Fullscreen HTML5Med Res (??MB)HTML4

3-Way Fully Adjustable Stereo Active Crossover for Loudspeakers This Stereo 3-Way Adjustable Active Crossover is not only a fantastic tool for loudspeaker design and development but it can also be integrated into a 2-way or 3-way Active (powered) loudspeaker. The crossover points and levels for tweeter, midrange and woofer are fully adjustable with separate controls for each driver. By JOHN CLARKE 24  Silicon Chip siliconchip.com.au FEATURES: • • • • • • • • Stereo crossovers 3-bands (Bass, Mid and Tweeter) or 2-band use (low pass and tweeter) Optional use of the bass output as a subwoofer output in 2-band mode Adjustable crossover frequencies Individual level controls for each band Overall volume control Balance control Limiter for Bass output (optional) Of course, passive crossovers can be designed with steeper roll-offs, but these are more complex and expensive. Another drawback with passive crossover design is that loudspeakers are not simply resistive, even though their nominal impedance may be 4Ω or 8Ω, for example. Impedance varies with frequency so an 8-ohm loudspeaker may only have an impedance of 8Ω at one frequency. At other frequencies, the impedance can be lower or higher; maybe much higher than the nominal impedance. So why does the impedance value vary? Because all loudspeakers have inductance. Loudspeaker impedance also varies because of cone resonances and in the case of the woofer, due to the air loading on the speaker cone inside the box. These need to be compensated for if the crossover is to work correctly. (The lowest impedance value for a loudspeaker will typically be just above its cone resonant frequency and will be close to its DC resistance). This why you cannot take a passive crossover off the shelf and hope that it will work well with a random selection of drivers mounted in a given enclosure. Nor can you simply substitute a tweeter or woofer for the original drivers in a loudspeaker system with a passive crossover network – it is not likely to work well! Solving the problems M ost hi-fi loudspeaker systems have passive crossover networks to separate the audio signal into different bands, to suit the tweeters, midrange drivers and woofers. Passive crossovers comprise inductors, capacitors and resistors. This approach can be simple and economical for a 2-way loudspeaker (ie, with tweeter and woofer) but it can be much more complex and expensive for 3-way loudspeakers (ie, with a midrange driver added), especially if there are big disparities between the efficiencies of the different drivers and if quite steep crossover roll-off slopes are required. With active crossovers, it’s easier to produce steeper roll-off rates and the signal level can be optimised for each driver via its own amplifier. siliconchip.com.au In more detail, one of the major disadvantages of a passive crossover is that the changeover between the separate frequency bands is usually not very sharp. A typical crossover slope is only 6dB/octave or maybe 12dB/octave, in theory. In practice, as we shall see, the slope can be much less and that means there is a wide frequency range over which the two drivers will be both producing the same sound frequencies. That can mean that a woofer will be fed with higher frequencies than it ideally should (eg, above 1kHz) and the tweeter may be fed with lower frequencies (eg, below 1kHz). This means that both drivers are operating outside the regions where they produce the lowest distortion. By contrast, active crossovers can solve many of the above problems. Firstly, the frequency overlap between two loudspeaker drivers can be minimised by steep roll-off slopes. Secondly, the impedance of each driver does not affect the crossover frequency. Nor is there any interaction between the crossover components, as can be the case in passive crossover networks. Thirdly, the electrical damping of the driving amplifier is not reduced by the impedance of the components in a passive crossover. This means better damping of woofer cone motion, ie, lower distortion and less boominess. OK, so active crossovers do have advantages but most designs are not easily adjustable without changing lots of components. Our new design is fully adjustable September 2017  25     Fig.1: the stereo audio signal is split into three separate stereo signals covering different frequency  ranges, to suit the woofers, mid-range drivers and tweeters. For a two-way system, the third signal can optionally be used for subwoofer(s). for