Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG Snoring through the night Dave Thompson* Because I supposedly “know everything about electronics”, I am often called upon to solve all sorts of technical dilemmas. But on this occasion I thought that someone must have already thought of the solution to this problem. Evidently, not. So could I come up with an electronic solution to the age-old problem of snoring and its effects on matrimonial harmony? You be the judge. Over the years, my oldest friends and acquaintances have come to know me as a fix-it man, someone to turn to when something breaks down. However, there was also a time when I got calls to stand in for musician friends in their bands, or for assistance changing a head gasket on a friend’s old banger that really should have been driven to the local scrap yard and left there. Sometimes I’d get a call to help out at the local model aircraft derby and I can still smell the burnt nitromethane and ethanol and hear the scream of high-performance model plane engines. These days, people come to me when their computer doesn’t work properly or their amplifier stops amplifying. And that’s a good thing; pulling allnighters, playing rock and roll and wearing leather pants are for young people (unless you’re a member of the 58  Silicon Chip Rolling Stones) and all that loud music and those model plane engines contributed greatly to the hearing problems I have today. Thankfully, I still sometimes get asked to fix something a bit outside the square, and the following tale illustrates one such occasion. A while ago, an old friend I hadn’t seen for a few years dropped by for the obligatory bull session. We’d talk about lost youth and about how it is wasted on the young and discuss the meaning of life; the usual stuff. When it came to how things were going at home, it turns out that there were problems and it was all down to my friend’s snoring. This is a familiar tale in shared bedrooms everywhere. As us blokes age, we tend to start snoring. Often, it’s down to weaker throat muscles and/or weight gain that often arrives with middle age. Whatever the cause, snoring doesn’t bother the snorer; for some odd reason, they don’t hear it and they’ll often swear they aren’t snoring, even if it sounds like a chainsaw chorus, leading to spouses sometimes making recordings to prove their case. It is only when snoring leads to health issues like sleep apnoea that it has any impact on the snorer; by far the most misery is imposed upon the person sharing the bedroom! Surely, my friend lamented, all this fancy technology and electronics voodoo that I do would have a solution to this problem. He’d searched the inter- net and asked everyone he knew, including his doctor, about what could be done. Lose weight, was the medical advice. Or have surgery. Avoiding dairy after 6pm was a solution for some. Pharmacists and health store workers sold him expensive contraptions; one pulled his tongue out while he slept, while another is an elastic sling that wrapped around his head and held his jaw forward. Others swore that peg-like devices that closed his nose off or opened it up were the answer, but none of these snake oil products stopped his snoring. People on internet forums offered all sorts of traditional suggestions, such as sewing tennis balls into pockets on the snorer’s pyjamas to prevent him sleeping on his back, or advised placing peeled garlic or a raw onion beside his pillow. I’m sure all this advice was given with the best intentions, and no doubt some of these methods worked for some snorers, but unfortunately none of them worked for my friend. Was there a gadget I could make that could help, perhaps? I felt his pain. I’ve been known to snore the odd night myself (apparently) and could empathise; it is no laughing matter. Some snoring is quite dangerous and could be symptomatic of a wider problem, which is why a doctor’s opinion should always be sought. I thought about it a lot over the following days and weeks and came up with an idea that I thought might work. I’ve long since discovered that no matter what bright idea I might have, it will have been thought of before. The Chinese have a proverb, “Nothing is ever new, only what has been forsiliconchip.com.au Items Covered This Month • • • A do-it yourself snoring solution It’s just not cricket Sherwood CD player *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz gotten.” If someone has thought of it before, then the internet would know about it. My idea was simple enough; make up a ‘VOX’, a voice-operated switch that when triggered, would send a warning signal to a pillow speaker placed in or under the snorer’s pillow. This would wake him up (or her; women also snore, though of course they’ll never admit it), and thus stop the noise. Over time, the snorer would be trained to respond to the sound and stop snoring as soon as he starts. Although this method obviously wouldn’t stop a snorer snoring altogether, theoretically at least if the device worked as expected, the impact of the snoring should be greatly reduced. After searching the web I found a single oblique reference to a similar idea but like many historic forum posts, the