Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG The stereo recorder that wasn't Dave Thompson* A friend popped in the other day to talk about video production. He’d heard I was contemplating starting a YouTube channel and wanted to compare notes as to what software and hardware I’d be using. In video production, as in everything else in life, there are many different ways of achieving the same thing. On the face of it, all you need is a computer, a camera and a good idea, yet many of the various guides on the web related to starting a YouTube channel imply that I’d need a raft of expensive video equipment and associated gadgetry if I was to have any chance at all of success. Of course, all the "tools" featured in these guides and videos are affiliatelinked in the video descriptions. Those who click through to purchase what they’ve just been advised as essential kit end up bankrolling the guide authors’ own success by lining their pockets with referral commission cheques. Nice work if you can get it and this explains why there are hundreds of such guides recommending everything from the best cameras, lenses, lights, filters, microphones and recording gear down to the software we can use to edit our masterpieces. There’s nothing wrong with all this as the internet is as valid a platform for marketing and carrying on business as any other medium. What’s difficult for the beginner is sorting through the huge amounts of available information to determine what it is we actually need in order to get a viable channel up and running. Items Covered This Month • • • Dodgy stereo recorder A failed Bose SoundDock Repairing a LED fluoro batten *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz 58 Silicon Chip Celebrating 30 Years siliconchip.com.au My friend and I agreed that we’d need a camera capable of taking at least high definition video (HD-720p, though Full HD-1080p would be better) and preferably with decent audio capture. While most modern smart phones have suitable cameras and would probably do quite nicely, a DSLR or similar purpose-built camera is likely to be better at coping with the different environments and situations my videos will most likely be shot in. A simple high-quality web-cam with built-in audio capture would be adequate if I was just going to sit in front of the computer and do a piece to camera. However, I intend to be filming in the workshop, in various locations which will all have different focal lengths and lighting requirements, so a "proper" camera will theoretically cope better than any smartphone version. tor to allow for recording telephone conversations. All of this is packed into a plush, velvet carry bag; not too shabby for the money. Specifications-wise, it has eight gigabytes of on-board storage, features so-called lossless WAV recording at 192kbits/s, voice-activated recording and selectable high/low quality settings, which also affects how much data can be stored on-board. There is no facility for adding storage space via removable media, but eight gigs should be more than adequate for the sort of work he wants to do with it, with the bumf promising around 48 hours of recording at high quality and 68 hours at the lowerquality setting. All in all, not too bad and while inexpensive, if it lives up to those specifications, it should be a useful tool for his online video production. Separate audio recording The main problem he had was recording in stereo. The built-in mics are by design mounted close together but even with an external stereo mic plugged in, he could not get a proper stereo-sounding spread, with both channels appearing audibly the same when played back. On looking more closely at the device, the dual microphones are clearly marked L and R on the case and while fixed, there are silver mouldings with the usual rows of holes and slots to indicate where they are. Similar-looking microphones are seen on much-higher-end devices like the well-known Zoom range, except that those microphones can usually be extended or otherwise manipulated to point where the recordist wants them to capture sound from; usually a crossed-over formation is used to ensure a nice, even stereo capture. However, the mics on this recorder are fixed facing forward, which means the user must take care to physically position the recorder towards the sound source if they want the best sound capture. The lapel mic has a 3.5mm stereo plug on the end of a 600mm-long cable, although the mic itself appears to be a mono/single capsule type, while the plugs on the line-in lead are also the same 3.5mm stereo versions. Because of this, I assumed the device is capable of stereo recording. Another consideration is audio capture. While most half-way-decent cameras also have stereo recording via built-in microphones, some have surprisingly limited audio recording capability, so an external microphone or even a separate recorder must be used. My friend had such a camera and had bought himself a digital audio recorder online to experiment with. He had brought it with him for me to take a look at because he was having some problems with it. He’d purchased it online from an eBay link in one of those "how-to" videos I mentioned, and while it worked reasonably well, he suspected it wasn’t performing as it should. On the face of it, the digital voice recorder looked like a hot little gadget. He’d only paid something like US$30 for it but it certainly looked the business. It is palm-sized and boasts dual microphones, a built-in speaker and has various sockets allowing for headphones, an external microphone and USB charging/data transfer cables to be attached. It also doubles as an MP3/ audio player. The LCD is clear and reasonably easy to understand and the device comes with a basic set of bud-style headphones, a clip-on lapel microphone and assorted leads for line-in recording. It even comes with an adapsiliconchip.com.au No stereo spread Celebrating 30 Years However, on checking the specifications in the supplied "user manual" (a laughably-small 2-page folded sheet of paper), nothing mentioned stereo recording. I also found this exact device