It's surprising how design defects like these manage to make their way out of R&D departments to the production line and retail. Great job explaining the way the Class D amp works and how to fix it. Probably going to save a lot of these from becoming e-waste.
@@Sydney268 Yeah like they need to test it at full volume for a long period of time to make sure nothing fails or malfunctions. Unless they don't care about the longevity of their product, which nowadays is sadly becoming the norm.
The application notes were for 8-ohm speakers, but 4-ohm speakers were used in the product. This would explain the different inductor value. Strangely the circuit diagram showed the capacitor and inductor being the same for both 8 and 4 ohms, clearly incorrect. The resistors that burned was part of the Zobel network in series with the capacitor. It probably burned because the filter was connected incorrectly. I assumed you replaced them, probably about 10 ohms. Its function is to keep the amplifier stable. The output inductors do tend to run warm, but of course not as hot as they were. I doubt the amplifier puts out more than about 10 watts with only 15 volts input. Well done in sorting it out.
Thanks Michael 👍Yes I replaced the resistors but the inductors still got hot. I've not really done anything with class D amplifiers before so it's all new to me. I'll have to look up Zobel network. Thanks for commenting 🙂
Yes, I had a look at the data sheet and noticed the filters were the same when the text was clearly implying they would be different. The filters are purely for EMI/RFI control and necessary because the connections to the speakers are long.
They did indeed try to implement a hybrid filter, problem is, when doing that you need to calculate the components to deal with the 2x the maximum output voltage and also the current and that is only for starters, if you look at an RC snubber for a power relay you'll see it is very similar the problem is, even those resistors were totally under speced for the output circuit being driven under modulation, for the basic power rating of the amp it would most probably have been ok, typically, this is referred to as a Zoebel network, and should be designed so that the amp is kept 'stable'........ But, this was inadequate, when combined with modulation..... Very well done sir, another brilliant video!
The filter network is not required for amplifier stability. It is required to control EMI/RFI. If the connections to the speakers were very short the filter could be as simple as a couple of ferrite beads and small capacitors to deal only with very high frequencies. The inductance of the speakers themselves would be the major "smoothing" components. Long connecting wires act as radiating antennas so the high frequency components due to the switched voltage at the amp outputs must be kept off the wires.
@@d614gakadoug9 Gap could have said the Zobel network had to be designed correctly so that it didn't cause instability. They just said it in shorthand the other way round, but was still correct. You are both right.
No problem 👍I usually try and do this on most devices I look at. I've been there myself where a chip has blown on something and I've been trying to find pictures or videos of the board to try and identify it.
I very much admire packaging engineering and design. Back in the 90's, I used to hang around a couple of people that were attending Penn State for packaging engineering degrees. Some of the projects that they had to do for class were very interesting. Like being given an object and one sheet of material to create a box to encompass the thing. It's always fascinated me.
Me too!! Many years back, a friend & I where having a beer and for some reason ended up peeling apart a Marlboro packet. We spotted that if they tweaked the design a bit and changed the nesting, they could save something like 10% of the card used per carton. We wrote to them with our idea......and never heard back.
Excellent fix. It just goes to show you that even Philips makes mistakes. They were once a top electronics producer with great quality and reliability. However, like many others, their quality has now become inconsistent.
I have to say, I did enjoy watching the issue identification process and also the applying of a fix. I'm a software developer and most of my work consist of analyzing strange happenings, go throught the rabbit hole, and eventually propose or apply a fix that could solve the problem. You video shows why knowledge and experience, but also information retrieval is so important. Never assume the others did everything right, and trust your feelings. Sometimes thing are just not right, no matter how pretty or elegant they look like. At this point just a quick shout-out to the internet, that made my live much much easier, having many active communities worldwide to use for information search. I'm sure you feel the same.
Thanks👍 I used to do a bit software too :) ...and reverse engineering. I did some of that with a solar inverter as after I repaired it I needed to adjust a setting so needed a code to access the service menu and the company were not helpful at all! So debugger...etc. I did a video on that too "No password No problem" and ended up writing my own tool to generate passwords to help others in the same situation 🙂 Also I did some messing with the Medical Ultrasound scanner (video too) where I unlocked a ton of hidden features...
Subbed, excellent detective work. It seems that Phillips knew about the problem but rather that go through the expense of recalling and reworking them back from the retail supply chain told them to send them off to e-waste. Seems bonkers from a sustainability point of view.
Thanks 👍I've had things like this happen before with the FLIR thermal camera I repaired and they were unwilling to help, or the solar inverter I repaired, then needed a service password to change a setting in it. I emailed the company for a password and they said pretty much it's an old model, buy a new one. Ended up having to reverse engineer their software and make my own password tool. I did videos on all of these with the emails that I sent them. I emailed them back telling them I fixed it myself and here's the passwords that I made myself 😂😂😂
Everything is disposable, which drives me nuts. Having to discard an entire product due to one component failing is shameful. worse, many manufacturers deliberately sabotage products by making key components out of the wrong material and that is always the part you can't buy. They also add unnecessary components that are designed to fail at the end of the "warranty" period. In my opinion it should be a law that manufacturers provide a lifetime warranty against any defects with no cost to the consumer, instead of being allowed to charge "processing" fees that actually cover their expense to replace the product. If we had such a law, products would be much better quality.
