goes out.

Note that the signal is tapped off  across the doorbell coil via R6  and the pair of diodes connected  in inverse-parallel (to limit the  signal,  par ticularly  when  the  bell is rung). The signal is then  filtered by R2/C1, before being  used by IC1, which is wired as a  comparator with hysteresis. The  trigger threshold is adjusted by  P1, using a pair of inverse parallel diodes as a voltage reference  (positive or negative according  to the output state):

 

For the detection to work, there  has to be continuity in the bell-push circuit this is generally  ensured by the little lamp illuminating the bell-push. Resistor R1  is added just in case the lamp is  blown or not present. To keep things simple, the circuit is powered directly from the  doorbell transformer itself (230 V  / 8 V). The author managed to fit  the little circuit within the door-bell unit, with the LED poking  through a hole in the casing so  it is readily visible in the hall of  his house. 

Author : Philippe Temporelli (France) – Copyright : elektor electronics