attachments referred to had long-since disappeared. siliconchip.com.au I did find a very complicated circuit diagram of a device similar to what I was considering, but I could find nothing else about it and decided I’d just have to build up something myself. Like all things in hobby electronics, it could be as simple or as complicated as the builder wants to make it. The premise was simple enough; there were no end of sound or voiceactivated switch circuits out there, and even a few kits for sale at the usual hobby electronics outlets. These are very reasonably priced given that buying all the separate components would likely cost more than the kit, and with a kit you might also get a pre-drilled box and screenprinted PCB, two components that can make a project much more complicated (and costly) if you have to make them up yourself. I was sure S ilicon C hip had produced a VOX in years past and so the magazine’s website was my first stop. I found one project listed; published way back in March 1994, before I started buying the magazine and while the article was available from the Silicon Chip store, there was no PCB, I decided to look at other options. Most of the circuits on the web used an electret microphone insert, maybe a potentiometer, a handful of resistors and capacitors, an op-amp IC and/or a couple of transistors and a relay, all of which I had in my assortment of bits and pieces. However, as I wasn’t certain the idea would even fly and because I didn’t know whether any of those circuits on the internet were proven runners either, I decided to start with a kit; at least I’d know it worked and with so many other unknowns in this project, it would give me a good foundation to build on for the rest of it. It didn’t take long to find and assemble the kit and as expected, it worked perfectly straight out of the box. I made a few simple mods, adding a potentiometer instead of their method of trialling various fixed resistors to determine microphone sensitivity and added a capacitor in order to hold the relay closed for a few seconds longer February 2017  59 Serr v ice Se ceman’s man’s Log – continued than the ‘factory’ setting once noise triggering ceased. I then turned my attention to the annunciator part of the project. My original idea was to use a simple oscillator wired through the VOX relay’s normally open contacts to generate a tone and send it to a small pillow speaker which would theoretically wake the snorer. To accomplish this, I built a twotransistor multivibrator oscillator I’d found a while ago on the Talking Electronics website and had previously used as a basic square-wave generator. It is simple and works well, and it is easy to change the output frequency if required. I built it onto a small piece of Veroboard and by changing the value of two capacitors, fixed the output frequency at around 2kHz. I didn’t measure the frequency exactly, but as it was a rather piercing tone – and all without having to use an additional amplifier stage – I was sure it would do the job of waking anyone very close to the speaker. I also added another pot to allow volume adjustment of the oscillator’s output. The relatively small oscillator board would fit snugly inside the jiffy box I got for the project and I would mount everything into it once I’d proven the concept. All that remained was to feed 9V from the power supply to the VOX board and wire the oscillator output through the relay and out to a mono 3.5mm jack socket that would eventually be mounted on one end of the box. I would then plug my speaker into that. The pillow speaker I had on hand is a commercial model with a speaker of about three inches diameter mounted in a tapered, circular slim-line plastic case so it can easily slip under a pillow without creating an uncomfortable lump. With oblong holes spaced around the circumference, it looks a bit like a UFO from one of those old Gerry Anderson TV shows that were all the rage in the 1970s. With the project working electronically, it was time to put it to the test, so I placed the whole mess of circuit boards, flying leads and pots on a side table beside our bed and rigged up the pillow speaker, warning my wife of the upcoming trials. It would be fine lying in a mess on the table – as long as it worked! I powered it up with a surplus 9V DC plugpack and settled down to try some sensitivity tests. However, right off the bat I could see there was going to be “issues”. In the dead of night, when all through the house, not a creature was stirring, not even a mouse, when the oscillator went off, it sounded like an air-raid siren to anyone in the room! If I backed the volume off to be so quiet that my wife couldn’t hear it, it was too quiet and I doubted it would wake up the snorer either. Another issue with my prototype was the relay; with it clicking away every time the sound switch was activated, even that mechanical noise in the middle of the night was far too disturbing for others. And when I cranked up the sensitivity in order for the microphone to pick up some fake ‘snores’, it would activate the relay and the sound of the relay would activate the VOX and it would then get into a loop and chirp like crazy. If this thing was going to fly, it was going to need further thought! The theory was sound or at least, I thought it was sound. The obvious thing was to find a quieter relay and I’d have to ditch the oscillator. Instead of an oscillator and speaker, I could use a vibrating buzzer, like the ones used in mobile phones. I had a few of these tiny devices sitting in the workshop and so I set out to see what I could do with them. These phone buzzers are simply a tiny DC motor with an imbalanced weight mounted on the rotor shaft. When 5V is applied, the motor spins quite quickly and as it is solidly mounted into the phone’s frame, it vibrates to indicate 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. 