Or even worse, actually repairing a product, then needing to adjust something but they refuse to give you a password or code (such as the solar inverter video I did) where I ended up having to take matters into my own hands 😂😂😂😂
@@RealBLAlley Better yet, the manufacturer should be responsible for e-waste that is sent right back to them at reduced or no cost to the buyer. Make the deadline something like 10 years. They'd fix their act rather fast if it became their problem.
Philips hasn't had facilities to do rework on consumer audio since at least the Gibson Innovations bankrupcy, but realistically probably also before they sold off to Gibson since most of the products were already outsourced and their own facilities brought down to a few hundred employees. They might have gotten some financial compensation from their supplier and that's it.
Another brilliant piece of work. One would assume that the circuit design would be thoroughly tested before entering into production. So this show's that even a large company like Phillips can get things wrong. Thank you for sharing. Really impressed with your work.
The long time subscriber hit this one out of the park! I don't know what they're worth, but looks like a fairly simple modification will make them usable again! Great problem solving skills!
Great video. Thank you for taking the time to record the process. There is one very nice thing about date formats that are YYYY-MM-DD They sort properly in any system as the most significant numbers are to the left like in text. There are also no formats starting with the year that have the day in the middle. I always store dates in text files in this format because they will always be imported correctly by a spreadsheet application without day month ambiguity.
Amplifier Classes: Class A - Amplifies the whole wave cycle, it needs a lot of bias to do this and is conducting all the time. Great sound quality, but around 25% efficient, 75% dissipated as heat. Class B - Amplifies half of the wave. Not used singly in audio amplifiers, but used in RF amplifiers. Efficiency in the 70%+ region. Class AB - Probably the most common audio amplifier configuration. It has the efficiency of a class B, but two active devices are used, one for the positive cycle, and one for the negative, so both halves of the wave cycle are amplified like in a class A. They have to be biased to reduce crossover distortion to a minimum, this happens when one device switches off and the other switches on. Class C - Amplifies less than half of the wave cycle. Not used in audio amplifiers, but is used in RF amplifiers. Technically 100% efficient because there is no bias and the device is off when inactive and not passing a signal, in practice they are about the same efficiency as a class B. Produces a lot of distortion, but this is filtered out by the tuned filters of the RF output stage of the transmitter. Class D - Uses PWM (Pulse Width Modulation) where the active devices are either fully on or fully off, i.e. the device is switching. In this configuration, the active devices only dissipate power during the transitions from on to off and vice versa like digital circuits do, which increases efficiency. Efficiencies of over 90% are common with class D amplifiers, and amplifiers of hundreds of Watts can be squeezed into small enclosures that you can pick up with one hand. The downside is that they are more complex, but their power to weight and size factor negates this.
Class D are very efficient at the rated power but at low powers surprisingly inefficient. Most of the inefficiency being lost in the output filters as heat, as they still need to remove the switching frequency, even with no output.
I am sticking to my speakers that each contain one 15" woofer, 2 dome mid-range drivers, and 4 dome tweeters. Plus my Heathkit AA-1506 audio amplifier is still good enough for me. The audio amp is from 1976, and I built the speakers myself back in 1976 as well. Yep, things from the past last a lot longer than today's stuff.
great fix....i bet the bloke who has the job lot / container load has now set up a repair production line to fix the lot ..following ur youtube video....
Class C amplifiers are for RF transmitters. They run in cutoff at idle, and produce positive “peaks” (when RF drive signal is present) intended for resonant filters and antenna.
I would love to have seen the result using low resistance inductors . Ie: also higher quality inductors. Higher saturation current. loved the video. thanks for all the effort.
How you do all of this thinking and research in 40 minutes I will never know ! so have you 24 more to fix now ? or half a cargo ship !! great stuff .....cheers.
Thanks Andymouse 👍. Unfortunately a lot longer than 40 mins... I actually did a lot more research than is on the video including downloading a tool from Texas Instruments to work out the values for the filters, but I didn't want to make the video an hour long. Then the time to edit it, as I had 5 different camera sources on this video. (An iPhone for the start and finish (as I didn't want to unmount my overhead camera), the thermal camera, microscope camera, camera on the scope and the usual overhead camer). It took me a good couple of days to edit and watch back, re-edit etc. The guy who has the other 24 just wanted to know how to fix them. I think he's already done 1 as we chatted over whatsapp 🙂👍
Thanks Jonny. I'm kind of semi retired. Although we run a small farm with around 70 sheep, and I also do TH-cam a bit although both are more like a hobby than a job.
Heya, nice design fault wel no problem just let re-design it as the manufactor ic shematic tell's us and wolla it's fixed pretty pretty cleaver of you well done
Well done for finding they messed up in thier design 😮😮. And with your mod made it nearer to what it should have been. I would buy a modified one off from you.