on AliExpress – at half the cost of what my friend paid for it, though I didn’t pass that on to him – and none of the specs there mentioned stereo recording either. While there is obviously stereo playback – it actually plays MP3 files quite well – my subsequent tests showed it January 2018  59 did not record in stereo from any of its input sources. To prove this definitively, I used the supplied stereo line-in cable to connect the recorder’s line input to my computer sound-card’s line-out socket; it definitely rings out as a bona-fide stereo lead according to my multimeter. I set the recorder to HQ, which according to the specs records at a bit rate of 192,000 bits per second (192kbits/s) and played a song on my computer using VLC media player. The source material is definitely stereo and was "ripped" by me from a CD. After hitting record on the recorder, I played the song and when finished, played it back into my computer after swapping the line-out cable at the recorder to headphones out and at the computer end to line-in. While the headphone out signal is usually a lot "hotter" than the typical line input can handle, I just made sure the recorder’s output volume was very low and this kept everything under control. I run a piece of software called Audacity – an excellent freeware program I heartily recommend for anyone into manipulating audio data on their computer – and analysed the track as it came out of the recorder. Both channels were identical, proving the recorder was mixing the two channel sources together into a single, monophonic signal. Boo! Hiss! 60 Silicon Chip Celebrating 30 Years Just to be sure, I did one more experiment, this time using the microphones to record the music as played from my computer through a pair of stereo speakers. The speakers are only about a metre apart on my desk but the stereo separation is readily apparent and any stereo-capable recorder placed in the centre of the spread should pick up and record the stereo sound quite easily. After recording the song and playing it back through Audacity, once again the channels were merged, and listening to it played back through the headphones doubly-confirms there is no stereo spread at all in the resulting recording. So it definitely isn't and cannot do stereo recordings. Double boo! Double hiss! That said, I shouldn’t be too harsh; none of the bumf mentions stereo recording and since this was apparently only going to be used for voice work, recording in mono won’t be too much of a problem. The rather mediocre-by-today’sstandard 192kbits/s bit rate is also fine for voice capture (192kbits/s is only considered medium quality nowadays). Added noise as well However, even if only used for voice, my main concern with any recording device is the quality of the recorded material and especially the signal-to-noise ratio. siliconchip.com.au In other words, the real test is how well the audio is captured, recorded and reproduced. A sound track full of hiss, clicks or pops is no good to anyone. While many software programs are available that make a reasonable attempt at removing hiss and other noise from recordings; it's much better to reduce any noise at the source rather than try to electronically remove it after the fact. The problem with any recorder, no matter the cost or quality, is that noise is produced all the way along the recording chain. The tiny mic capsules add noise; amplification adds noise; any equalisation or normalisation circuitry adds noise and even the headphones or speakers add noise. It’s a miracle any recorded sound is discernible at all! After a few simple recording tests, I found the quality of the captured audio to be quite good and while there is a little background hiss at higher volumes, it is at an acceptable level – at least to my years-of-rock-and-rolldamaged hearing. However, a bigger problem with digital recorders such as this one is handling noise. If using the built-in microphones, the recorder would ideally be used sitting on a desk or other stable surface. Even so, anything bumping or touching the platform while recording can result in unwanted noise being transmitted through the chassis of the recorder, which is why many "better" digital recorders at least try to isolate the mics from the rest of the unit. These mics might be rubber-mounted or have some other method of sonically-isolating them from the recorder so that it can be carried around or otherwise handled without lots of added noise but not so with this recorder! If held in my hand while recording, any tiny movement of the case against my skin is amplified and is very apparent on playback. If sitting on the bench and the bench bumped, the sound is transmitted directly to the microphones. Scratching the table’s underside far away from the mic results in a clear reproduction of the noise on playback, proving it is transmitting through the plastic case to the capsules. None of this is very scientific and any audio engineers reading this are probably shaking their heads but all this leads me to conclude there is no microphone isolation used at all. However, simply using an external mic should resolve this problem. Plugging in an external mic usually disables on-board mics, so my friend could just use the supplied lapel mic clipped to his shirt collar and have the recorder in his pocket. This is all good in theory but when I plugged in the lapel mic to try this out, I could still hear the handling noise, seemingly at the same levels, and it turns out that the on-board microphones are not disabled when Inside the cheap recorder you can see the sole electret microphone aimed towards the right speaker grille. siliconchip.com.au Celebrating 30 Years the external mic is plugged in. Triple boo! Surely it would be easy enough to use a set of contacts in the external mic socket to disable the onboard mics? Every other recorder I have used, that has provision for an external microphone, has the internal mics disabled whenever an external microphone is plugged in; it just makes sense to have this happen. Perhaps in this one the socket hadn’t been wired correctly or something else was