Fullscreen HTML5Med Res (??MB)HTML4

Phil Prosser’s compact and high-quality Microphone Preamplifier If you use microphones for stage, recording or testing, you will be familiar with the need for a preamp to get a usable signal. Many microphones also need ‘phantom power’. This small box runs from a plugpack and offers a flat frequency response, very low distortion, low noise and adjustable gain. Background image: https://unsplash.com/photos/ALM7RNZuDH8 T his small microphone preamp is ideal for use in the studio, workshop or on the stage. It allows you to boost the gain of your microphone to line level and delivers a balanced or single-ended signal. The main version of this Preamp fits into a small, standard-sized enclosure that is widely available, as shown in the photos. This diecast aluminium case makes it tough enough to survive the worst abuse. If you want to integrate this design into a larger project, we have a version of the board that omits the cutout for the XLR connectors and drops one of the switching regulators, making it easy to run it from existing ±15V DC rails. That would make sense if integrating it into a power amplifier, preamplifier, mixer or similar. When built as a standalone unit, it runs from 9V DC, as widely used on stage. Those plugpacks generally have 2.1mm plugs with a positive ring and negative tip. We have included reverse polarity protection, so no damage will occur if the wrong plugpack is used. We set this requirement as it is a fair bet that things will get mixed up on the stage. You don’t want to be fiddling with equipment while the crowd waits for the concert to start! Therefore, it should ‘just work’. Performance The performance of the Microphone Preamplifier depends on various factors. Having low noise is important; the noise level is significantly affected by the source impedance and gain setting. For a source impedance of 560W with 50dB gain and a 1V RMS output, the signal-to-noise ratio is 70dB. At the same output voltage but a gain of 20 times, the SNR is 85dB. Features & Specifications Operates from a 9-15V DC plugpack (9V DC is common for stage equipment) Fits in a compact 120 × 93.5 × 35mm diecast enclosure Adjustable gain from -15dB to +50dB Switchable 20dB attenuator for high-level sources Switchable 48V phantom power Drives in excess of 5V peak-to-peak (1.75V RMS, 13dBu) into a 600Ω load Balanced or single-ended output Frequency response: ±0.1dB, 12Hz to 20kHz (gain=26dB/20×) (see Fig.1) Signal-to-noise ratio (SNR), Zi = 560Ω, Vout = 1V RMS: 85dB (gain=26dB/20×), 70dB (gain=50dB/320×) » Total harmonic distortion (THD): <0.002% (see Fig.2) » Built-in power protection, including reverse polarity » Inputs and output protection against most abuse » » » » » » » » » 28 Silicon Chip Australia's electronics magazine The frequency response with a gain of 20 times (26dB) is within 0.1dB from 12Hz to 20kHz – see Fig.1. As shown in Fig.2, the distortion (THD+N) is entirely determined by noise. The underlying distortion is significantly lower, in the region of -95dB (0.0018%) to -105dB (0.0006%). There is some evidence of noise from the switch-mode regulator at the output, but it is 70-80dB down, depending on the gain setting. That is a low enough level that it is not a concern. Given that the distortion is so low, it’s the SNR that’s going to be the performance limit. 70dB is pretty much the worst you can expect as long as your input signal level is sufficient to achieve at least 1V RMS output at the maximum gain setting of around 50dB. As you reduce the gain to 26dB, it will improve to 85dB, and it should improve further at even lower gain settings, exceeding 90dB. That’s assuming your microphone/signal source is high enough in level to still provide a useful output with less gain. Some challenges This design is a little tricky because microphone phantom power needs to be 48V DC to be universal. That is a lot higher than 9V DC. To provide users with headroom of 10-15dB over 0dBu, we want to be able to deliver an output signal with peaks above ±8V. That is needed for people using the mic closer than expected and to deal with loud passages. Stage equipment must have headroom; the siliconchip.com.au sound engineer can deal with levels at the mixing desk. That means we need supply rails of 48V DC plus dual rails sufficient to get this ±8V from an op amp. We want this in a small box and for the circuit to be as tough as a cheap steak. If we start with 9V DC and drop 0.5V across a reverse polarity protection diode, then budget another 0.5V for the plugpack output drooping, we only have a poorly-regulated 8V supply to work with. We considered using switched capacitor inverters/doublers using 555s but found that gave marginal supply rail headroom. After some thought, we decided to take a more industrial strength approach, using two LM2577 boost regulators and a cunning trick to sneak in a negative rail. These regulators are more powerful than we