both crossover frequencies and driver signal levels – just use the control knobs! Low pass, high pass Before we go any further we should explain some terms which often confuse beginners: low-pass, high-pass and band-pass filters. Exactly as its name suggests, a lowpass filter is one that allows low frequencies to “pass” through it and it blocks the higher frequencies. Hence, a circuit to drive a subwoofer would be called a low-pass filter since it only delivers frequencies below 200Hz or thereabouts. Similarly, a high-pass filter is one that allows high frequencies to pass through it and it blocks low frequencies. The part of a crossover network which feeds a tweeter is said to be a high-pass filter, even though it may consist of only one capacitor. You would probably realise that as the frequency drops, the impedance of a given capacitor increases, hence blocking the higher frequencies. (Incidentally, the ultra-handy S ILICON C HIP Inductance/Capacitance Ready Reckoner Giant Wall Chart (see www.siliconchip.com. au/l/aaek or www.siliconchip.com. au/Shop/3/3302) demonstrates this perfectly – you nominate a capacitance value and as you move up the frequency scale, you can see that the impedance increases. If you’re designing filter circuits, this chart is a must!). If we cascade (ie, connect in series) a high-pass filter with a low-pass filter, the combination will pass a band of frequencies and we then refer to it as a “band-pass filter.” We use a bandpass filter for the midrange output in this active crossover circuit. Other points you need to know about high and low-pass filters are the so-called cut-off frequency and the filter slope roll-off. Typical filter slopes are specified in dB/octave where the dB (decibel) term is the attenuation. Typical slopes are -6dB/octave (quite gradual), -12dB/oc- Fig.2: eight active filters are used to produce the signals for each channel, along with four variable attenuators, a bass limiter. The stereo volume and balance controls operate on both channels.  26  Silicon Chip siliconchip.com.au   Fig.3(a): this is the configuratio of each second-order low-pass filter, which is known as a Sallen-Key type. Its expected frequency response is shown at right. Note that the variable resistances required are of the same value. tave, -18dB/octave and -24dB/octave (quite steep for a crossover network). The filter slope applies for frequencies after the cut-off frequency. The cut-off frequency is where the signal output is -3dB down on the normal level. For example, in a low-pass filter we might have a cut-off frequency of 1kHz (ie, -3dB point) and at slightly above that frequency, the slope will be -12dB/ octave. And for the filters described here, this means that the response at 2kHz (ie, one octave above) will be -12dB and at 4kHz it will be -24dB. Two or three filter bands? Fig.1(a) shows the three filter bands available with our new Active Crossover. While it may not be immediately apparent, this involves two crossover points and no fewer than four filters. Starting from the left-hand side, we have a low-pass filter for the bass frequencies and it “crosses over” to a high-pass filter for the midrange frequencies. Further up the audio spectrum, we have another low-pass filter which blocks out higher frequencies and then it “crosses over” to another high-pass filter which handles the frequencies fed to the tweeter. Note that when we shift the low crossover frequency, we are actually simultaneously changing the cut-off  frequencies of the respective low-pass and high-pass filters – they are ganged together. Similarly, when we shift the high crossover frequency, we simultaneously change the cut-off frequencies for the midrange low-pass and upper high-pass filters. Fig.1(a) shows the new Active Crossover used in a 3-way configuration, with bass (woofer), midrange driver and tweeters. But Fig.1(b) shows that it could be used in an alternative configuration as a 2-way system with a midrange/ woofer and a tweeter, together with an optional subwoofer. The circuitry remains the same but the way you connect is a little different. We will talk about that later. Block Diagram Fig.2 shows the block diagram for the 3-Way Adjustable Active Crossover. Only the left channel is shown; the right channel is identical. It actually comprises four low-pass and four high-pass filters. Hmm, we just mentioned that only four filters were needed to produce the three bands shown in Fig.1. Why are there now eight filters involved? Patience, now – all