amiss inside the unit. It was time to take a look. The recorder was held together with six small, self-tapping screws and though I had to break through one of those screw-cover type security stickers, my friend said he was OK with voiding the warranty as he was unlikely to go through the hassle and expense of returning it, regardless of the outcome. I wasn’t too surprised when I discovered the reason why we could only get monophonic recording with the mics; there is only one capsule! While the shape, design and printing on the exterior of the case implies dual microphones, the sole electret capsule is mounted on the right side only. To make matters worse, it isn’t even mounted pointing forward, but is instead soldered at a 90-degree angle to the sound source if the case is pointed straight ahead. No wonder the recording isn’t that flash, as capsules are designed to be either omni or uni-directional; with no part numbers evident on this capsule there was no way, other than carrying out some complex testing, to accurately determine which type of capsule was used. I considered it might be a dual capsule (if such a thing even exists?), so I checked for multiple leads coming from the capsule. As I suspected, there were only two leads; one ground and one live lead to the circuit board. So what about the possibility of rewiring the mic socket so that the internal mic is disabled when an external mic is plugged in? Unfortunately, that option wasn't possible and the accompanying pictures make it bleedingly obvious why that simple modification just isn't possible. January 2018  61 There simply isn't any room inside the case to make any mods at all; not even to re-orient the mic capsule to possibly improve sound pickup. What an absolute con! Yes, it is a nice looking little package but it just goes to show that appearances can be really deceiving and you cannot take anything for granted. If anything, when you decide to purchase something cheap online from Asia, you probably should be prepared to be disappointed. On the other hand, if a cheap purchase turns out to be a good product, you are a winner. One can hope that more expensive recorders would not rely on deception, but either way I would read the specifications very carefully and not take cheaper goods at face value before buying anything. I called my friend and told him and while he was disappointed and a little annoyed at being duped, he was philosophical and agreed that he didn’t pay a lot for it and it would suffice for his needs, at least until he started making money (if any) from his channel, after which he could invest in some better gear. To be fair, the recorder does an OK job of recording him talking and the noise is manageable as long as the device sits somewhere where it isn’t likely to be bumped or handled whilst recording. He posted a test video privately and the results were actually better than I thought they’d be. Throughout all of this, I started thinking that the better way to go for myself would be to use a camera with a good-quality external mic. However, these mics can be costly, which led me to start thinking about making my own, possibly using one of the 34mm microphone capsules I’ve had my eye on purchasing online for a while now. These capsules are used in highquality broadcast microphones, and with the right pre-amp and phantom power supply circuit, I could make a mic that would be comparable with the commercially-available models. Partial circuit diagram of the power supply for the Bose SoundDock. It would certainly be a good video project, however, I’d better do some more due diligence before I go buying anything… Bose SoundDock repair J. W., of Hillarys, WA managed to keep his daughter happy by repairing the power supply for her (no doubt pricey) Bose SoundDock after it failed. Here's the story in his own words... My daughter recently tried to play some music on her Bose Series 1 SoundDock but could not get it to work. I tried it with another iPod but still no music. I noticed that the iPod was not charging either, so I checked the separate power supply (±18V). Nothing was coming out of it. I cracked open the power supply case and removed the board, a typical switchmode power supply running straight off the mains. There was no circuit available on the 'net so, before I began checking the board, I plugged the unit into my mains isolation transformer. It always pays to be safe. The main storage capacitor (C1) was charged to approximately 340V, so that proved the mains input side of the circuit was working OK. I checked the gate of Q1 for switching pulses but found none. Servicing Stories Wanted Do you have any good servicing stories that you would like to share in The Serviceman column? If so, why not send those stories in to us? We pay for all contributions published but please note that your material must be original. Send your contribution by email to: editor<at>siliconchip.com.au Please be sure to include your full name and address details. 62 Silicon Chip Celebrating 30 Years A common problem in switchmode power supplies is electrolytics which have gone low in value, so I checked C2 but found it too to be OK. I noticed there was a UC3843 IC on the board and discovered from the datasheet that it was a current-mode switchmode controller chip. The data sheet showed that the DC supply went into pin 7 of the IC and was filtered by C2. A +5V reference should be coming out of pin 8, when the DC supply (across C2) to the IC (pin 7) was above 8.5V. When the under voltage lockout (UVLO) circuit detected a voltage lower than that, the 5V reference was disabled. I checked both voltages, only to find no 5V and the supply at pin 7 was only 8.3V. After reading the datasheet more thoroughly, it became clear that during normal operation, power for the IC came from the auxiliary winding on the transformer, through D2 and R4. R1 provided a small bootstrap current (about 1mA) which was supposed to get the circuit running initially. This is because Q1 must be switched on and off for an AC current to flow through the transformer's primary winding and without this, there is no voltage at