need, but they are widely available and can handle 60V on their output, enough for the phantom power rail. The resulting power supply fills a significant proportion of the PCB, as we shall see in more detail later. While this solution is hefty, it is very tolerant of input supply variation; even if the output is loaded with a very low impedance, the rails will stay up. If you are wondering if this could be run from a 9V battery, the answer is not for any length of time. The current draw is far too high to expect a decent lifespan from the battery, and it will go flat exactly when you don’t want it to. Full load current draw is about 120mA, which will flatten a 9V battery in short order. Don’t think that all this talk about the power supply means we’ve forgotten that the preamp part must also have decent performance. We’re using the same hybrid transistor/op amp balanced microphone preamp found in the Loudspeaker Test Jig (June 2023; siliconchip.au/ Article/15821), developed by audio guru Douglas Self. It gives excellent performance with low distortion and noise, plus a wide range of possible gain settings. Fig.1: we had to make the vertical scale very small to see the variations in frequency response as it is so flat. Fig.2: any distortion produced by the circuit is well and truly buried in the noise. Thus, the SNR is the primary determinant of the performance at any given gain setting. Circuit description Fig.3 is the block diagram for the Preamp, while the main (analog) part of the circuit is shown in Fig.4. S1 switches phantom power for the microphone via header CON10. Noise is filtered out of the 48V DC supply by a 100W/220μF low-pass filter (LPF). We siliconchip.com.au Fig.3: the Mic Preamp block diagram shows the somewhat complicated power supply at the top, with the superficially simple attenuation and preamplification circuitry below. Australia's electronics magazine February 2024  29 have used 6.8kW resistors for the two bias resistors; these should be matched as close as possible. We selected two resistors that measured within 0.1% from our collection of 6.8kW 1% resistors. You could buy 10 resistors and choose the bestmatched pair. The 47μF/100nF parallel capacitor pairs block DC from the microphone signal as it’s fed into the attenuator. These prevent the full 48V phantom power from being applied to the attenuator when the microphone is unplugged, so they must be rated at a minimum of 63V. This Preamp has a 20dB pad at the front end. It can be switched in to avoid the Preamp clipping with higher-­ level input signals. The pad uses two 1.8kW resistors in series with the input signals and a 430W resistor connected between the terminals of RLY1. Fig.4: the main analog section of the Preamp circuit. It is based on two dual op amps and two transistors; the transistors lower the noise floor substantially. The second op amp drives the balanced and unbalanced outputs. Relay RLY1 switches in a resistive attenuator so it can handle higher level input signals. 30 Silicon Chip Australia's electronics magazine siliconchip.com.au When the attenuator is switched out, the relay shorts out the 1.8kW resistors, and the 430W resistor is out of the circuit. When switched in, the 430W resistor is connected between the downstream ends of the 1.8kW resistors, forming a voltage divider. These relatively low values minimise additive noise from the attenuator and keep the impedance driving the following preamplifier low. To calculate the attenuation of this stage (when activated), add a mental ground connection in the middle of the 430W resistor, splitting it into two 215W resistors. These resistances are in parallel with the 4.7kW resistors to ground, so the dividers are formed with resistances of 1.8kW plus the microphone source impedance and 205.6W. Assuming a low source impedance, the resulting attenuation is -19.8dB. Note that we need closely-matched values for the 1.8kW and 4.7kW parts to ensure good common mode rejection performance when the attenuator is switched in. We have used a relay for this job as our experience with switching small signals with miniature toggle switches wired to the board is not great. A telecom relay gives better long-term reliability and lower noise for a modest increase in cost. The 6.8V zener diode across the relay protects it in case someone runs the Preamplifier from a higher voltage than expected. The series resistor will get quite warm, but it should survive, provided this abuse is not continuous. Preamp gain We have provided a variable gain Two versions of this project allow it to fit into a small box (as shown) or a larger chassis with dual-rail power available. that allows you to set the level from a range of microphone types and situations. When VR1 is set to minimum resistance, the gain is 47.8dB, calculated as: G = 1 + 2.7kW ÷ (10kW || [22W ÷ 2]) G = 1 + 2.7kW ÷ 10.98W G = 247 (47.8dB) When VR1 is set to its maximum of 10kW, the gain is 5.1dB: G = 1 + 2.7kW ÷ (10kW || [(10kW + 22W) ÷ 2]) G = 1 + 2.7kW ÷ 3338W G = 1.8 (5.1dB) Using a reverse log taper potentiometer for VR1 results in the attenuation being ‘linear’ in dB terms as the potentiometer is rotated. Otherwise, most of the potentiometer’s range will result in relatively low gain, with the last fraction of the rotation ramping the gain over 20dB or so. So make sure the pot you choose has a ‘reverse log’ or ‘reverse audio’ (C) taper. The small signal diodes in the preamplifier (D4-D8) ensure the op amp inputs are not overdriven. We have included a buffer following the preamplifier that also produces an inverted Calculating the total current draw The phantom power supply needs to provide about 10mA to the LM317HV (REG1) and a maximum of 14.1mA into the 6.8kW resistors if they are shorted to ground. That is 24mA at 55.3V, which will require ~166mA (55.3V ÷ 8V × 24mA) at the input of the REG3. The dual rail power supplies must supply up to about 40mA to the NE5532 op amps and input circuit and about 10mA each for REG3 and REG4. That is a total of 100mA, given there are positive and negative rails, meaning a draw of up to about 225mA (18V ÷ 8V × 100mA) at the input to REG4 in the worst case. That means the Preamp could draw something in the region of 350mA, although that would only happen if it were driving a shorted load. Most 9V plugpacks can supply this, but most 9V batteries can’t. The most we saw in our tests was 150mA from 9V. Note that the worst case current is at startup, when the switch mode regulators are charging the 56V and ±18V supply filtering capacitors. We have included a power LED, powered from the negative rail. We chose this rail because if a user connects the Preamp to an 18-24V DC plugpack, the boost regulator for the positive rail will likely shut down, and the negative rail will not be generated. No damage should occur, but the user will be informed that it is not operating by the power LED being off. siliconchip.com.au Australia's electronics magazine output. This allows the output to be single-ended or drive a balanced line at a high level. We have also added small signal diodes to the positive and negative rails on the outputs (D14-D17) so that if someone inadvertently connects this to a piece of equipment with a large DC offset on its input, they will protect the NE5532 (IC1). We have incorporated 100W series resistors on the outputs to ensure the op amp remains stable even when driving difficult loads or long cables. Those will also help to limit the current flow in the case of a misconnection. You can use the positive buffered output at pin 2 if you only need a single-­ended output. Power supply The power supply portion of the circuit is shown in Fig.5. The overall design comprises two switch-mode pre-regulators that drive LM317/337 linear regulators. This generates very clean power rails, including the phantom power rail. The phantom power supply uses the LM2577 (IC3) in a textbook configuration. Its input is bypassed with a 220μF low-ESR capacitor and a 100nF capacitor. 220μF is quite low, but the maximum current we need to supply is less than 30mA. That is little more than idling for the LM2577. We have increased the compensation capacitor in series with the 2.7kW resistor at its pin 1 from a suggested value of 1μF to 10μF. That slows the startup of the boost regulator. Our small 500mA switchmode plugpack went into current limiting without February 2024  31 this; that would not be a problem with a larger plugpack (or a linear type). The output voltage is set by the resistors connected to the feedback pin (pin 2). With the 33kW/750W feedback divider and IC3’s internal 1.23V reference, the result is an output of 56.25V (1.23V × [33kW ÷ 750W + 1]). A 10W/10μF low-pass RC filter on the output reduces the remnants of the 52kHz switching frequency. The following LM317HV-based linear regulator drops the output close to the 48V required for phantom power while removing most of the remaining switch-mode noise. The 330W and 12kW feedback resistors set its output to 46.7V (1.25V × [12kW ÷ 330W + 1]). Switch-mode regulator IC4 produces the +18V rail (dropped to +14V by linear regulator REG3) and is set up similarly to REG3. It uses the recommended 1μF compensation capacitor rather than the higher 10μF value used for REG3 to reduce its startup current. A lower value inductor of 100μH is used due to the much lower boost ratio required, under 2:1. You must use toroidal inductors. Its output voltage is set by 33kW and 2.4kW resistors to about 18.4V (1.25V × [33kW ÷ 2.4kW + 1]). It also has a 10W/10μF low-pass RC filter on its output, and the following LM317based linear regulator has its output voltage set by 3.9kW and 390W resistors, resulting in about 13.75V (1.25V × [3.9kW ÷ 390W + 1]) for the positive op amp rail. Now to the cunning trick. Being a boost regulator, LM2577 (IC4) switches its pin 4 to ground to establish a current in L1. When pin 4 