will be revealed! The left and right channel inputs are fed to a stereo volume control (VR1a  and VR1b) and the signal is then buffered with op amps IC1a & IC1b and their outputs connect to the balance control, VR2. After further buffering by op amps 1C2a & IC2b (for the right channel), the signal is passed to two adjustable high pass filters involving IC4 and IC5 (signal path in green) and also fed to two adjustable low pass filters involving IC3 (signal path in blue). The signal from the high-pass filters is fed to the tweeter level control and then to the tweeter output, CON2a. The signal from the low-pass filters is fed to a second pair of adjustable highpass filters involving IC7 & IC8 and to a second pair of adjustable low-pass filters involving IC6. The output from the second pair of high-pass filters is fed to the midrange level control and then to the midrange output, CON3a. The output from the second pair of low-pass filters is fed to the bass level control (signal path in red) and then goes via the bass limiter (can be switched in or out) to the woofer (or subwoofer) output, CON3b. Why do we need a bass limiter? Because we envision that in some applications, the bass output will need to be boosted substantially and that could lead to overload of the woofer or woofer driver amplifier on loud pas-  Fig.3(b) & (c): the Sallen-Key high-pass filter requires two different resistances, however, the circuit at right shows how we have reconfigured it for identical resistance values so that ganged pots can be used.   siliconchip.com.au September 2017  27 The equation for calculating the fc for the filter is shown (in Fig.3(a)) though this calculation only applies to a Butterworth filter. High-pass filter By swapping the resistors and capacitors in the circuit of Fig.3(a), we can obtain a high-pass filter, as shown in Fig.3(b). Once again this arranged to have a Butterworth response with a Q=0.7071 but instead of having capacitors with values of C and 2C, we have resistors of 2R, between the non-inverting input of the op amp and ground, and R at the output of the op amp. Both these resistive elements are adjustable using potentiometers and that presents a big problem since our Active Crossover uses an 8-gang potentiometer for each crossover output; each potentiometer element needs to have the same value, eg, 10kΩ. To solve that problem, we use an exFig.4: the simulated response of a single pair of Sallen-Key low-pass/high-pass tra op amp, as shown in Fig.3(c). The filters with a corner frequency of 1kHz (red) and the cascaded pairs of Sallensecond op amp is connected as a unity Key filters (red), known as a Linkwitz-Riley arrangement. The flat green line gain buffer and is driven from a voltage shows the overall response when the signals are acoustically summed. divider connected to the output of the sages (hint: see page 33!). filter which gives a roll-off slope of first op amp, to drive the bottom end The bass limiter will prevent this 12dB/octave. of the potentiometer (R). while having negligible effect on the The basic design is referred to as This resistor now has half the sigsignal at other times. a Sallen-Key filter (after R. P. Sallen nal current through it and so acts as and E. L. Key of MIT Lincoln Labora- though it has a value of 2R – which is Two-way configuration tory in 1955). what we want. As noted above, this Active CrossoThe graph to the right of the circuit So that shows the configuration of vers can also be built as a 2-way system shows the roll-off slope beyond the all the low-pass and high-pass filters with an optional subwoofer output. In cut-off frequency (fc). The passband in the circuit but it does not explain that case, you would have a tweeter region refers to the frequencies below why we using four of each. output (CON2a), the midrange/woofer fc where the signal level is mostly unThe reason is that the circuits of output (CON2b) and the subwoofer affected by the filter. Fig.3 are second-order filters and their output (CON3b). The circuitry for IC6, For this particular circuit, the filter filter slopes are equal to 12dB/octave IC7 & IC8 could then be omitted. has a Q of 0.7071 and has a Butter- which is not particularly steep – we So now let us explain why we need worth response. The Q value means want twice that: 24dB/octave. So we eight active filters in each channel that the frequency response below fc use identical cascaded low-pass and rather than four. remains as flat as possible rather than high-pass filters to get