Good analyses. Looks to be a Friday night design. The electrolytic at 12:54 is above a ceramic cap and an-off-the-top-of-my-head calc shows 1.5A into the speakers at 20W through what looks to be 1A capable inductors. Oops. Your gonna need a bigger mat.
@@markturner7459 My approx calcs were for 4 Ohm, with 8 Ohm you halve the current (sort of). But, just shoving in a hybrid filter to then trim the top end increases the current (more damping) and that's probably why the resistors burnt. The datasheet looks odd, it seems to show the same filter components irrespective of impedance. I would be very surprised if Philips designed that (unless a few pints of Jenever were involved).
@@BuyitFixit Indeed, though the impedance will make a marginal difference, the cutoff seems to have been changed from 44kHz (the 33kHz in the datasheet is wrong, that would be true for only one inductor) to 36KHz (if the ones on the PCB are 10uH), closer to what the datasheet was aiming for. Pure guesswork on my part but I suspect they intended 8 Ohm speakers then fitted 4 Ohm without compensating for the increased current, frequency response etc. Your workaround should be ok, but if your subscriber has all the time in the world then consider doubling the power handling capability of all the components in the output circuits, as well as investigating that 'flying electrolytic'.
Brilliant Mick I knew you would get the problem sorted out, there no end to your talents 😊 P.s haven’t forgotten be in the post in the next day or so thanks Mick 😊
I have been loving them! Appreciate the time you take to explain the circuits. A lot of repair vids seem to go straight to the fault or diagnose mostly of camera. It's great to see you work it out methodically even if it might be slower to repair.
I love watching your channel, you are so knowledgeable and logical in your approach to your repairs, my only comments are: when you put text up on screen I have trouble finishing reading it, could you leave it up a little longer?, my second thing is: you find and fix all of these things, why haven't you found a nice de-soldering station? I have a Pace micro portable which I love but there are some much less expensive ones that I think would really help you in your work. Keep going with the wonderful videos , I sure do enjoy watching you tackle all of these items, you get right into the component level which is so helpful. Oh yes, and I forgot, maybe a nice cordless screwdriver (does Milwaukee make one that you can repair?).
Thanks 👍I've been on the lookout for a decent rework station. The soldering iron I'm using is ok (although I may get some better tips). The rework station is just a cheap 858D which is not really up to the job on some things. I'm not a fan of cordless drivers, I've used them before and find it hard not to strip screws especially in plastic. I always like to give them a turn manually the opposite way until they click before screwing them in so they don't cut more threads in the plastic. Thanks for commenting 🙂👍
Hey , good Video 👍👍👍👍👍🍻 The heatspot isnt a faulty design from the Ingenieurs. Smaler components which a little bit heat radiate are a sign of power🤣🤪🤪 Circuits that work at border area are cost effective and guarantee a faster failure after warranty. No mistake, a good obeselezens calculation.
I had a great 5.1 Philips CSS5530B soundbar for some years (out of warranty now), that suddenly refused to turn on (with a light smell of "too hot"). After viewing your video I decided to open it, but I couldn't find any sign of a component having melted or burned... I have a multimeter, but much to basic skills to use it smartly like you do. So the mystery remains unsolved. I am pretty upset with Philips, who sells over 500€ soundbars, which after a few years are good to throw away... My Samsung TV is much older than my Philips soundbar, but still works perfectly well...
If it "smells too hot" and you don't have a thermal camera, then you could try using some isopropyl alcohol (brush it on parts of the PCB and then power it on then see where it evaporates quickly, switch off and repeat until you see what is getting hot). Does it have any power indicator? If it lights / doesn't light it may give some idea of where to start looking.
@@BuyitFixit Thanks a lot for your reply ! The device does not light on (no more red power indicator). By luck I do have a thermal camera (that I put on my phone) : I powered the soundbar on, then some spots are rather warm (let say around 45° celsius), but 2 spots are REALLY hot : - a very large square chip rises to 95° celsius, - and a tiny chip on the other side rises to more than 110° celsius ! I don't know if that means something
@@BuyitFixit Sorry for the late answer, I think I won't be able to give you any code or figure about the small hot chip (it is so tiny I think there is no figure on it)... Thanks a lot anyway ! 👍
Great job Class D amplifiers are high frequency switching devices and must have very good sharp filters on their output to reject the switching frequency I'm afraid like a lot of stuff these days poor quality control and design errors yet they get past QC and are for sale
Well technically as the switching frequency should be far above any audio frequency you could run a class D amplifier without an output filter. Only then the switching frequency noise would be dissipated as heat in the voice coil of the loudspeaker instead of in the filter. That and speaker wire becomes an antenna radiating the switching frequency as rf noise.