subsequently goes open-circuit, that current continues to flow and charges the output capacitor to our target of 18V DC. That is repeated at 52kHz by this device. Therefore, we have a node at pin 4 switching between about 18.7V and ground. Our trick is to generate the negative rail is piggybacking off this node using a 2.2W resistor, 47μF capacitor and ultrafast diode D9. When the output of IC4 reaches 18.7V, that capacitor is charged to around 18V via D9. When IC4 switches pin 4 to ground, the positive end of that capacitor is pulled to 0V, so the negative end goes to about -18V. That charges the following 47μF capacitor via diode D3, creating our negative rail. The negative rail is not directly regulated, but the positive rail regulation will ensure the negative rail is about right. LM337 linear regulator REG4 has its output set to -13.75V, so even if its input is a little lower in magnitude Fig.5: the power supply circuitry uses two switch-mode regulator ICs, one charge pump and three adjustable linear regulators to generate a 48V DC phantom power rail plus regulated ±14V rails for the op amps. Those are all derived from a single 9V DC input. 32 Silicon Chip Australia's electronics magazine siliconchip.com.au than that of REG3, the final regulated rails will still be close to ±14V. While the negative rail can only provide a modest current, we only need about 40mA total to power a few op amps. PCB layout INPUT INPUT PROTECTION PROTECTION AND A ND A TTENUATOR ATTENUATOR MICROPHONE MICROPHONE PPREAMPLIFIER REAMPLIFIER OUTPUT OUTPUT BUFFERS BUFFERS + + + + DC 48V DC 48V LINEAR LINEAR R EGULATOR REGULATOR + Australia's electronics magazine + +56V DC DC +56V BOOST BOOST REGULATOR REGULATOR + + + + ±18V ± 18V DC DC BOOST BOOST REGULATOR REGULATOR ±14V ±14V DC DC LINEAR LINEAR REGULATOR R EGULATOR + + Fig.6: how the various circuit sections have been arranged on the PCB. This configuration allows it to fit in a compact case while keeping the noisy switchmode ICs away from the sensitive analog preamplifier circuitry. + siliconchip.com.au + + + The Microphone Preamp is built on a double-sided PCB coded either VR1 XLR MIC OUTPUT SOCKET XLR MIC INPUT SOCKET + Construction The ‘box’ version of the PCB requires some more components due to the dual-rail generation circuitry. In our prototype, we used bobbin-style inductors, but we found that toroidal inductors provided such a great improvement in performance that we had to change them to the design presented. COIL We have laid the board out so that it is a neat, if tight, fit into a standard 120 × 93 × 35mm diecast aluminium enclosure. It is just large enough to accommodate the PCB, two XLR connectors and the switches, but small enough not to get in your way in use. The aluminium is tough enough to take some abuse without getting ratty or cracking. Due to the fairly packed board, it was important to put the switch-mode regulators at one end and the preamp circuitry at the other and use extensive ground planes to keep the noise down. The resulting board configuration is shown in Fig.6. To get it to fit, we had to lay the board out with cutouts for the integral pillars in the corners of the enclosure and a cutout into which the XLR connectors sit. That allowed us to use through-hole parts exclusively, so it’s straightforward for anyone to build. Suppose you are integrating this into a larger enclosure, such as an existing preamp. In that case, we have designed a separate ‘embedded’ version of the board without the LM2577 that generates the positive and negative rails (IC4). That means you can run it from external ±15V rails instead. At the same time, we filled in the cutout as it would serve no purpose in such an application. Everything else is basically identical, so you can use the same overlay diagrams regardless of which version you build. Just leave out the parts that don’t exist on the embedded version (in case that is not obvious!). When buying your board, make sure you choose the version that suits your needs. The only other difference in components is that the 150W resistor next to CON10 is increased to 330W and two of the 3.9kW resistors have been reduced to 3.0kW so that the LM317/337 regulators will not go into dropout with their inputs at ±15V rather than the ±18V generated by the switching regulator in the other design. February 2024  33 VR1 VR1 Mic In 34 Silicon Chip G ND − 15V +15V CON 1 UF4002 + 220mF 2 5V 220mF 63V 22pF 2.2kW 4148 D6 4148 D4 100n F 100n F D26 47mF + 10mF 3.0kW 10mF 10mF 33kW 750W + 1 00n F R E G3 D28 10 0n F 4148 4148 4.7kW D7 D8 R E G4 390W 47mF LM317 390 W 100nF + 10W LM337 100n F D27 3.0kW 2.7kW 100nF IC3 LM2577T D13 10 W + Q1 BC559 10kW 1 00 W D1 5 47mF 10 W 47mF 4.7kW 1nF 1nF 100nF D29 IC2 NE5532 + 10W 1 D23 47mF + 47 m F 4148 4148 4148 D17 4148 GND 47mF D16 CON4 10 0n F 47 k W 10W 1 00 W 10 W Mic Out R E G1 330W 63V 4.7kW 10k W ZD4 ZD3 ZD2 47kW ZD1 1nF 4 30 W CON3 Atten. LM317HV 3.0kW A K 6.8kW 100nF 