the desired reFig.3 a, & b show the basic circuits with any amplitude ripple or peaking. sult. for the low-pass and highWe simulated the filter filters used in our Active circuits using LTspice to Crossover. obtain the actual responses Let’s talk about the lowfor the filters. If you wish to do some calculations of responses for these pass filter first, as shown If you use LTspice or are filters, an excellent website is available. This calculates the filin Fig.3(a). This consists of following our series on this ter responses for the Sallen-Key configuration and shows plots a single op amp together in SILICON CHIP, you may and filter Q for values of R and C. with two identical (adwish to use the SPICE file. For the low pass filter C1 is the capacitor that needs to be justable) resistors R and This file (Active filter.asc) twice in value to C2. R2 is double the resistance of R1 in the two capacitors, C and 2C. will be available from the high pass filter. (2C is actually two identiSILICON CHIP website. For a cut-off of 1kHz (fc), use 22nF for C (44nF for twice the cal capacitors in parallel). Fig.4 shows the results value) and 5.11543kΩ for R (10.23086kΩ for twice the value). The op amp is connected for the low-pass filter when as a unity-gain buffer and the cut-off frequency is For the high pass filter see: siliconchip.com.au/l/aaei because it uses two RC net1kHz. The response for the works, it is a second-order single stage Butterworth For the low pass filter see: siliconchip.com.au/l/aaej siliconchip.com.au 28  Silicon Chip Calculating R & C siliconchip.com.au September 2017  29                                       Fig.5: the main portion of the Active Crossover circuit, built around 22 LM833 dual low-noise/low distortion op amps. The layout is similar to that of block diagram Fig.2, so you should be able to identify the corresponding sections. VR3-VR6 are four eight-ganged 10kΩ linear potentiometers which allows the corner frequency of each set of four active filters which makes up a crossover network to track. So only two adjustments need to be made to change the crossover point for either bass/midrange or midrange/tweeter. The bass limiter and power supply sections of the circuit are shown separately. 30  Silicon Chip siliconchip.com.au                                         siliconchip.com.au September 2017  31 filter is 3dB down at the cut-off frequency. At 10kHz (one decade away) the response is down by 40dB, as expected. That’s a 40dB per decade (or 12dB/octave) roll-off. When the two filters are cascaded, we get a response that is referred to as “Butterworth squared” (also called a Linkwitz-Riley) filter. The combined filter Q is 0.5; obtained by multiplying the Q (0.7071) of each Butterworth stage together. The cascaded filter response is 6dB down at fc and 80dB down at 10kHz. Putting it another way, the combined filter slope, beyond fc, is 24dB/ octave. Similar results for the low-pass filter are also shown in Fig.4; -3dB down at 1kHz for the single stage and 6dB down at 1kHz for the cascaded filters. At 100Hz (one decade away), response is 40dB down for the single stage filter and 80dB down for the cascaded filter. We use the Linkwitz-Riley filters because when both the low and high pass filters are summed, acoustically the response is flat. Using the Linkwitz-Riley filters means that there are no dips or peaks in the frequency response across the crossover frequency region. For more information on LinkwitzRiley filters, see siliconchip.com.au/l/ aaeh The left and right channels have separate frequency adjustments. Ideally, both left and right channels should be able to be adjusted together for the same crossover frequencies. However, we were not able to do this easily and we shall see why later. Main circuit The main circuit of the Active Crossover is shown in Fig.5 and again, this only shows the left channel. Just so you can recognise the various low-pass and high-pass filters, dual op amps IC4 and IC5 are the cascaded first and second high-pass filters while dual op amp IC3b and IC3a are the cascaded first and second low-pass filters. All op amps in the circuit are LM833s for very low noise and distortion. Similarly, dual op amps IC7 and IC8 are the cascaded third and fourth second high-pass filters while dual op amp IC6b and IC6a are the cascaded third and fourth low-pass filters. Also note that all the potentiometer elements for the filters of IC3, IC4 and IC5 are part of the same 8-ganged potentiometer, VR3. Similarly, all the