60  Silicon Chip a call or other haptic feedback events. It should be simple enough to fashion a vibrating ‘pillow speaker’ and this also had the advantage of simplifying the device, as I could dispense with the oscillator side of it altogether and simply pick 5V from a point in the VOX circuit to run through relay contacts and the ‘speaker’ cable to power the motor. With that in mind, I set about modifying the snore machine. I tried various relays, most of which had different pin connections and so wouldn’t fit in the board. Most clicked, some louder and some quieter but none I tried were suitable. No doubt there are ultra-quiet mechanical relays out there or I could utilise an electronic relay, but the cost and complexity of going down that road started overtaking the original simple idea of the project. I ended up sticking with the original relay but made a ‘cosy’ for it out of foam rubber and insulation tape. Fortunately, I had room on the PCB to add the cover and a couple of layers of foam and tightly-wound tape soon had the activation noise reduced to almost nothing. I also mounted the whole thing in the jiffy box and made up a remote mount for the microphone, removing it from the board and instead soldering a 120mm length of heavy, enamelled copper wire to a PCB pin I’d soldered in the earth side, with the other end to be soldered to the mic’s earth contact. I then ran a length of insulated hookup wire up the side of the copper wire from the live pad on the PCB to the mic’s ‘live’ pad. Before I soldered the mic on, I pushed the wires through a hole I’d drilled in the jiffy box for them, then reinforced the junction with some epoxy resin and covered the exposed part of the ‘stand’ with a couple of lengths of heat-shrink tubing. With the mic soldered on, I covered that with pieces of larger diameter heat-shrink and shrunk it down to cover all the connections. It ended up looking pretty neat, with the mic on a now-movable boom, and it was more isolated from whatever mechanical noise came from the jiffy box. It could also now be directed toward the snorer by gently bending the boom. I modified the pillow speaker by taking it apart and removing the speaker, then gluing the vibrating motor to the bottom of the plastic case with another dab of epoxy resin. Finally, I soldered the old speaker wires to the motor siliconchip.com.au Editor’s Note We actually published a much more recent VOX project in the July 2011 issue, titled “Build A Voice-Activated Relay (VOX)”, by John Clarke. An Altronics kit is available (K5542) and we also stock the PCB in our Online Shop (completed PCB shown below, not to size). In this specific example, we suspect Dave Thompson missed the newer VOX project because it was described as a “voice-activated relay” rather than “voiceactivated switch”. But searching for “VOX” would have found both projects. If readers want to try Dave Thompson’s approach to detecting and stopping snoring, the July 2011 design can be used without the relay since the BC337 used to drive the relay could easily switch the phone buzzer motor directly. However, we would suggest changing the 10kΩ base resistor for the BC337 to 4.7kΩ and also connecting a 47Ω 1W resistor in series with the buzzer motor. No other changes to the circuit should be necessary. before re-assembling the top cover. Trials were much more satisfactory and I was able to get the mic’s sensitivity up high enough to pick up the slightest snore without noise from the relay setting it off. The unit, when assembled and sitting beside the bed, was almost silent in operation, and my wife could no longer hear the relay clicking, nor could she hear the buzzer going off under my pillow. But it sounded quite loud to me and had no problem waking me when something set the device off (it can’t have been me snoring!). One downside, which could perhaps be worked around with a more directional microphone, was that sounds such as the neighbour slamming their car door or giving a quick toot on the car horn when driving off also triggered the device, though those sounds might have woken me anyway. All it needed now was some field testing, and so I passed it on to my friend, who reported last time I saw him that it worked well and there was a lot less stress in the household now that they were both getting a reasonable night’s sleep. Fixed! It’s just not cricket D. P., of Faulconbridge, NSW siliconchip.com.au employed all sorts of fancy methods in an effort to track down an elusive