6.8kW 100nF 22pF 2.2kW 4148 D6 4148 D4 ZD5 10 0 W CON10 3 30 W 10mF 63V 22pF D14 22kW Q2 B C 559 2.7kW 10 0n F 22kW 1 0m F D22 100nF IC3 LM2577T Fig.7: this version of the PCB suits the diecast metal case, with a cutout at the top for the XLR sockets to fit. The diodes, electrolytic capacitors and ICs are all polarity sensitive, so make sure they are orientated as shown here. 01110231 (full version) or 01110232 (embedded version) and measuring 85 × 110mm in either case. The two layouts are shown in Figs.7 & 8. The main difference is the omission of the dual rail generation circuitry in the ‘embedded’ version. Most other parts and locations remain the same. We will describe building the full PCB that fits in the small case. You simply skip the missing parts for the embedded version that operates from dual rails. The only added part is the three-pin header for power input CON1 rather than the barrel socket. Start by fitting all the resistors. The pairs of 6.8kW, 4.7kW and 1.8kW resistors in the input section at upper left, need some care. These parts should ideally be matched to better than 1%; we bought 10 of each and chose the two that measured the closest for each pair. That improves the common mode (noise) rejection. Now move on to the diodes. There are five different diode types, so don’t get them mixed up and ensure that the cathode stripes are orientated as Phantom Power 63V + 22k W 2.7kW UF4002 6 3V 2.7kW 47mF 10 W 220mF 25V 220mF D27 3.9kW 63V LED CON5 22pF IC1 NE5532 1.8kW 220mF 4.7kW 10m F UF4002 33kW 750W 1mF LM317 390W RLY1 5V 4.7kW 100nF D7 4.7kW 4148 100nF + 10mF 3.9kW 10mF 10mF 100nF IC4 LM2577T 100nF REG3 47mF 10m F D13 100nF 220mF 25V + D26 D28 10mF 47mF 100nF REG4 390W 4.7kW 1.8kW 100nF 4148 4.7kW D8 10W L2 330mH + LM337 UF4002 2.4kW 33kW 3.0kW A K 100W D15 D29 10W 10W D2 Q1 BC559 10kW 4.7kW 10kW 1nF 1nF 100nF L1 100mH 2.2W D3 U F 4002 + 1N5819 or 100nF 47mF 100W 10W + 47mF + 10W 1 2.7kW U F 4002 1N5819 or 220mF 63V 100nF 47mF 1N5819/UF4002 D9 10W 220mF 63V + 47mF 4148 4148 4148 D17 4148 GND 47mF D16 CON4 D23 CON1 D1 1N5819 ZD3 ZD 4 47kW REG1 330W 63V 9V DC IN 47kW 10W LM317HV Mic Out 100nF 10mF Phantom Power 63V D22 12kW 150W 6 3V ZD2 ZD 1 1nF 430W + 47mF D14 10mF + 100W CON10 + 63 V CON3 Atten. 4.7kW 220mF 22pF 63V 63V IC2 NE5532 + 47mF 12kW Mic In 6.8kW 100nF 6.8kW 100nF ZD5 10m F 1.8kW Q2 BC559 2.7kW 100nF 22kW + 47mF 22kW COIL COIL RLY1 5V 22kW 2.7kW + 10m F + 4.7kW 1.8kW 22pF IC1 NE5532 + 63V 4.7kW + 63V + + 47mF + + 47mF LED CON5 470mF L2 330mH 22W CON2 470mF + CON2 + XLR MIC INPUT SOCKET + XLR MIC OUTPUT SOCKET 22W VR1 VR1 1 Fig.8: the ‘embedded’ version of the PCB removes the split rail generators so it can run from ±15V DC rails (or similar) that might already be available within a mixer, preamplifier or power amplifier. shown in the overlay diagrams. Note that the 400mW zener diodes look similar to the 1N4148 small signal diodes, so be careful with those. While diodes D2, D3 & D9 can be either UF4002 or 1N5819 high-speed types, D13 must be a UF4002. Next, mount all the non-polarised capacitors, ie, the ceramic and plastic film types. Follow with the electrolytic capacitors, which are polarised. They all face the same way, with the positive (longer) lead to the right and the stripe on the can to the left. We have marked the 63V-rated capacitors on the PCB, although if you use the parts specified in the parts list, they will already have the correct ratings. Now install the power socket, twopin and three-pin polarised headers, the two toroidal inductors (which are not polarised) and the potentiometer. The orientations of the polarised headers are not critical, but if you use our suggested orientations, you’re less likely to make mistakes following our wiring instructions. Australia's electronics magazine We can now fit the two LM2577s and test the boost regulators. Depending on whether yours come with staggered or straight leads, you might need to bend the leads to fit the pads. Ensure that the regulators sit close to the PCB and do not hang off the edge. You can put a dab of neutral cure silicone under the inductors. Initial testing To test the switching part of the board, connect a 9V DC plugpack and check the voltages on either side of D1, the protection diode. There should be 9V on the anode and over 8.5V on the cathode. If not, check for shorts and things getting hot, and verify that your plugpack has negative on the tip and positive on the ring (the opposite of many that you’ll find). Check the voltage on either end of the 10W resistor immediately next to the 33kW resistor (it’s all by itself on the embedded version, to the left of that 33kW resistor). You should get readings at both ends of 55V ±5V. Do not touch this with your fingers as it siliconchip.com.au is a high enough voltage to bite. If that is not correct, check the parts in the lower-right corner, especially IC3, and verify the orientation of D13. For the dual rail voltage generator on