potentiometer elements for the filters of IC6, IC7 and IC8 are part of the same 8-ganged potentiometer, VR4. However, that means that this Active Crossover is not able to simultaneously adjust the crossover frequencies in both channels; each channel must be done separately. If we wanted to do both channels simultaneously, we would need 16-element pots and that is simply not practical. However, the level adjustments for each channel output are made using dual ganged pots, so these are done simultaneously. Now let’s track the signal through the crossover circuitry. The input signal is applied to an RF suppression network comprising ferrite bead L1, a 100Ω stopper resistor and a 10pF capacitor. The signal is then coupled to the volume control VR1a via a 22µF non-polarised capacitor. The signal from the wiper of VR1is buffered by IC1a and its output is con-                        Fig.6: the bass limiter circuitry, which prevents bass drivers which are driven with significant levels of gain from being overloaded. It uses pairs of LEDs and LDRs to form a variable gain amplifier for each channel, similar to a compressor but with a much longer attack and decay times. 32  Silicon Chip siliconchip.com.au Coming soon: a 3-way active dipole loudspeaker One of the main reasons why we have produced this highly flexible 3-way active crossover is that we are developing a 3-way active dipole loudspeaker with some most unusual features. For a start, there is no enclosure. All three drivers are mounted on a simple baffle. How can that possibly work? Don’t you need some sort of enclosure in order to produce adequate bass response? Normally, the answer is a resounding “yes!” but we have taken a similar approach to speaker design in producing a dipole loudspeaker – it radiates equally from the front and rear of the baffle. Doesn’t that lead to bass cancellation? Yes it does but a dipole enclosure can work well in a small room provided there is considerable bass boost. That is just not possible with a passive crossover but our new 3-way active crossover makes it quite simple to achieve, because it allows large differences in the signal power applied to each driver. We hope to feature this most interesting loudspeaker system in a few months. Watch out for it! nected to one side of the balance balance control, VR2. The balance control has a limited range of action and it works as follows. When centred, there is an equal loss in signal level for both channels that amounts to -1.42dB. When the pot is rotated off centre, more signal is shunted to ground in one channel than in the other channel. When the balance pot is rotated fully in one direction, it causes a loss of 8.3dB in one channel and slight increase in the other. So there is an overall 8.9dB change in level between one channel and the other. Following the balance control, the signal is again buffered by IC2a and then fed to the first high-pass and first low-pass filters involving IC4 and IC3, respectively. So the signal progresses through the first and second high-pass filters of IC4 and IC5 and also to the first and second low-pass filters of IC3b and IC3a. Then the respective tweeter and midrange signals are fed to the respective level controls, involving VR7b and VR8b. These are Baxandall circuits which give a logarithmic response when using a linear potentiometer. This is highly desirable since we want to use linear dual ganged pots and these have far better matching and tracking between channels than logarithmic taper pots. Two op amps are involved for each level control. The tweeter control, VR7b, involves op amp IC15a, configured as buffer, and IC16a, an inverting op with a gain of 4.5. Hence the overall gain range of the circuit is from unity to 4.5 which is  more than adequate for this application. Another advantage of this Baxandall level control is that it reduces noise at the lower gain settings. Further filter stages The output of the second low-pass filter involving IC3a is also fed to the third and fourth high-pass filters involving op amps IC7 and IC8 and also to the third and fourth low-pass filters involving IC6b and IC6a. The output of the fourth high-pass filter IC8a is fed to the midrange level control VR9b involving op amps IC19a and IC20a. Finally, the output of the fourth low-pass filter IC6a is fed to the bass level control VR10a involving op amps IC21a and IC22a. However, the bass level control can also be fed to the bass limiter which can        Fig.7: the power supply section of the circuitry, which