Fullscreen HTML5Med Res (??MB)HTML4

ASK SILICON CHIP Got a technical problem? Can’t understand a piece of jargon or some technical principle? Drop us a line and we’ll answer your question. Send your email to silicon<at>siliconchip.com.au How to measure harmonic distortion I recently found out about your magazine, and I bought several of your older issues that were very helpful. Do you have a magazine where you described how to measure THD (total harmonic distortion) on audio amplifiers? (B. G., Neu-Isenburg, Germany) • We have not published a dedicated article on the topic, although we have touched on it several times over the years. To measure THD of an amplifier, you need to use a distortion analyser and a low-distortion signal (sinewave) generator. We have published several articles on these in the past. The most recent article is the USB SuperCodec (August-October 2020; siliconchip.com.au/Series/349) and the matching Balanced Input Attenuator (November-December 2020; siliconchip.com.au/Series/349). Those articles also included details on analysis software that can be used in combination with the SuperCodec for measuring THD+N, signal-to-noise ratio, channel separation etc. In terms of a build-it-yourself instrument for distortion measurements, the USB SuperCodec will be hard to beat. We use an Audio Precision System Two, which cost quite a bit new (and they still go for a fair bit of money used). What is a PC stake? I am building the High Power Ultrasonic Cleaner (September & October 2020; siliconchip.com.au/Series/350), and I have a question about the parts list. What is a PC stake? What does it look like, what is it made of and what is its purpose? Can I make one myself (out of thick copper wire)? I can find no mention of them on Jaycar’s website, for instance. (D. P., Noumea) • A PC stake is a small hardened pin (typically 0.9mm to 1mm in diameter) that fits into a hole in a printed circuit board (PCB). It is used to allow wires to be easily soldered, or as a point siliconchip.com.au for connecting multimeter or oscilloscope probes. They are generally optional, and in this case, can be left off if you are happy to directly solder wire or hold a probe to the PCB pad. Jaycar sell PC stakes, Cat HP1250 (they call them “PCB pins”, which is the same thing). Identifying charred component for repair I have an Ozito SNG-956 Staple/Nail Gun that has a burnt-out resistor R1. The circuit appears to use the typical capacitor/resistor in series from the mains to a bridge rectifier to provide DC to the rest of the circuit. I wonder if anyone knows where I can find a circuit diagram for it, or the value of R1. This is quite an old tool, and unfortunately, Ozito can’t get the circuit diagram from the manufacturer. so at the moment, I’m stuck with not being able to repair it. (B. P., Dundathu, Qld) • If the resistor is in series with the capacitor feeding the bridge rectifier, it’s likely to be a relatively low value like 1-10W. Its primary purpose would be to limit the inrush current when power is first applied, and perhaps act as a ‘fuse’ of sorts (it sounds like it did...). It might also exist to drop some of the voltage (although we don’t think that is a very wise design decision), in which case using a low value could cause other components to overheat. Increasing feedback for Motor Speed Controller I purchased an old gem facet machine with a ¼hp induction drive. It shook, rattled and totally stuffed up soft gem faceting. I have modified it to work with a small universal motor and your Full Wave 230V Universal Motor Speed Controller (March 2018; siliconchip.com.au/Article/10998). Actually, I’m using two universal motors, both running via the same type of speed controller. One starts and runs perfectly; I just needed to Australia’s electronics magazine adjust the feedback pot (VR2). The feedback works well, maintaining a fixed speed. This is important for a faceting machine when cutting soft gemstones; unlike diamonds that can handle speed variations, the speed is more critical for soft gems. But the other universal motor will not start, no matter the setting on VR2. I connected a desk lamp in parallel with the motor, and off it went. The speed control was great; remove the lamp load, and the motor continues to run. However, there is a slow-speed point where this motor stops rotating. The only way to get it going again is to have the lamp connected in parallel with the motor. I note that the slow speed point on the other motor is much slower. It appears to be a back-EMF sensitivity constraint; this motor requires a lower feedback threshold to start and rotate as slowly as the other motor. Can I change the CT feedback loop components to improve the sensitivity for motors such