the non-embedded version, measure the voltage on either end of two more 10W resistors in the power supply section. One is just to the left of D26, while the other is just above D29. These should be ±18.4V ±1.5V. Again, if these voltages are not correct, stop and work out why. The likely culprit is incorrect diode or capacitor orientation. If IC3 or IC4 is not working, put a scope probe on pin 4 of IC4. You should see a switching waveform at around 52kHz. If not, it might not be getting power. Now fit the LM317HV, LM317 and LM337 devices (REG1, REG3 & REG4). After that, check the voltage on CON10, the phantom power header. It should be 48V ±4V. Also check the voltage on pins 4 and 8 of the (still empty) IC1 and IC2 locations. You should measure +14V ±1V on pins 8 and -14V ±1V on pins 4. Again, if one of these is off, there must be a problem around the associated regulator, so check the input voltages, and the orientations of the regulators and associated protection diodes. With the power supply now fully operational, mount the relay (watch its orientation), the two BC559 transistors and the two NE5532 op amps, which can be soldered directly to the board or socketed (although using sockets could reduce its robustness). Double-check their orientation before soldering, as desoldering op amps or relays is hard. If you have to remove one, cut off all the legs and desolder them individually. Re-apply power and check that the relay works by shorting the pins of CON3; you should hear the relay click. If not, check that the relay is the right way around and that you have ZD5 orientated correctly. You can now plug in a microphone or oscillator, with a maximum input level of 100mV, to the CON2 input and check that it is amplifying the signal correctly and delivering correct output signals at the pins of CON4. If you don’t get an output, check that you have phantom power on if required. Place a shorting block across CON10 if necessary. There should be close to 48V on the CON10 pins and a siliconchip.com.au Parts List – Compact Microphone Preamplifier 10 double-sided PCB coded 01110231, 85 × 110mm 1 120 × 93.5 × 35mm diecast aluminium box [Altronics H0454, Jaycar HB5067, Mouser 546-29830PSLA] 10 9V DC 700mA+ plugpack with 2.1mm ID plug 10 100μH toroidal inductor (L1) [Altronics L6522] 1 330μH toroidal inductor (L2) [Altronics L6527] 1 9mm 10kW reverse log potentiometer (VR1) [Mouser 858-P091NFC25CR10K or 652-PTD9012015FC103] 1 knob to suit VR1 (D shaft), around 13mm in diameter 10 PCB-mounting 2.1mm inner diameter barrel socket (CON1) [Altronics P0620] 2 8-pin DIL IC sockets (optional; for IC1 & IC2) 2 3-pin polarised headers, 2.54mm pitch, with matching plugs and pins (CON2, CON4) 3 2-pin polarised headers, 2.54mm pitch, with matching plugs & pins (CON3, CON5, CON10) 1 3-pin female chassis-mount XLR socket (CON11) [Altronics P0850] 1 3-pin male chassis-mount XLR socket (CON12) [Altronics P0852] 2 SPDT chassis-mount mini toggle switches (S1, S2) [Altronics S1310] 1 5V DC coil DPDT PCB-mounting telecom relay (RLY1) [Altronics S4128B] 1 panel-mount green 3mm LED with bezel (LED1) [Altronics Z0240] 8 M3 × 16mm panhead machine screws 4 6mm-long M3-tapped Nylon spacers 10 M3 shakeproof washers 6 M3 hex nuts 4 stick-on rubber feet [Altronics H0940] 3 1m lengths of light-duty hookup wire (eg, white, red & black) 1 short length of 3mm diameter heatshrink tubing Semiconductors 2 NE5532 dual low-noise op amps, DIP-8 (IC1, IC2) 21 LM2577T integrated switch-mode regulators, TO-220-5 (IC3, IC4) 1 LM317HV or LM317 adjustable linear regulator, TO-220-3 (REG1) [Altronics Z0545] 1 LM317 adjustable linear regulator, TO-220-3 (REG3) 1 LM337 adjustable negative linear regulator, TO-220-3 (REG4) 2 BC559 low-noise PNP transistors, TO-92 (Q1, Q2) 5 6.8V 400mA axial zener diodes, DO-35 (ZD1-ZD5) [Altronics Z0320] 10 1N5819 40V 1A schottky diode, DO-41 (D1) 30 1N5819 40V 1A schottky or UF4002 100V 1A ultrafast diodes, DO-41 (D2, D3, D9) 1 UF4002 100V 1A ultrafast diode, DO-41 (D13) 8 1N4148 75V 200mA diodes, DO-35 (D4, D6-D8, D14-D17) 6 1N4004 400V 1A diodes, DO-41 (D22, D23, D26-D29) Capacitors 1 470μF 25V radial electrolytic; 5mm pitch, max. 10mm dia. & 21mm high [Altronics R5164] 42 220μF 63V radial electro; 5mm pitch, max. 10mm dia. & 21mm high [Altronics R5148] 21 220μF 25V radial electro; 3.5mm pitch, max. 8mm dia. [Altronics R5144] 10 47μF 63V radial electro; 2.5-3.5mm pitch, max. 8mm dia. & 21mm high [Altronics R5108] 87 10μF 63V low-ESR radial electrolytic [Altronics R4768] 10 1μF 50V/63V radial electrolytic [Altronics R4718] 1512 100nF 63V/100V MKT Microphone Preamp Kit (SC6784, $70): 1 10nF 63V/100V MKT includes the standard PCB plus all 3 1nF 63V/100V MKT onboard parts, switches and mounting 3 22pF 50V C0G/NP0 ceramic hardware. Case, XLR connectors, bezel low-ESR types are preferred but not required LED and wiring not included. Resistors 2 47kW 6 4.7kW 2 1.8kW 3 100W 21 33kW 20 3.9kW 1 750W 1 22W 3 22kW 