is on the same PCB as the rest. Power can come from either an AC plugpack or centre-tapped mains transformer. The transformer output is rectified, filtered and regulated to produce the ±15V supply rails for the op amps. siliconchip.com.au September 2017  33 Parts List – Three-Way Active Crossover 1 1 1 1 2 2 1 2 4 6 2 1 1 1 2 2 1 1 8 4 4 4 2 4 main PCB, coded 01108171, 284 x 77.5mm front panel PCB, coded 01108172, 296 x 43mm rear panel PCB, coded 01108173, 296x 43mm 16VAC 1A (or higher current) plugpack DPDT PCB mount push button switches (Altronics S1510) (S1,S2) knobs to suit push button switches S1 & S2 (Altronics H6651) two-way vertical stacked PCB-mount RCA socket (Altronics P0210) (CON1) four-way vertical stacked PCB-mount RCA sockets (Altronics P0211) (CON2,CON3) knobs to suit VR3-VR6 (Mouser 5164-1227-J) knobs to suit VR1,VR2,VR7-VR10) (Jaycar HK-7734) TO-220 heatsinks, 19 x 19 x 9.5mm (Jaycar HH-8502) PCB-mount 2.5mm DC power socket (Jaycar PS-0520, Altronics P0621A) (CON4) 2.5mm DC line plug (Altronics P-0635A, Jaycar PP-0511) 3-way PCB-mount screw terminals with 5.08mm spacing (CON5) 5mm ferrite suppression beads (L1,L2) ORP12 (or equivalent) LDRs (Jaycar RD-3480) 50mm length of 6mm diameter black heatshrink tubing set of black Acrylic case pieces (SC4403) 16mm long M3 tapped spacers 9mm long M3 tapped Nylon spacers M3 x 32mm machine screws M3 x 5mm black machine screws M3 x 6mm screws & nuts self-adhesive or screw-on rubber feet Semiconductors 25 LM833D SOIC (SMD) dual op amps (IC1-IC25) 1 7815 +15V three-terminal regulator (REG1) 1 7915 -15V three-terminal regulator (REG2) 2 1N4148 diodes (D1,D2) 2 1N5819 Schottky diode (D3,D4) 1 W04 1.2A bridge rectifier (BR1) 2 5mm 7500mcd green LEDs (Jaycar ZD-0172) (LED1,LED2) 1 3mm blue LED (LED3) Capacitors 2 470µF 25V PC electrolytic 1 100µF 16V PC electrolytic 10 22µF NP 50V PC electrolytic 12 10µF 35V (or greater) PC electrolytic 20 120nF 63V or 100V MKT polyester 25 100nF X7R 50V SMD (1206) ceramic 20 22nF 63V or 100V MKT polyester 11 100pF X7R 50V SMD (1206) ceramic 2 100pF 50V ceramic Resistors (0.25W, 1%, through-hole or 1206 SMD as specified) 2 100kΩ 7 100kΩ SMD 8 22kΩ 2 10kΩ 1 5.6kΩ 8 2.2kΩ 2 2.2kΩ SMD 2 1kΩ 2 620Ω 8 150Ω 2 100Ω 26 10kΩ SMD 38 1kΩ SMD Potentiometers and trimpots 1 10kΩ log dual 9mm potentiometer (Jaycar RP-8756) (VR1) 1 10kΩ linear single 9mm potentiometer (Jaycar RP-8510) (VR2) 4 10kΩ linear 8-gang 9mm potentiometers, Bourns PTD9081015FB103 (VR3-VR6) (Mouser) 4 10kΩ linear dual 9mm potentiometers (Jaycar RP-8706) (VR7-VR10) 1 5kΩ 25-turn top adjust 3296W style trimpot (VR11) 34  Silicon Chip be switched in or out using switch S2. Limiter circuit operation The Limiter circuit is shown in Fig.6 and it acts on the signals from both channels, left and right. In essence, the bass signal from each channel (left from IC22a; right from IC22b) is fed to a passive attenuator comprising a 10kΩ resistor, a 100kΩ resistor to ground and a paralleled light-dependent resistor (LDR). LDR1 is used for the left channel and LDR2 for the right channel. Normally, the LDR resistance will be very high and the reduction in signal level will be less than 1dB. Op amp IC23b buffers the signal from LDR1, while IC23a buffers the right-channel signal from LDR2. Each LDR is located next to a LED and both are encased in a light-proof housing (made of heatshrink tubing). So light from LED1 can reduce the resistance of LDR1 and LED2 does the same for LDR2. Both LEDs are driven with the same current so that the signal level in both channels is reduced by the same amount. The drive signals to LED1 & LED2 are derived by dual op amps IC24 and IC25. The bass signals from IC23a and IC23b connect to the inverting inputs of IC24a and IC24b via 1kΩ resistors which mix the signals from both channels. These amplifiers have a gain of 100 by virtue of their 1kΩ input and the 100kΩ feedback resistors. The amplifiers also have their noninverting inputs connected to separate voltage references formed using a resistive divider across the ±15V supply. The attenuator comprises a 10kΩ resistor from the +15V supply, two 2.2kΩ resistors and another 10kΩ resistor to the -15V supply. The centre point of the attenuator where the two 2.2kΩ resistors meet is connected to the ground (0V). A 5kΩ trimpot (VR11) connects across the two 2.2kΩ resistors and can be used to adjust the voltages at TP1 and TP2. With VR11 set for 5kΩ, the voltage at TP1 and TP2 will be +1.57V and -1.57V respectively. This voltage can be reduced down to 0V, with