the secondary and hence no current flow through D2. So the current from R1 has to be sufficient to start Q1 switching, at which point D2 takes over and powers the IC with greater efficiency. I could not get a decent signal on my CRO to see what was happening at start-up so decided to cut a track and put a switch in the power line to the IC. This enabled me to toggle power to siliconchip.com.au Reparing a modern LED "fluoro" batten G. C., of Tawa, in New Zealand loved the increased light output from an array of new LED battens but one of them failed straight away. Fortunately, the repair was easy. He writes... I recently purchased 14 40W 230VAC 1.2m long LED light fittings to replace in like-for-like fashion, a similar number of old twin 40W fluorescent light fittings. These LED fittings were of the Philips brand and were easily installed in place of the original fluoro light fittings, after drilling a new cable hole in the fitting base to suit the original cable installation. Although I do not have any actual light output readings, the new (38W measured consumption) LED fittings were clearly much brighter than the old 80W fluoro fittings, even though they were less than half the wattage. In fact, the LED’s light output without the diffuser fitted was too bright to look at. But we could live with that brightness because of the general increase in lighting levels with the diffuser fitted. Anyway, after turning on the second group of seven newly installed LED lights for the very second time, there was an ominous “phut” noise from one light fitting and the unit was now dark. Well, that seemed to prove the old adage that one could expect an 5% failure rate of new electronic equipment; after installing 14 fittings, that would be about right. But I actually expected better of Philips-branded equipment even though the light fitting was clearly made in China. So, after opening up the LED fitting again to gain access to the supply input terminals, the 230VAC supply was found to be present so the fault had to be with the light fitting. the IC without turning the mains supply on and off. I could now see what was happening when the IC was first supplied with power. There was a 20ms time frame where the +5V reference appeared and pulses came from output pin 6 to switch the primary of the transformer via Q1. siliconchip.com.au These light fittings have a layout of one long row of 96 white SMD LEDs, nearly 1m long. Electrically, two series strings of 48 LEDs were in parallel and were powered from a Philips-branded 230VAC to 160V DC switch-mode power supply unit mounted inside the LED light fitting. With no voltage reading on the power supply output, it was clear that the power supply unit had failed in the first instance. But had a fault with the LEDs caused the power supply to fail? Other than borrowing a power supply from another LED fitting, I had no easy means of providing a 160V DC power supply to test the LEDs with, so that idea was put aside for the meantime. So what to do? Should I claim a replacement power supply (probably not available) or claim a complete new light fitting, under warranty? Or should I take a look at the power supply unit with a view to fixing it, as I was naturally curious to know what the failure was for future reference? In any event, it was easy to pry the metal lid off the power supply and ease the clipped-in PCB out of its case and examine the electronics. My thinking was now that if the fault was not obvious, I could always put the unit back together again and claim the warranty replacement (as there were not going to be any access screws with broken seals to indicate the unit had been tampered with!). But with the power supply unit out of its case, the cause of the initial failure was immediately obvious – one 10mm long PCB “fuse track” in the incoming 230VAC Active mains supply line was blown and clearly violently blown at that. The auxiliary feedback winding then showed pulses which should have been rectified to provide charging current to C2 and so allow the IC to continue to run. Upon testing D2, I found it to be open-circuit. So the IC was doing its job and getting the switching started Celebrating 30 Years That would account for the “phut” heard at the time of the failure. As the unit was rated at 200mA max, I soldered a slim strand of a flexible conductor wire that I deemed to be about the same size as the printed circuit track across the blown track; not an elegant fix but perhaps a practical one, to allow for further testing. But what had made the fuse track blow? A look at the 230VAC input circuitry revealed some suppression equipment and a 4-diode full-wave diode bridge. It was easy to see these plastic diodes had the markings of the venerable 1N4007 1000V 1A type and a quick check with a DMM showed that one of the diodes was shortcircuited. With a bit of luck, this should be an easy fix, as I had my own stock of 1N4007 diodes. The old diode was de-soldered, a replacement diode fitted and then I was ready for the smoke test. This was an anti-climax as the power supply now gave out the expected open-circuit voltage of about 200V DC, with no spectacular fuse failure display. In due course, the power supply unit was re-fitted into the light fitting base and the LED tray re-installed on top. Then when the covers were replaced, the installation was again powered and it all worked, as expected. So, that appeared to be it; a faulty 1N4007 diode had caused the problem. So should I have claimed a replacement power supply unit under warranty? Some would say I should have. But at least I had the satisfaction of knowing what the fault was and that I had saved an otherwise good piece of modern lighting equipment from the rubbish tip. Probably all the supplier would have done was give me a replacement light fitting and throw the faulty light fitting away. but then could not continue with just the bootstrap current, as it requires 17mA for continued operation. I replaced D2 and was rewarded with ±18V at the output connector; a great end to a few hours of circuit fault finding as my daughter has music again. SC January 2018  63