as this? Also, can some of the feedback bridge’s values and its associated RC network be adjusted to increase the voltage feedback to the PIC? (J. T., Teneriffe, Qld) • You would need to add more turns of the mains wire through the transformer. That might be difficult as the hole is a small diameter. Select 10A mains wire that will allow more turns through. You could also increase the 510W loading resistor that is across the AX1000 transformer coil. A larger value will increase the output voltage. We used 510W but, for example, a 2.2kW resistor would give a higher feedback voltage. There is a limit to the output versus current response, and it becomes non-linear with greater resistance values. Note: J. T. got back to us and said: “I increased the 510W loading resistor that is across the AX1000 transformer coil to 1kW, and doubled the number of winding through the CT. It works perfectly now.” March 2021  107 Larger display for RPi Tide Clock I want to build the Raspberry Pi Tide Chart from July 2018 (siliconchip.com. au/Article/11142), but I want a larger display than the 2.8in TJCTM24028 screen you used. Do you know if there is a larger screen that is compatible, which could be plugged or wired in to replace the existing small LCD? I would like to put it in a frame that can be placed on the wall for everyone to see. (R.W., Mt Eliza, Vic) • 3.2in and 3.5in variants of the ILI9341-based 2.8in display we used in that project are available, with the same pinout as the TJCTM24028. We suspect that the mounting holes and SD card reader would not match, but the 14-way header for the LCD and touch appears to be the same in each case (which is all that is needed for the Tide Clock). For example, see siliconchip.com. au/link/ab6y and siliconchip.com. au/link/ab6z We haven’t tested any of these, so we can’t comment with any certainty that they would work. Using any other display controller (instead of an ILI9341) would require a major rewrite of the code. Note, though, that you would have many more options for larger (and cheaper) screens if you used one with an HDMI input (ie, a small computer monitor), which is natively supported by the Raspberry Pi. Effect of changing crossover inductor The recommended inductor for the Majestic loudspeaker crossover (June & September 2014; siliconchip.com. au/Series/275), in series with the woofer, is a 2.7mH inductor. The Jaycar Cat LF1330 inductor that was recommended is no longer available. With COVID restrictions, most European suppliers aren’t exporting down under. I can get a 2.5mH air-cored inductor locally. Would this be suitable, or would I need to make other changes? (P. S., Hamilton, NZ) • We doubt you would notice the difference. That is only a 7.4% difference in value, and the tolerance of these inductors is probably ±20% anyway. There might be slightly more midrange getting to the woofer (it has the bandwidth to reproduce up to a few 108 Silicon Chip kHz). In the unlikely event that you can hear the difference, and it is bothersome, you could add 180µH or 220µH air-cored inductors in series with the 2.5mH types. Just make sure they are mounted at right-angles, so their magnetic fields don’t interact. Increasing DC-DC Converter soft-start time I am building the DC-DC Converter to power the CLASSiC-D Class-D amplifier (May 2013; siliconchip.com.au/ Article/3774). Would increasing the value of the 47kW resistor connected to IC1’s pin 4 and the 10µF capacitor be a suitable way to increase the circuit’s soft-start time? I have found that connecting the Mk.3 power supply board with six 4700µF capacitors is too much all at once, resulting in blown STP60NF06 Mosfets, and I am hoping that a softstart modification would make the two projects compatible. Also, the TL494CDR switchmode controller is out of stock at element14 and Mouser at the moment. Are the TL494IDR or TL494CD (both from Texas Instruments) suitable alternatives? (E. B., Viewbank, Vic) • You can slow down the soft-start by increasing the value of the capacitor (originally 10µF) at pin 4 of IC1. 