13 3.0kW 1 430W 7 10W 1 12kW 10 2.4kW 2 390W 10 2.2W 2 10kW 43 2.7kW 12 330W 2 6.8kW 1 2.2kW 10 150W 🔹 🔹 🔹 For the embedded version, add: 1 double-sided PCB coded 01110232, 85 × 110mm 1 3-pin polarised header, 2.54mm pitch, with matching plugs and pins (CON1) num digit indicates how many to use for the embedded version Australia's electronics magazine February 2024  35 If you decide to build the version that suits a case (shown right), it is a neat and tight fit. Because only the pot shaft needs to go through the case, assembly is not as hard as it might look. The embedded version of the PCB is a bit simpler (shown left). ‘reasonable’ DC voltage at pins 2 and 3 of the input connector. This will vary depending on the microphone; expect it to be between about 5V and 43V. If you still have trouble, use an oscillator to drive the ‘hot’ input (middle pin of CON2) and: ● Check the input voltage with a scope. It should be set to 100mV. ● Check the voltages on the 1.8kW resistors immediately on either side of RLY1. Assuming the use of a single-­ ended oscillator, one of these should have your test voltage. Switch the attenuator in and out; you should see a 20dB (10 times) reduction in voltage level at one end. ● Check the base voltages of Q1 & Q2. They should be about 0V (a ‘touch’ above, to be precise!) ● Check the Q1 & Q2 emitter voltages; they should be about 0.6V. ● There should be about 10V across the two 10kW resistors right next to the XLR socket cutout, on either side of the 470μF capacitor, and about 4.7V across the 4.7kW resistors immediately to the right of D7 and below D8. Check the orientations of D7 and D8 if these voltages are not right. These voltages should be identical as they connect to the inverting and non-inverting inputs of the same op amp (IC2a). ● Check the voltage on pin 1 of IC2; it should be close to 0V with no signal applied to the Preamp. If it is pegged to one of the supply rails, look for something amiss in the feedback loop through IC2a, IC2, Q1 & Q2. 36 Silicon Chip If it’s working, check that the gain control provides about 48dB of range. You will need to drop the input voltage at high gain settings to avoid clipping. You should be able to achieve more than 8V RMS between pins 2 and 3 of the output connector into a 600W load. Case preparation The 120 × 93.5 × 35mm (119mm from some sources) diecast enclosure is available from a range of suppliers. All our measurements assume the use of 6mm standoffs for mounting the PCB, which provide clearance for the attenuation and phantom power switches and taller low-ESR capacitors. If you want to use different standoffs, verify that everything will fit, especially the 63V capacitors and switches. Standoffs taller than about 8mm are unlikely to work. Start by drilling and deburring the holes in the side walls of the enclosure, as shown in Fig.9; hold off on the mounting holes in the base. We used a stepped drill bit to make the XLR connector holes. These are a real boon for making larger holes. We bought several types of XLR connectors and found they were all similar Fig.9: the drilling details for the XLR sockets and holes for the potentiometer, LED, DC socket and switches. Leave the small XLR mounting holes until you have the sockets ready to install so you can position them accurately. Australia's electronics magazine siliconchip.com.au Fig.10: while you can expect the PCB mounting holes to be in these positions, you should use the PCB assembly to mark them exactly before drilling them to ensure everything will fit. Fig.11: by attaching the standoffs like this, we get a robust result while also allowing us to finagle it into the case. ◀ The PCB is designed to accommodate the XLR connectors and just fit inside the case. The board is a tight fit, but the parts are not squished together too much. ◀ but differed in the required cutout. You might need to fine-tune your metalwork for your connector. We also recommend that you hold off drilling the smaller fixing holes for the XLR connector until after you have made the main hole. Once the connector fits OK, mark and drill these holes so they are in the ideal locations. The two lower holes for the XLR connectors will need to be drilled and tapped for a 3mm thread (drill to 2.5mm first), as there is no room for nuts inside the case. An alternative is to use a long 3mm pop rivet, an approach we have tried and found to work well, especially if you get a hole slightly crooked. Once you have the side holes drilled, present the PCB to the case without the standoffs attached, and mark the locations of the mounting holes. They are shown in Fig.10 but you should use the PCB to mark them more accurately. Drill these to 3.5mm and deburr them. This method is easiest since getting those measurements perfect inside the box is not easy. Install the standoffs to the case by putting a 16mm M3 machine screw and M3 shakeproof washer through the panel from the outside, then screw the 