Calling a PWM amp a digital amp is not totally wrong. In the end, it just flips a switch very quickly and this switch only knows two states: On an off. And that's in fact a digital behavior. The analog wave forms are not generated by the amplifier itself but by the LC resonator (capacitor and inductor) following it and how the wave form looks like is controlled by varying the switching times of that digital switch. And the switching time is controlled by the analog input signal. This is BTW exactly how a 1-bit DAC (digital to analog converter) works, which is the most common form of DAC in use today. People think that DACs must be 16, 20 or 24 bits, as their audio recordings are in that format but instead of fabricating a 16 bit DAC that can produce 65536 different output levels, it's much easier to produce a DAC that only knows 2 output levels and then convert the x-bit value into a switch time that controls how long this 1-bit DAC will remain in one state before switching to the other one. The result at the final output will look the same either way when done correctly. There may be a tiny ripple with a 1-bit DAC but it's at such a high frequency, way higher than what a speaker can produce (so it would never transform into an actual audio signal) and also easy to filter out using an active or even passive low-pass filter, that cuts off after 20 kHz.
You wondered what a class C amplifier was. Class A amps the active element conducts all of the time, Class B uses a pair of active elements with each element conducting on either the positive or negative half cycle of the waveform (lots of distortion around the point where elements turn on due to non linearity), Class AB is similar to Class B except that both elements are biased to have a small current flowing when then they would otherwise be non conducting in a Class B system (better compromise between distortion and power efficiency. Class C amplifiers have the active elements conducting for much less than half of the output cycle, higher efficiency, considerable distortion of the waveform. The were primarily used for RF circuits where there is very significant narrow band tuned circuit filters to eliminate the distortion. However the narrow band tuned circuits severely limit frequency response etc, hence the limitation to RF circuits rather than audio circuits. These 4 classes of amp have existed since the vacuum tube triode was invented and are applicable to any active element family (vacuum tube, bipolar transistor or FET). Class D really only became viable as a design option when FET and particularly MOSFETs become commonly available as bipolar transistors generally either lacked the current capability but had the speed, or had the current capability but significant turn on/turn off delays.
I had an issue with a Vizio sound bar I got on clearance years ago and it worked great until it didn't, I would turn on my TV which would turn on my soundbar and it didn't turn on as usual. So I hit the power buttons on the remote and on the bar itself and still no life, I tried power cycling it and it still didn't work. I figured it was dead, so I left it unplugged for about a day and then it worked. I began to suspect that the built in power supply was faulty and was going to change it but Vizio doesn't make parts available for that model so while it still worked I got the power output of that power supply (12v) and just removed that junk and put an external 12v, 5 amp power supply and it worked great for years after that until I upgraded to a newer Vizio soundbar.
Nice workaround 👍I did something similar to a radio someone brought into the repair cafe where I volunteer at. The internal transformer had failed that produced the 9V DC it used. I just found an old external 9V DC adapter and wired that to the board inside 🙂
Makes you wonder if they're skipping out on proper QC testing somewhere with these things if enough get made with the same problem? Also you'd think for environmental testing they'd consider some people like the heat on rather than having the AC set to ice-box room temperatures, and they should use both ends of the range while also pushing the device to its limits. If not that, then somewhere in the pipeline people are signing off on defective stuff because of things like production delays. Wouldn't be too surprised if that happens too.
It's everything I have several BlackVue dash cams I find 2-5 have issues that require replacement. QC is probably the majority of the issue with modern electronics. That build them so fast.
I expect the subscriber is going to want you to do this modification on his other soundbars so he can sell them on... Great work, and nice explanation of audio amplifier design. Class D amps are everywhere these days - it is how all smartphones can produce so much audio in a small form factor, efficiently.
Thanks 👍 He's quite capable of doing it himself as he repairs bluetooth speaker. This one he was just having trouble working out what the cause of the problem was 🙂👍
"it is how all smartphones can produce so much audio in a small form factor, efficiently" Have to give credit to Digital Signal Processing too. Without a DSP, phones, TVs and even sound bars would sound much worse. The DSP lets manufacturers improve the sound of a speaker that's either constrained by space, size or price.
I saw someone use an E-cigarette and fill it with rosin to create a white coating on the circuit bord and them powered up the bord to find the hot componnents that wet out the rosin vapor film. Neat trick when you don’t have a thermal camera.
Yes I bought one of those a while back too, although I've not used it yet. At a push you can also use IPA and see where it evaporates first when power is applied 👍
@@BuyitFixit nice trick, I’ll try it. I really like your videos there aren’t a lot of electronic diagnostic videos out there and you have a really simple methodology for fixing the problem.
Excellent video, and superb diagnostic skills.. Just out of interest, could you put thermal pads on top of the inductors to bring the heat dissipation to a level that suited the schematic?.. thanks and all the best, kimbo
I think they just screwed up the design on the output, either that or a wrong value component was used on all of them. It was just easier to alter it to the chip manufacturers recommended design instead of trying to fix their mess.