intermittent fault in a Beyonwiz DP-P1 set-top box/PVR. Here’s his story . . . It all started one warm, humid day, when my wife sat down to watch a movie, only to find the projector displaying a very strange, multi-coloured image. It was then that she noticed water dripping from our air conditioner onto the cabinet containing our audiovideo equipment. She quickly turned everything off and called me in to take a look at the problem. The air conditioner, a split-system type, has a trough running along the bottom of the indoor unit to catch the water condensing on the heat exchanger. Unfortunately, the drain hose attached to the end of the trough had become blocked, causing the trough to overflow. Most of the water had gone down the back of the entertainment cabinet and into the set-top box (STB), a Beyonwiz DP-P1. This unit had copped the worst of the flooding, with copious quantities of water lying on top of its case and flowing out the bottom. I quickly disconnected the power to the cabinet and proceeded to mop up as much of the water as I could. I then took the STB outside, removed the cover and emptied out the water. The motherboard still had water on it but someone must have been on my side, because the water had all stayed at one end of the board, well away from the switchmode power supply. And as far as I could see, there was no water on the underside of the board. There didn’t appear to be any damage from the flooding, so I carefully soaked up as much of the remaining water as I could and left the unit in the sun to dry. I then brought the STB back inside and powered it up. To my relief, it booted up normally and all seemed well. It didn’t stay that way though. After a few hour’s operation, we began to notice that off-air images were occasionally pixellating. This problem got worse as time went on, the signal breaks eventually becoming long enough to cause a “No Signal” warning to appear on the screen. My first thought was that maybe I hadn’t pushed the antenna plug in properly and so, on the next signal break, I gave the antenna plug a wriggle. It seemed to have been properly seated but in any case, the signal immediately came good and seemed stable. However, I thought that water might have found its way into the antenna plug/socket, so I pulled the plug out and sprayed it and the socket with WD40. All seemed well after that, with no more pixellation or signal breaks. Unfortunately, it wasn’t long before the problem appeared again. Once again, wiggling the antenna connector “cured” it, so I thought that I would take a closer look at the antenna connector combination. The first thing I did was to try feeding the STB from a different antenna outlet. I ran a coax from another room and once again, the problem appeared to be cured. I then pulled the original antenna plug apart, expecting to find an intermittent short or open circuit. However, everything looked fine and it tested OK with a multimeter, so I put the connector back together again and plugged it back into the STB. Once again, I was greeted with a clean signal with no breaks but the fault returned with a vengeance just a few days later. In fact, the STB now became so unreliable that it was now pot luck as to whether or not it would work on a particular day. And even if it did, it wouldn’t work for long. February 2017  61 Serr v ice Se ceman’s man’s Log – continued Further testing revealed that a TV receiver plugged into the STB’s antenna feed performed perfectly, so the fault was evidently in the STB itself. It was time to set it up on the bench for some serious troubleshooting. I began by establishing that video files recorded on the HDD before all this had happened could be played normally, even when the STB was in the fault condition. This indicated that the problem was confined to the RF section but with the unit on the bench, the fault stubbornly refused to appear . . . most of the time. I thought that this might be because the cover was off but replacing it didn’t have any effect. On the rare occasions that the fault did appear, I tried heating, cooling and flexing the motherboard. Sometimes these actions cleared the fault and sometimes they didn’t. The “cure” always seemed vague but I did eventually get the impression that it was more sensitive in the vicinity of the tuners. The DP-P1 has two tuners so that one program can be watched while a different program is being recorded. These tuners are both branded LG and are housed in small tin-plate boxes with snap-on covers, with a row of pins along the bottom edge. One of the tuners had two Belling-Lee connectors, for RF input and RF output, while the other tuner had no connectors. I later discovered that the tuner with the connectors is designated the “master” and the other tuner, the “slave”. Apparently either tuner can be used for direct viewing or recording, the logic circuitry in the STB sorting out which tuner will be used for which task on any particular occasion. By now, I thought it was a fair bet that I had a tuner fault, although other possible culprits were the power supply and the logic controlling the tuners. Snapping the covers off the two tuners revealed that they were quite similar