6mm standoff onto the machine screw – see Fig.11. Do not fully tighten it, as you need to be able to jiggle the PCB onto the M3 screws. Once the PCB is in place, tighten the screws onto the standoffs. Pushing the PCB onto the standoffs will help you do that. We placed slotted holes at the connector end of the PCB so you can present the board to the case with the connector end tilted down, allowing the gain control pot shaft to go through the front panel. You can then jiggle the M3 screws through the slotted holes. Once the board is in place, use shakeproof washers and an M3 nut to secure it, as shown in the photos. Installing the XLR connectors The input connector is next to the input header, with the output XLR next to the gain control. Solder three differently-coloured 100mm wires to these and twist them together neatly. Trim these back to allow a neat installation, and crimp or solder pins to the pluggable headers. Refer to the wiring diagram, Fig.12, to connect the ground, hot and cold wires to pins 1-3, respectively. The bottom fixings for the XLR siliconchip.com.au Australia's electronics magazine February 2024  37 HEATSHRINK SLEEVES Switches and LED 10mF 220mF 25V IC4 LM2577T 2.7kW 100nF 220mF 63V 4148 4148 4148 D6 D4 100nF 4.7kW 4148 2.2kW + 47mF + LM317 390W D27 3.9kW 2.7kW 100nF IC3 Fig.12: how to wire it all up. The switches, connectors and LED all connect to the PCB via polarised headers, so you can wire each up one at a time and then plug it all together once the PCB is in the case. 100nF D26 10mF 3.9kW 10mF 10mF 220mF 25V 47mF 100nF REG3 D28 33kW 750W 1mF 100nF + UF4002 REG4 390W D7 D8 10W 100nF D13 + L2 330mH + LM337 IC 2 NE5532 4.7kW 2 2 pF UF 4002 2.4kW 33kW 3.0kW A K Q1 BC559 Q2 BC559 2.7kW 100W D15 1nF 10kW 4.7kW 10kW 1nF D29 10W 10W D2 L1 100mH 2.2W D3 U F4 0 0 2 + 1N5819 or 100nF 47m F 100W + 47mF 22kW 10W 1 100nF U F4 0 0 2 1N5819 or 220mF 63V 100nF 47mF 1N5819/UF4002 D9 10W 220mF 63V 2.7kW 47mF 4148 4148 4148 D17 4148 GND 47mF D16 CON4 D23 CON1 D1 1 N5 8 1 9 Mic Out REG1 330W 63V 9V DC IN ZD3 ZD4 47kW LM317HV 100nF 10mF Phantom Power 63V D22 12kW 63V 150W 100W CON10 + 47mF 22kW D14 10mF + HEATSHRINK SLEEVES 63V + 1S S1 CON3 Atten. 4.7kW 220mF 47kW 10W 1.8kW LED C O N5 2 2 pF 100nF 22kW 10 m F 100nF RLY1 5V 10W 430W ZD5 10m F ZD2 4.7kW 1.8kW 4.7kW IC1 NE5532 ZD1 1n F 6.8kW 6.8kW 100nF 63V + 100nF + 47mF 63V COIL 1S S2 + 47mF + BOTH SWITCHES TURNED BY 90° TO MAKE CONNECTIONS CLEARER 1 + K F 0n 2 2 pF 470mF A + LED1 INPUT 2 3 SOCKET 1 CON2 VR1 V R1 22W Mic In 1 XLR MIC OUTPUT PLUG (XCLO 2) RN M1IC OUTPUT 3 2 1 SOCKET + XLR MIC INPUT SOCKET (XCLO 1) RN M1IC socket are pretty close to the case base, so we simply drilled and tapped ours. Solder a 10nF capacitor between the case lug on one of the XLR connectors and the ground wire on pin 1. This will effectively ground the case for AC signals. The connections for the switches are made with light-duty hookup wire. Use twisted wire (any colour will do) and assemble to the two-pin pluggable headers, as shown in Fig.12. Similarly, use two pieces of twisted light-duty hookup wire for the LED. Apply heatshrink tubing over the solder connections to it. We used red for the anode and black for the cathode. These connect to pins 1 and 2 of the pluggable header, respectively. Now attach a knob for the gain control. Make it small, as it will be next to the output XLR connector. You should have tested the board already, so you will be set to go. We found that the lip of the lid hit the M3 nuts that secure the XLR connectors. To solve that, we used a file to notch the lip on the lid to clear the nuts, and the lid was then a perfect fit. You will find that the case is very full. The capacitors and TO-220 devices fit with a couple of millimetres of clearance to the lid. We think this is about as good packaging as we could have achieved. If you are using the ‘embedded’ version, we will leave it to your creativity on where and how you mount the Preamplifier. It is a relatively modest PCB, so it should fit in most places. We would supply the board with ±15V, but you could probably run it from up to ±30V without the regulators getting hot, as the current drain on the linear rails is quite low. You will need to check this detail in your application. We kept our labelling simple in line with the utilitarian intended use of this device (see Fig.13); you can be creative with this if you wish. Finally, stick some rubber feet on the bottom so it won’t damage the surfaces it’s on and won’t slide around too much. Using it Fig.13: print out and attach this lid panel artwork to the top of the box so you (or someone else) will remember what everything does. The Preamp should generally be run from a 9V DC plugpack. It will work fine from 12V DC. While it will not be damaged by a higher voltage, up to 24V DC, it likely won’t operate as the negative rail will not be generated. SC Australia's electronics magazine siliconchip.com.au 38 Silicon Chip