22µF or 47µF capacitors would be suitable. The 47kW resistor value should not be changed. As for the TL494CDR IC specified, the following types are also suitable: TL494CN, TL494CNE4, TL494IN or TL494INE4. In fact, any 16-pin DIP version of the TL494 should work. CLASSiC-D overheating and motorboating I built two of your CLASSiC-D ClassD amp modules (November-December 2012; siliconchip.com.au/Series/17) from Jaycar kits, Cat KC5514. My construction experience is extensive, having been employed by a competitor for six years as the national production manager for local manufacture and kit assembly and tech support. After many years of continual use, my Series 5000 150W modules have died with the 2SK49 and 2SK134 transistors failing, so I decided to upgrade the modules to the cooler/more efficient Class-D type and get more power, 240W into 4W. Australia’s electronics magazine I used a multimeter to check all resistor values and used my phone camera to zoom in on the diode and capacitor markings to make sure I had the correct values in the correct locations on the board. The larger components are easily read, so I completed the PCB assembly and also drilled the heatsinks at the full 75mm height. The steps to confirm the board setup is correct worked as per the instructions, and I can get a clear sound from the amp modules. I’m using a ±50V supply rails from a 35-0-35V toroidal transformer. The amplifiers have tested with the correct voltages, and when I plug in an RCA male to 3.5mm jack cable into my mobile phone, I get clear sound. However, the heatsink is very hot with no signal applied and no speaker connected. When I connect a signal and 8W speaker, the sound remains clear for about one minute, but then I can no longer touch the heatsink, and the amp has distortion until it is turned down and cools a bit. Also, when I plug in a DJ mixer or other preamp device, the amp gets a low-frequency oscillation at full power and the speaker is thumping at full volume. Unfortunately, I lost a lot of gear in a bushfire, so I currently do not have an oscilloscope or signal generator. I’m hoping you can provide some insight into what steps I can take to resolve these problems. (D. F., Perth, WA) • The heatsinks are probably running hot due to the dead time not being sufficient for the Mosfets being used. You can initially lift the 5.6kW resistors between pin 9 and pin 12 of IC1 on each amp board to get the maximum dead time setting (DT4). If the heatsinks run much cooler, that tells you that it was definitely the dead time setting at fault. Our original design uses the DT2 setting. If DT4 works OK, you might like to try DT3, which will give lower distortion. To test this, change the 5.6kW resistor to 8.2kW and the 4.7kW resistor, from pin 9 to ground, to 3.3kW. You will need to verify that the heatsink temperature is still OK with this setting, but if so, it will give you better performance. As for the low-frequency oscillation, that’s possibly due to the power supply cycling up and down in voltage when delivering a high power output. This is explained on pages 21 and 22 of the IRAUDAMP5 Refersiliconchip.com.au ence Design document (siliconchip. com.au/link/ab2a). The recommendation to solve this is to reverse the input and output phases of one of the amplifier modules. This is catered for on our modules by op amp IC2 and link LK2. Simply move the LK2 shunt on one of the modules to the alternative position, then swap the speaker wires to CON3 on that same module. Using Bridge Adaptor with Class-D amplifiers I built a couple of your Bridge Adaptor For Stereo Power Amps (July 2008; siliconchip.com.au/Article/1887) from Altronics K5566 kits, and they work perfectly. Can this adaptor be used with a Class-D amplifier? (P. N., via email) • It depends on the amplifier but probably not, because most Class-D amplifiers already run in bridge mode. If you can’t tell from the amplifier specs/data, check to see if there is continuity between either of the output terminals and ground (generally if there is continuity, it will be with the black/negative output). Continuity to ground suggests that the output is not bridged and you could use a bridge adaptor. Lack of continuity suggests that it is already bridged. You can also tell looking inside the amplifier as a bridged ClassD amplifier usually has two filter inductors per output (ie, four for a stereo amplifier). 12V to 15-35V Inverter output dropping I have just built the 12V 100W Converter With Adjustable 15-35V DC Output (May 2011; siliconchip.com. au/Article/1009). At the top of page 79, there is a graph which shows at 25V you should get 3A. I have hooked up a 12W LED floodlight (Jaycar SL3931), tested on my bench supply at 25V DC as drawing 600mA. But the inverter output drops from 25V to 9V and the current increases to 1.5A, which is not good. I can adjust the output voltage from 12V to 30V. The pin 5 voltage is 1.25V but does change on varying the output. The voltage at the gate of Q1 is very low, less than 1V. I cannot get a steady 10V reading. Reading the project notes, it says I should get 10V at siliconchip.com.au the gate of Q1. I do not understand whether this is with the circuit under load or not under load. At pin 2, I measure a 32kHz signal. Do you know why it can’t drive the floodlight with 24V DC at 600mA? (M. T., Upper Swan, WA) • The lack of output power can be due either to the input supply not being able to deliver