Philips hasn't made any consumer stuff since 1999, as its sub contracted out under brand franchising. 2022 Philips completely sold its consumer compaines. Philips only does a military and medical client base now. I suspect its a class N amp (digital ic amp), and using a trick to decrease audio bandwidth to achieve a louder audio output. Other trick I seen them use 2ohm or even 1ohm deep throw speakers, on a 8 ohm nominal outputing amp, and they just use these hybrid crossover filters for control. Creative Lab has a simular problem with the 1st and 2nd revision main pcbs on the flagship Gigaworks T40 series 1 and series 2. Just like SMPSU's, the boards with N class (one chip simple design digital amps) ic amps tend to blow throu, as the fault stress of one going faulty component stresses out multiple compoments making them a different value, making the system to snowball sprial to another fault shortly after some time of a working fix. The T40 pcb are unfixable, thus why so many 2ndhand T40 's on sale on ebay. In my opinion its these cheap psu plugs and smpsu chocoboxes that allow mains spikes throu, and so on the limit for wattage to save money that the psu output regulation wobbles and usually to a higher value for a moment taking or weakening these all in one amp ics. These pcb designs also drives these amp ics hard, to get the maximum audio loudness out. So these ics are run hot with no heatsink and in a non vented enclosed box. They are simply made to cheap.
The inductor measurement probably seemed off due to the limitations of your tester. I'm guessing that it only goes down to 0.01mH otherwise it should have autoranged to µH at that point (I'm guessing it can't measure that low). And it probably isn't calibrated well enough so what should have shown as 0.01mH showed as 0.02mH. Does the tester have a specification sheet of its range capability?
Yes I was thinking the same, it's only a cheap tester which is why I assumed the inductors were probably ok, but it's the only tool I have that measures inductors currently.
😁"Better than Factory" (as Alex would say) ---- a lot better, as it now actually works. Could be a wrong value capacitor landing in the assembly line component carrier.
Sorry for my reaction i am a fan of this channel and still new to it. I know its not you but that board looks like they have used very cheep nasty ass components as well.
I used to work for Combined Electronic Services way back in 70’s ! C.E.S. Was the service side of Philips at that time! The factory was in Holland! C.E.S. Covered all the Philips and Pye equipment. TV, audio, comes, were all covered ! I worked on audio, radio’s etc! They had their finger in almost everything like washing machines, microwave s and they had a huge market in office equipment! Dictating machines , language labs, voice recognition etc ! They took out the micro cassette for dictating m/c’s and pocket recorders ( tiny gears , pulleys, etc,,.) that lawyers, police, business, people used! I think you can still get them but all digital now! Many people carried that pocket m/c’s on their hip pocket so many ended ended in the toilet!😂 Fraser
You should build your own products mate atleast people could rely on its good quality build and performance if was designed by you as someone who loves fixing faulty electronics myself i really enjoy your content keep the videos coming mate amazing work and logic 👍
Thanks 👍I'm no expert. I haven't really designed much apart from an internet connected incubator for hens eggs a while back, and a couple of arduino button bashing games and a shooting gallery type game.
Haha, you're not an expert you say in the description. I know many "experts" with much less skills. Great approach to get to a possible root cause. And great to experience you can think out of the box (a design fault is not a common path to follow)
The filter has to dissipate everything above 20.000 Hz. With a 48kHz carrier pulse width modulated signal that's a lot of energy that has to be dissipated by the filter coil. Of course it heats up. That's the whole idea.
I didn't buy any Philips labeled stuff since over 20 years by now. The reason for it is that I remarked how much of their stuff does just survive the warranty.
Just started watching, good content interesting. Can i ask you put a link to your phone usb camera , yiu used for the heat detection. Is it android? Regards
I wonder if Phillips and Polk sound bars are made in the same facility? I have a Polk with Alexa that came in the exact same box. The remote even looks the same.
Yes they are very handy! I put some links in the description of the tools that I use and also did a video review on that camera VS the FLIR thermal camera that I repaired in a previous video.
I remember fixing several Phillips portable radios in the '70's, the volume control was mounted on the edge of the circuit board and the corner of the board and control used to snap off. Awful design even in them days.
My guess is that they were either finding problems with EMI/RFI interfering with the wireless stuff in the box or they were failing to meet CE or FCC or other standards for emissions.
The values of the inductor and capacitors in the filter on the datasheet are different then what you measure on the board because the datasheet schematic is using an 8 ohm speaker and the speaker bar is using 4 ohm speakers. The excessive heat might be due to the inductors and resistors in the filter circuit being under sized. As the inductors heat up their current carrying capacity drops, their inductance value changes and being in a small enclosed space probably forces a thermal runaway that eventually damages the inductors and resistors.
Also, in the datasheet there is a ground in the filter design but there doesn't have to be. The filter can be designed just between the two outputs of the H-bridge. Edit: But you need to use higher voltage caps and higher watt resistors as the caps are seeing a higher AC voltage. The filters are difficult to get right and those original capacitors in the original configuration probably didn't have a high enough AC voltage rating to be in that particular filter, they looked really undersized for that even at ~20-40w because of the switching frequency of the amp.
Interesting. On the datasheet it shows the 4 ohm filter just under the 8 ohm one. However both show the same value. I did download a tool from TI to work out the LC filter and it said that a 1.2uF would be better. A 1uF gave a slighty higher cut off frequency. The calculations for all of it were way over my head 😂 luckily they had that spreadsheet that worked it all out for you.