internally, the main difference being that the slave lacked the two BellingLee connectors. It also lacked the amplifier circuit which evidently fed RF to the output socket and (presumably) to both tuner circuits. However, most of the circuitry in the two tuners looked identical, so I thought that some voltage comparisons might give me a clue. I started with voltage measurements at the pins along the edges of the tun62  Silicon Chip er boards. In the non-faulty condition, the voltage readings were virtually identical between the two tuners. However, it was a different story with the fault condition. Most of the readings were the same but one was very different. In the “good” tuner, the reading on a particular pin was about 6V but in the “bad” tuner, it was around 0V and varying slightly. I traced the circuit back from this pin and this led me to believe that it was probably the tuning voltage but I really needed more data. A internet search for information on these LG tuners draw a blank. However, both were based around a UN6034 IC and I had more luck finding information on this device, a quick search revealing a comprehensive data sheet and an application note. It turned out that the UN6034 IC is virtually a complete digital TV front end. It contains no less than three separate voltage controlled oscillators (VCOs) and associated mixers (one for each of the three TV bands), the logic circuitry for switching between them, a programmable phase locked loop (PLL) for referencing the VCOs, a charge-pump circuit to generate the relatively high (33V) VCO tuning voltage, and an IF amplifier capable of driving a surface acoustic wave (SAW) filter. In addition, the chip incorporates various associated functions, including AGC and lock detectors. As far as I could see, the circuitry around the UN6034 in the tuners conformed quite closely to the application note, so by using the information in the application note, there was a chance I could track down the cause of the volt- age difference between the two tuners. Assuming that what I had measured was indeed the tuning voltage, then according to the data sheet, I should have measured somewhere between 0.4V and 33V over the full VCO tuning range. I was now beginning to suspect that the VCO in the master tuner wasn’t running. Either that or it was well out of range. But how to check it? The digital TV channels in my area are in the high VHF range, so the local oscillator (or VCO) should be running at the RF input frequency plus or minus the IF. My old analog scope would not be in the race at these frequencies but an RF probe that I use with my Heathkit VTVM (now converted to solid state) would probably tell me if the oscillator was running (but not if it was on the correct frequency). The SAW filter in each tuner was marked EPCOS X7253D which, from an EPCOS data sheet, I determined to have a 36MHz centre frequency, so I figured that this must be the IF. I reckoned that if the STB was tuned to our local ABC transmitter on 226.5MHz, then the local oscillator (LO) would be on either 262.5MHz or 190.5MHz, depending whether it was above or below the RF input. I then figured that I should be able to detect the local oscillator signal with my hand-held scanner. With the STB tuned to the ABC frequency, I took a punt that the LO frequency was above the input. My guess was correct – with the scanner tuned to 262.5MHz and a short piece of wire pushed into the scanner’s antenna socket as a “sniffer”, I immediately found a strong signal at that frequency in the vicinity of the IC and the oscillator coil in the nonfaulty condition. By contrast, in the fault condition, This photo shows the two tuners in the Beyonwiz DPP1 PVR. The red arrow indicates the position of the cricket body that was wedged under the master tuner. siliconchip.com.au there was no LO signal and the RF probe indicated no oscillator activity. So what was killing the VCO in the faulty tuner? According to the data sheet, there are a number of legitimate operating conditions that can stop the VCO. In addition, a component fault around the VCO, a fault in the IC itself or a power supply problem could do it. In addition, the VCO locks only when an input signal is present. Loss of RF would thus also kill the VCO, so the problem could be due to a loss of RF into the UN6034. This would need to be investigated but how to do it? My trusty scanner could certainly tune to the RF input frequency but it has a 50Ω input impedance, so it would be unsuitable for signal tracing the high-impedance circuitry in the tuner. In any case, I found that tuning it to a digital TV frequency with an antenna connected produced nothing useful. Even on its widest-bandwidth setting, it could not resolve anything meaningful. What I needed was a wide-band receiver that could decode a digital TV signal. Pondering this, I suddenly realised that I had one on the shelf right in front of me – an ordinary digital TV receiver! But what about its input impedance? 