the required current and so dropping the voltage, or the current detection resistance is high (R1 on the circuit). Check the input supply and note that it will need to provide over twice the output current when delivering a 25V output with a 12V input. If the input supply is holding up, possibly R1 (the 0.025W resistor) is the wrong value or the connections to the PCB are high resistance. Check the value and also the soldering of this component to the PCB. You might have a dry joint. Power factor correction and mains-borne noise Leo Simpson’s March 2011 editorial (siliconchip.com.au/Article/921) claimed that power factor correction circuitry won’t reduce your energy usage or save money. Yet in this IEEE article, they point out that smart meters can misread when dirty power is fed into them: siliconchip.com.au/link/ab70 As power factor correction reduces noise (aka dirty power), why wouldn’t it reduce your power bill? (M. C., via email) • Power factor correction (PFC) doesn’t usually reduce mains-borne noise. In fact, it can increase noise on the mains supply. Capacitive PFC shifts the current phase to be closer to the voltage waveform, to compensate for inductive loads. It might provide some noise filtering, but that is mostly incidental to how it works. On the other hand, active PFC, which improves the power factor of switching supplies using switching techniques, can inject more noise due to its switching action. Also, power factor correction would typically be applied on the load side of the meter. It’s unlikely to do anything to affect incoming noise from external sources, which must be significant if it is passing through the low-impedance mains distribution network. Australia’s electronics magazine The primary way to reduce noise is filtering. Mains filters are simple and readily available. If smart meters are misreading, that suggests they do not have adequate filtering on the input side and their metering circuitry. Freq/voltage converter for RPM counter Have you published a project or projects that shows how to create a DC voltage directly proportional to frequency, for example, using the LM2917 IC? I want to make an RPM counter for the tail shaft of an irrigation engine. (P. H., Gunnedah, NSW) • Try the Twin-Engine Speed Match Indicator for Boats from the November 2009 issue (siliconchip.com.au/ Article/1622). It could be used for a single engine by tying the pin 10 non-inverting input of IC3c to ground and deleting IC2 (LM2917) and its associated components. There are kits available for this project from Jaycar (Cat KC5488) and Altronics (Cat K6220). Amplifier and power supply kits wanted Do you happen to sell a kit for the 20W Stereo Class-A Power Amplifier (September 2007; siliconchip.com. au/Article/2341), including the chassis? If not, do you have the PCBs and the chassis? Also, do you have a linear DC power supply kit that is not a bench type? I want a supply with 5V, 9V, 12V and 15V outputs, either variable/switchable or a single output with 2A capability. Preferably with chassis. (D. S., via email) • The only kit available for the 20W Class-A amplifier with a chassis was Altronics Cat K5125, but unfortunately, it has been discontinued. We believe that the case is no longer available. You would need to make your own chassis from a standard vented rack case or similar. We do have the PCBs for that project, which you can purchase via this link: siliconchip. com.au/Shop/?article=2283 As for the power supply, we don’t have a non-bench supply that meets your requirements. However, you might want to take a look at the 4-Output Universal Voltage Regulator (May 2015; siliconchip.com.au/ Article/8562). March 2021  109 This has 5V and 3.3V fixed outputs and adjustable positive and negative outputs up to 22V. It does not have 2A capability, however, replacing the LM317 with an LD1085 would likely mean that you can draw over 2A (and possibly as much as 3A) from the positive adjustable output, given a sufficiently beefy DC input supply and enough heatsinking. Failed LC Meter from 2008 I built the LC Meter described in your May 2008 issue (siliconchip.com. au/Article/1822) from an Altronics kit that same year. I was so happy with the result that I have not bothered to build the updated versions described since. However, when I went to use the unit the other day, I noticed the capacitance reading was high. I checked the readings against several capacitors of known value and found that all readings were out by the same amount. I could get useful results from the readings by measuring a known capacitor first, calculating a fudge factor to correct the error in the readings and then applying that factor to the unknown capacitor’s reading. While this allowed me to get on