@@BuyitFixit Yeah, I got that spreadsheet too. Class D LC Filter design tool. The "differential mode" design is usually used for BTL/PBTL modes because it can be cheaper with fewer components but it can be dangerous if you don't choose the right components. The capacitor in the middle can look like a dead short at higher frequencies. The "hybrid mode" is pointless to me as it is just as dangerous as the "differential mode" version and doesn't give much more benefits. With the "common mode" filter design is best as the capacitors are not stressed as much and don't need 2 to 3 times the AC voltage rating of the output.
An inductor running at 50 °C is perfectly OK unless the core material is something very strange with an exceptionally low Curie temperature. Any change in inductance due to temperature would be of no consequence in a circuit like this.
@@d614gakadoug9 Yes, the main thing is the AC rating of the capacitors at the output frequency of the class D amp. The filter needs to filter out the PWM frequency and using cheaper capacitors with a series resistor does not help the capacitor handle the AC voltage or frequency of the circuit. Might help with DC voltage but not really AC voltage. The capacitor can still look like a short circuit at the higher frequencies so its a bad idea. The next problem is the current capacity of the inductors but really only when the capacitors are underrated. I've had to replace inductors before because underrated capacitors caused excessive current and heating in the inductors. There was a Corsair 2.1 speaker set that was bi-amped stereo + a subwoofer, three chips total. Two chips in BTL stereo mode and one in PBTL for subwoofer. Hi+Mid amp for left and another Hi+Mid amp for left. It was notorious for power supply failures but the main cause was the design of the Class-D output filters (aside from other issues with the heatsink). They chose to use underrated caps in the output and put resistors inline thinking it would help. The burning of the adhesive they threw around the output filters showed that it was literally burning itself to death at regular output levels. The filter components should never be getting that hot. The solution was to redesign the output filters with new caps and also new inductors for the subwoofer section as there was usually at least one bad inductor in the subwoofer section. The subwoofer section was PBTL mode using a hybrid filter that meant a capacitor directly across the + and - outputs and, of course, it was not rated with a high enough AC voltage for that circuit and would explode or meltdown and then kill an inductor or two. The inductors would have been fine had they chosen a capacitor with the correct AC voltage rating for the hybrid mode design. And even TI doesn't have any design notes for us on how to choose a proper rating for those filter capacitors. :/ I forget which exact capacitor I went with for the hybrid filter center cap in the subwoofer section, but I remember the one that didn't heat up or cause the inductors to heat in testing was a film cap rated something like DC 150v while the power supply for the amp was only DC 21-24v output. That is how I discovered the AC voltage rating was related to the amount of heating I observed. Test two different brand caps with the same DC voltage rating but with different AC voltage ratings, the one with the lower AC voltage specs would always heat up more.
Looks like they were attempting to use the more complex filter to filter out the remaining switching noise but got the layout wrong. They obviously didn't do any long term testing before sending them out the door.
Or a batch was made with the wrong component values or flakey components. Post-assembly inspection of boards with surface mount parts can be a pain because some components, notably almost all ceramic capacitors and many resistors have no markings of any sort. SM ceramic caps have "never" been marked. Resistors larger than 0402 used to be, but lots of manufactures quit marking them years ago. I've seen quite a lot of stuff where there was evidence of good engineering that has been overridden by some other consideration. In most cases the override appeared to be motivated by the desire to save a few pennies. I've had idiots who thought they knew better and didn't follow instructions make a complete mess of building products I've designed. One was doing fine in the hands of one company. Its parent company acquired a competitor and decided the manufacturing should be done by the competitor company because they did more manufacturing. But they were too stupid to follow some instructions for a method that had been carefully worked and made a mess of things. It killed the product. I had another design for which a small batch of prototypes were to be built. My client had picked an enclosure they wanted to use and it presented some challenges. I carefully worked out a method for handling the challenges. I wrote it all up in substantial detail. The job of actually doing the work was handed off to someone known to my client but not to me. The fool complained that I must not be used to working with professionals, then botched the job because he wouldn't let anyone tell HIM how to do anything. I had a client who manufactured a product of my design in-house. It contained several capacitors that were critical to performance and longevity. Some rep from some component vendor offered them something he claimed was a direct equivalent. They were CRAPacitors. I don't do any of this stuff anymore and I'm glad to be out of it.
It was a different sound bar he tried changing the IC on but it didn't make any difference. He sent me a fresh one that hadn't been messed about with 😉
It seems to be a common occurrence of products getting past r&d with stupid issues. At my work we had to swap out 600 Ethernet switches due to two issues. A batch of bad diodes, and a chip burning up due to insufficient heat sinking. Major pain in the butt trying to get down time to swap all of these out. Someone in engineering with a thermal camera should have spotted this right away. I can't blame the manufacturer for a bad batch of diodes. That can happen from any supplier. Not taking a thermal shot of the board is inexcusable. 🤬
@@BuyitFixit it took a year and each switch had at least 50 ports and probably three quarters of them 100 ports. Loved the density of the 100 port switches. Unfortunately our supplier no longer makes them. 24 or 50 port switches only. 😡. Thank you Extreme Networks
Well done Sherlock, another awesome bit of detective work. I guess we need to look out for a flurry of Phillips Soundbars hitting eBay over Christmas now he knows how to fix them 😉
I'm not too impressed with most of Philips' later tech. Although on TV they're still great. Bought a cheap soundbar, B5105, without sub for use in bedroom. Well.. Yes it is better than the TV speaker but even for the cheap price it's not really a good buy. I then gave a cheap Sony soundbar to my sisters kids, an HT-SF150, and that is worth every penny. Doesn't have deep bass but it has what you reasonably expect these days and great tonal balance - doesn't sound thin like the Philips. While it only has Dolby 5.1 (no Atmos) it is very convincing when you sit right and every streaming box has capability to down convert anyway. Tt was just pennies more in contrast to the Philips yet so much better.