75Ω was better than 50Ω but still not good enough. After some more pondering, I hit on the idea of using a 1/4wave matching section. If I stuck to one frequency and cut a piece of coax to a 1/4-wavelength of that frequency, it would transform the 75Ω input to a high impedance. I wasn’t exactly sure how high but maybe it would be high enough. The wavelength of an electromagnetic signal in free space is given by the formula: λ=v÷f where: λ is wavelength, v is the velocity (300,000km/s) and f is the frequency in Hz. Plugging the ABC transmitter frequency into this formula gives: λ = 300,000 ÷ 226,500,000 = 0.001324503km ≈ 1.3245m. So a 1/4-wavelength would be 0.3311m. I planned to use RG59 coax which has a velocity factor of 0.659. Applying the velocity factor gives 0.3311 x 0.659 = 0.2182, or approximately 218mm. Accordingly, I fitted a connector to a piece of RG59 coax and cut it to 218mm, including the connector. On the open end, I soldered a very short siliconchip.com.au clip lead to the braid to make a ground connection to the tuner. A 10pF capacitor with very short leads was then connected to the inner conductor to act as a DC blocker and probe. It was all a bit rough but it worked! In the non-fault condition I found that I was able to successfully trace the RF signal from the STB’s input connector, through the RF stages and right up to the input pin of the IC, all without upsetting the operation. Apparently the input impedance of my matching section was high enough to do the job. However, once again my sense of having achieved something was to be short-lived. The next time the fault appeared, I found that RF was not reaching the IC and as before, the LO was not running. At this point I was stumped; this was a real chicken-andegg situation. After all this, I still couldn’t tell whether I had a VCO problem or a loss of RF. At this point, replacing the tuner seemed to be the next logical step but new tuners were unavailable as far as I could determine. The only way out would be to purchase a non-working DP-P1 set-top box on eBay, maybe one with a dead hard drive, with the hope of salvaging a good tuner from it. And then, as I was staring gloomily at the motherboard, I saw something I hadn’t noticed before. Hidden just underneath the suspect tuner was a small black blob of something. It was hard to see because there were only a couple of millimetres of clearance under the tuner. I gently poked at the blob with a toothpick and it seemed quite soft, as though it was something organic. Further digging then brought out a sad little pile of exoskeletal remains and decomposing soft tissue. It was hard to tell what it was (or had been) but I am pretty sure it was the body of a species of tiny cricket that is common here during summer. These little critters are small enough to climb through fly screens and often come into the house in hot weather looking for water. Certainly, they are small enough to get into the STB through its ventilation holes. In this case, the little body had been lodged between the motherboard’s ground plane and one of the tuner’s pins. Guess which pin . . . yes, it was the tuning voltage pin! There was a nice film of corrosion where the body had been and with this cleaned up, there was no sign of signal breakup, so I set it up for a soak test. The Beyonwiz STB performed faultlessly for several days and so was put back into service. It has continued working for without fault for several months now. Evidently, the leakage path caused by the dead cricket was enough to kill the VCO. Ironically, the cricket itself was probably killed by the VCO in the first place. If you’re that small, maybe 33V is more than enough to do the job! And the reason for the intermittent nature of the fault? It seems that the resistance of the leakage path fluctuated with humidity. Sherwood CDC-5090R/G CD player A simple fault could have led to an expensive CD player being ditched. Instead, J. W., of Hillarys, WA fixed it for just a few dollars. A friend rang and asked if I could fix his Sherwood 5-disc CD player. When I got it, I plugged it in and found that although the display was working and the disc was being loaded and “played”, no audio was coming from the rear RCA sockets. I searched the net for a circuit diagram to no avail and then I remembered a business called High Country Service Data. I left a request on their website and Steve (the owner) rang me a short time later with the news that he had a circuit diagram but no service manual. This was certainly better than nothing and so I purchased a copy from them ($4) and it arrived five minutes later via email. The circuit showed that the RCA sockets were fed from op amp IC701 which was supplied with ±8V. So the first step would be to check the voltages around this IC. I removed the covers and found that I was able to stand the CD carousel on its end to get to the main PCB. I quickly identified IC701 and found that the -8V supply rail measured only -5V, so I then moved on to the power supply section of the circuit. This revealed that IC102, a 7908 voltage regulator, was supposed to supply the -8V rail. It looked OK and wasn’t getting hot due to overload so I replaced it with one from my box of spares. That solved the problem and I now I had a clean audio signal at the RCA sockets. My friend was delighted as a replacement for this particular unit would cost about $500. SC February 2017  63