with the work I was doing, I feared it might be the start of bigger problems. I went back to the instructions and re-ran the calibration procedure and found that it gave 0.00pF when started and 49435 with the jumper shunt in LK2, but the display vanished with the jumper in LK1. I checked all solder joints and reflowed a couple of suspect ones without any change in the performance. I cannot see any solder bridges. Any suggestions of what I should check next? (C. K., Parkhurst, Qld) • There isn’t a whole lot to go wrong in that circuit. The lack of display suggests that the microcontroller isn’t running. First, check that the output of REG1 is a steady 5V (4.75-5.25V). The fact that your readings shifted by a consistent amount before it failed completely suggests that there may be a problem with crystal X1. If its frequency changed then that could throw the calibration out, and if it failed entirely then the micro would not run. Check for a 4MHz signal at pin 15 of IC1 (eg, using a scope or frequency meter). 110 Silicon Chip If there is no oscillation then there is something wrong with either crystal X1 or microcontroller IC1. If you have a PIC programmer, it would be a good idea to attempt to reprogram IC1. While we find PICs very reliable, there is a slight possibility that your IC1 chip has failed. In that case, you can order a replacement programmed PIC from us; see siliconchip.com.au/ Shop/9/1277 If the voltage across the electro is low, as is in many coupling circuits where both ends of the capacitor are nominally at ground potential, the orientation doesn’t matter. Typical electrolytic capacitors can tolerate a small DC voltage of either polarity (up to say ±500mV) indefinitely. Modifying the Four Input Mixer I have a question about the FM Wireless Microphone project from your October 1993 issue (siliconchip.com. au/Article/5343). I have been trying to work out how the RF oscillator based around NPN transistor Q3 works, but I have not been able to. There needs to be capacitance across inductor L1 to form a resonant circuit. Is this the Miller capacitance between the transistor base and emitter? I cannot see where the feedback path is for the oscillator, either. Sadly, the article does not specify a value for L1. If you could help me understand how the oscillator works, I would be most grateful. (A. C., Gembrook, Vic) • The 1pF capacitor across inductor L1 forms part of the capacitance necessary for oscillation to occur, but is only a small contributor. Q3’s Miller capacitance would also make a small contribution. The rest is via the 15pF coupling capacitor, which is in series with 33pF and 15pF capacitors to ground. That combination has a total capacitance of around 6pF, and is effectively in parallel with the 1pF directly across L1. You also have to consider trace inductance etc which will significantly reduce the effectiveness of that extra capacitance at 95MHz. As for feedback to make Q3 oscillate, that would be the 33pF capacitor between its base and emitter. The base and emitter are effectively 180° out of phase, so that plus the phase shift introduced by that capacitor should be enough to sustain oscillation. Making RF oscillators work reliably and at a particular frequency is a bit of a black art. It must have taken quite a bit of tweaking for Oatley to come up with the circuit as presented. We could be accused of going into too much detail in our circuit descriptions these days, but your question makes it clear that there was far too little detail in these early articles. There’s continued on page 112 I want to build a variant of the Versatile Four Input Mixer from the June 2007 issue of Silicon Chip (siliconchip.com.au/Article/2256). I only want two inputs, one for a guitar and the other for a CD player. Can I delete the master volume control (VR8) and only use the headphone volume control (VR9)? I want this project to be heard on headphones only. Could you please tell me what other components need to be deleted or added, especially around the master volume control. Also, in this design, you have some coupling electrolytic capacitors that the input goes into the positive side, yet there are some where the negative side is fed a signal. How do you determine which way the cap is supposed to go in these cases? (J. R., Hoppers Crossing, Vic) • You could take the connection that goes to the top of the master volume pot (VR8) and connect this to the top of the headphones volume control VR9 instead. Remove the original connection from the main output. Then the output socket and master volume control can be removed. Electrolytic capacitors are orientated based on the expected DC voltage at either end, ie, with the positive lead to the more positive side. You need