I've got a Sony TV and sound bar and the sound is fairly good. The Philips one isn't that good at all. There's a number of reviews on Amazon that pretty much say that the TV speakers sounded better and it was more of a downgrade 😂
Wow very nice video you are professional engineer master of the master 👏 I have philips pronto tsu9600 touch screen remote control can you please kindly do a video of it as its not working ie not powering up it was working before I changed new 18650 battery full changed still it doesn't switching on? If make a video of this it will be the best video on TH-cam please kindly help me on this issue 🙏 waiting for your reply Thankyou ps how can I send this to you for repair 🙏please
Someone had a container full of a single product all with the same fault. That's not janky at all. They already replaced it with the B5306 which works well. It's a shame they didn't simply update the board and replace the faulty ones instead of designing and manufacturing an entirely new model while these were packed up as e-waste.
There is something strange afoot there. The over-temperature shutdown threshold for the amplifier IC is nominally 150 °C (die temperature). The top of the package will be cooler than the die, but not by a huge amount. I'd expect the top of the package to be in the range of 100 °C at shutdown. Since no testing to shutdown was done with board exposed we have no useful data. My suspicion is that shutdown was occurring due to something in the filter network causing short circuiting when hot, which also causes shutdown.
I see design errors like this very often. I always ask myself why a chip manufacturer specifies certain component values and their arrangement and wiring in their data sheet if something different is installed in the end. The probability that a chip will no longer work within the manufacturer's specifications due to different component values and wiring is very high. Is this planned obsolescence, or simply a design flaw? Are such devices no longer tested extensively as prototypes before they are mass-produced and sold?
Yes it is odd. I guess they were trying to improve the sound quality or get more out of the chip perhaps? Thanks for commenting and have a great Christmas 👍
It's surprising how design defects like these manage to make their way out of R&D departments to the production line and retail. Great job explaining the way the Class D amp works and how to fix it. Probably going to save a lot of these from becoming e-waste.
Thank you 🙂👍Yes hopefully it might help others with the same problem.
Yeah I was thinking that, the design must have worked at some point?!
I think the problem only starts showing when at loud volume after about 15 or 20 minutes.
@@BuyitFixit sure but one of the first tests I'd do is full power for an hour or so to see how hot it got - come on Philips R&D
@@Sydney268 Yeah like they need to test it at full volume for a long period of time to make sure nothing fails or malfunctions. Unless they don't care about the longevity of their product, which nowadays is sadly becoming the norm.
The application notes were for 8-ohm speakers, but 4-ohm speakers were used in the product. This would explain the different inductor value. Strangely the circuit diagram showed the capacitor and inductor being the same for both 8 and 4 ohms, clearly incorrect.
The resistors that burned was part of the Zobel network in series with the capacitor. It probably burned because the filter was connected incorrectly. I assumed you replaced them, probably about 10 ohms. Its function is to keep the amplifier stable.
The output inductors do tend to run warm, but of course not as hot as they were. I doubt the amplifier puts out more than about 10 watts with only 15 volts input. Well done in sorting it out.
Thanks Michael 👍Yes I replaced the resistors but the inductors still got hot. I've not really done anything with class D amplifiers before so it's all new to me. I'll have to look up Zobel network. Thanks for commenting 🙂
Yes, I had a look at the data sheet and noticed the filters were the same when the text was clearly implying they would be different.
The filters are purely for EMI/RFI control and necessary because the connections to the speakers are long.
@@BuyitFixityou can try to put a small fan to cool them down
@denisstefan2820 No need, works fine after the modification.
@@BuyitFixit oh ok
They did indeed try to implement a hybrid filter, problem is, when doing that you need to calculate the components to deal with the 2x the maximum output voltage and also the current and that is only for starters, if you look at an RC snubber for a power relay you'll see it is very similar the problem is, even those resistors were totally under speced for the output circuit being driven under modulation, for the basic power rating of the amp it would most probably have been ok, typically, this is referred to as a Zoebel network, and should be designed so that the amp is kept 'stable'........ But, this was inadequate, when combined with modulation..... Very well done sir, another brilliant video!
Thank you, and thanks for the info 👍
The filter network is not required for amplifier stability. It is required to control EMI/RFI. If the connections to the speakers were very short the filter could be as simple as a couple of ferrite beads and small capacitors to deal only with very high frequencies. The inductance of the speakers themselves would be the major "smoothing" components. Long connecting wires act as radiating antennas so the high frequency components due to the switched voltage at the amp outputs must be kept off the wires.
This. Great explination.