VR11at the opposite extreme. When the combined signal from IC23a and IC23b swings positive but less than the TP1 voltage, IC24b’s output will be high; ie, above 0V. When the combined signal from IC23a and IC23b swings negative but less negasiliconchip.com.au tive than TP2, IC24a’s output will be low; less than 0V. In effect, IC24b & IC24a operate together as a window comparator. The signal from IC24b is inverted by IC25b, change any negative-going signal to positive-going. Then the positive going signals from IC25b and IC24a are fed to diodes D1 and D2, respectively. So any positive-going signal from IC25b or IC24a will cause D1 or D2 to conduct and charge the 100µF capacitor via the 1kΩ resistor. IC25a monitors the signal across the 100µF capacitor and drives LED1 & LED2 (in series) and these LED control the resistance of LDR1 & LDR2 to limit the bass signals when the exceed the thresholds set by TP1 & TP2. The time constant for the 100µF capacitor to discharge via the 100kΩ resistor is ten seconds. This time-constant prevents the audio signal from being modulated by the limiter circuit. The associated 1kΩ resistor sets the attack time-constant to 100ms, so that limiting does not instantly occur with brief transients. Note that the maximum 1.57V threshold at TP1 and -1.57V threshold at TP2 will start signal limiting for a sine wave that’s 1.57V peak or 3.14V peak to peak. That is about 1.1V RMS. Power supply Fig.7 shows the power supply circuit. It can be powered using a centretapped 30V transformer or a 16VAC plugpack – either transformer feeds the bridge rectifier via switch S1. However, the bridge rectifier works differently, depending on which transformer is used. The 16VAC plugpack connects via CON4 with one side going to ground while the centre-tapped transformer connects to 3-pin CON5. The net result is only two diodes are involved when the power comes via CON4 and S1a and we have half-wave rectification for the positive and negative rails fed to the 3-terminal 15V regulators. When the power comes via CON5, the full bridge rectifier is involved. Either way, the rectified DC is filtered using 470µF capacitors. Next month . . . Have we whetted your appetite sufficiently with the description of the Three-Way Active Crossover? Next month, we’ll move on to the construction, setup and use of this project. MPPT REGULATOR + SOLAR PANELS PACKAGE Includes 1x 12-24V 40A 150V MPPT Solar Regulator + 4x FS272 72W Solar Panels. Charge 12/24V batteries at 30/15A: 280W!! $ IT118..... 249 FOR PICK-UP ONLY from WOY WOY (or maybe SYDNEY) LOOKING FOR A PCB? PCBs for most recent (>2010) SILICON CHIP projects are available from the SILICON CHIP PartShop – see the PartShop pages in this issue or log onto siliconchip.com.au/shop You’ll also find some of the hard-to-get components to build your SILICON CHIP project, back issues, software, panels, binders, books, DVDs and much more! So in the meantime, use the parts list opposite to start gathering the bits you’ll need (there are some that aren’t normally available from your local lolly shop!) and get the PCB from the SILICON CHIP online shop (they’re already available, priced at only $20.00 plus P&P) – and remember, if you’re a SILICON CHIP subscriber, you get 10% off all items from the shop (subscriptions and postage excepted). While you’re about it, why not order one of the giant L-C-R Wallcharts as well – you won’t believe how handy SC it will be in your workshop! 12V SOLAR PANELS AND REGULATORS Framed Polycrystalline 30W and 50W SOLAR PANELS. Also available is a 12/24V PWM 20A Regulator. 30W Solar Panel: IT119 .... $50 50W Solar Panel: IT120.... $80 20A PWM Regulator ........ $18 STEPPER MOTOR ARDUINO-ETC. EDUCATIONAL PACK 7W LED BARS IT117 $5 Ea. 2x small 5V 4-Wire 2-Phase 25mm Stepper Motors + 2x Driver modules + A 5V Universal regulated plugpack Pack of 6: IT117P.... $24 1/2M Long Bars... 36 high output Pure White LED's... Heat-sinked by the Aluminum Bar... covered by a diffuser strip... Around 700Lumens at 12V... 0.6A at 12V and 0.75A at 12.6V: CLEARANCE: PICK UP ONLY FROM THE CENTRAL COAST 54W SOLAR SKYLIGHT KIT Includes 3X Custom Made Oyster Lights $ (350mm Diam) and 1XFS-272 Solar Panel. 60! K401 ALL THIS FOR ONLY....$ 9!! K416 CLEARANCE: 72W SOLAR SKYLIGHT KIT Includes 8X 20W LEDs (45 X 45mm/33V) and 1 x FS-272 Solar Panel $ 50! PHONE/EMAIL/CALL FOR A FREIGHT QUOTE K415 P H O N E/S M S/E M A I L TO R E Q U E S T A CALLBACK 0428 600 036 branko<at>oatleyelectronics.com siliconchip.com.au September 2017  35