to do some circuit analysis to determine the DC operating conditions at either end, or run a simulation, or just build the device with a non-polarised capacitor and measure the voltage before substituting an electrolytic capacitor. One of the trickier aspects of this sort of calculation is taking into account op amp or amplifier input bias currents; analog IC inputs can source or sink current, or do neither, and sometimes that changes depending on certain factors. Australia’s electronics magazine Help to figure out how an oscillator works siliconchip.com.au Notes & Errata USB SuperCodec, August-October 2020: in the Fig.13 circuit diagram on page 88 of the September 2020 issue, pin 12 of IC7 (SDOUT) should not be shown connected to pin 9 of IC6. Instead, it goes to the I2S_ADC1 connection at the right edge of Fig.12 on p86. Car Altimeter, May 2020: the design is missing one schottky diode (D8) which connects from the cathode of ZD1 (schottky anode) to the positive terminal of the battery (schottky cathode). This is needed to charge the battery. It can be added to the underside of the PCB, as shown in the accompanying photograph. Advertising Index Altronics..................17, CATALOG Ampec Technologies................. 20 Analog Devices..................... OBC Dave Thompson...................... 111 Digi-Key Electronics.................... 3 Emona Instruments................. IBC Jaycar............................ IFC,53-60 Keith Rippon Kit Assembly...... 111 LD Electronics......................... 111 LEDsales................................. 111 Microchip Technology.................. 5 Mouser Electronics...................... 7 Ocean Controls......................... 19 SC Colour Maximite 2............... 71 Silicon Chip Binders............... 111 Silicon Chip Shop...............98-99 6GHz Touchscreen Frequency Counter, October-December 2017: in the circuit diagram on pages 30 & 31 of the October 2017 issue, a 1µF bypass capacitor is missing between the anode and cathode of REF1. Also, in the overlay diagram (Fig.3) on p86 of the November 2017 issue, the board shown is RevA; the final (RevB) board adds a 100µF capacitor just to the left of REG2, with its positive lead towards the regulator. The April 2021 issue is due on sale in newsagents by Thursday, March 25th. Expect postal delivery of subscription copies in Australia between March 23rd and April 9th. hardly any mention in that article of how the circuit works! Disconnecting the charger on full battery Some time ago, you advised me how to modify the “Add-On Regulator for 12 Volt Battery Chargers” published in Electronics Australia, June 1997, to charge a 24V SLA battery. I have used it for several years, but I am now planning to upgrade to a 24V Lithiumion battery. If I set the charge voltage to 28.5V, will I need to add extra circuitry to disconnect the battery at this point? Will one of the cut-out modules, as available from eBay, be suitable for this? (B. C., Dungog, NSW) • Yes, you would need to switch off the charger when the Lithium-ion 112 Silicon Chip battery is charged. You could use our Threshold Voltage Switch (July 2014; siliconchip.com.au/Article/7924), sold as a kit by Altronics (Cat K4005) and Jaycar (Cat KC5528). Any other similar device should also work. Graphic Equaliser level matching problem I have been using an Electronics Australia Graphic Analyser for many years, even though it spends most of its life in the cupboard. I drive it with an electret mic which is switchable between 600W and 50kW. The problem is that to get a decent level on the LED display, I must have the sound level in the room extremely high, to the point that I must wear ear protection and only do it when nobody else is at home. Australia’s electronics magazine Silicon Chip PDFs on USB....... 91 Switchmode Power Supplies..... 29 The Loudspeaker Kit.com........... 9 Tronixlabs................................ 111 Vintage Radio Repairs............ 111 Wagner Electronics................... 64 It has always been that way, but it seems that it would be best to do the process at a normal listening level. Is there something I can do to increase the mic preamp gain, or might there be some other problem? The original build did have problems with many dead or partially-dead quad op amps. Might there be more remaining undetected? (R. A., Hunter’s Hill, NSW) • We suggest that you use a preamplifier to boost the microphone signal. Then you won’t need to have the volume so loud. You could use our Multi-Role Champion Preamplifier published in the June 2015 issue (siliconchip.com.au/ Article/8609). It is inexpensive and easy to build, and its gain can be adjusted to suit your needs. We can supply the PCB for that project. SC siliconchip.com.au