@@d614gakadoug9 Gap could have said the Zobel network had to be designed correctly so that it didn't cause instability. They just said it in shorthand the other way round, but was still correct. You are both right.
Your persistence always gives me so much encouragement to try again when I'm stuck troubleshooting ! THANK YOU!
You're very welcome, and thanks so much for your kind comments 👍
Thank you for taking the time to identify the ic’s. Really appreciated.
No problem 👍I usually try and do this on most devices I look at. I've been there myself where a chip has blown on something and I've been trying to find pictures or videos of the board to try and identify it.
I very much admire packaging engineering and design. Back in the 90's, I used to hang around a couple of people that were attending Penn State for packaging engineering degrees. Some of the projects that they had to do for class were very interesting. Like being given an object and one sheet of material to create a box to encompass the thing. It's always fascinated me.
Yes, unusual shape box, but two if them fit together nicely as shown in the stack on the video thumbnail 👍
Me too!! Many years back, a friend & I where having a beer and for some reason ended up peeling apart a Marlboro packet. We spotted that if they tweaked the design a bit and changed the nesting, they could save something like 10% of the card used per carton. We wrote to them with our idea......and never heard back.
Never thought that one could fix a design fault like that, good to know!
Thanks 🙂👍
Excellent fix.
It just goes to show you that even Philips makes mistakes. They were once a top electronics producer with great quality and reliability. However, like many others, their quality has now become inconsistent.
Thanks 👍🙂
Philips doesn't exist as an electronics manufacturer. They only licence the name.
@@argusdenia well the Philips brand used to be good, now it's just hit and miss whether the product is any good or not
😠 ~ I think “PHILIPS” is now a zombie brand and being used by “Chineseum crap” factory’s - a name slapped on their polished turds
China shit with a philips name.
I'm dutch and just to tell sorry about it.
I have to say, I did enjoy watching the issue identification process and also the applying of a fix. I'm a software developer and most of my work consist of analyzing strange happenings, go throught the rabbit hole, and eventually propose or apply a fix that could solve the problem. You video shows why knowledge and experience, but also information retrieval is so important. Never assume the others did everything right, and trust your feelings. Sometimes thing are just not right, no matter how pretty or elegant they look like. At this point just a quick shout-out to the internet, that made my live much much easier, having many active communities worldwide to use for information search. I'm sure you feel the same.
Thanks👍 I used to do a bit software too :) ...and reverse engineering. I did some of that with a solar inverter as after I repaired it I needed to adjust a setting so needed a code to access the service menu and the company were not helpful at all! So debugger...etc. I did a video on that too "No password No problem" and ended up writing my own tool to generate passwords to help others in the same situation 🙂 Also I did some messing with the Medical Ultrasound scanner (video too) where I unlocked a ton of hidden features...
Subbed, excellent detective work. It seems that Phillips knew about the problem but rather that go through the expense of recalling and reworking them back from the retail supply chain told them to send them off to e-waste. Seems bonkers from a sustainability point of view.
Thanks 👍I've had things like this happen before with the FLIR thermal camera I repaired and they were unwilling to help, or the solar inverter I repaired, then needed a service password to change a setting in it. I emailed the company for a password and they said pretty much it's an old model, buy a new one. Ended up having to reverse engineer their software and make my own password tool. I did videos on all of these with the emails that I sent them. I emailed them back telling them I fixed it myself and here's the passwords that I made myself 😂😂😂
Everything is disposable, which drives me nuts. Having to discard an entire product due to one component failing is shameful. worse, many manufacturers deliberately sabotage products by making key components out of the wrong material and that is always the part you can't buy. They also add unnecessary components that are designed to fail at the end of the "warranty" period.
In my opinion it should be a law that manufacturers provide a lifetime warranty against any defects with no cost to the consumer, instead of being allowed to charge "processing" fees that actually cover their expense to replace the product. If we had such a law, products would be much better quality.
Or even worse, actually repairing a product, then needing to adjust something but they refuse to give you a password or code (such as the solar inverter video I did) where I ended up having to take matters into my own hands 😂😂😂😂
@@RealBLAlley Better yet, the manufacturer should be responsible for e-waste that is sent right back to them at reduced or no cost to the buyer. Make the deadline something like 10 years. They'd fix their act rather fast if it became their problem.
Philips hasn't had facilities to do rework on consumer audio since at least the Gibson Innovations bankrupcy, but realistically probably also before they sold off to Gibson since most of the products were already outsourced and their own facilities brought down to a few hundred employees. They might have gotten some financial compensation from their supplier and that's it.
Another brilliant piece of work. One would assume that the circuit design would be thoroughly tested before entering into production. So this show's that even a large company like Phillips can get things wrong. Thank you for sharing. Really impressed with your work.
Cheers mate 👍
The long time subscriber hit this one out of the park! I don't know what they're worth, but looks like a fairly simple modification will make them usable again! Great problem solving skills!
Thanks 🙂👍yes a little bit from the normal this one 🙂
anything to keep em out of landfill!
Great video. Thank you for taking the time to record the process.
There is one very nice thing about date formats that are YYYY-MM-DD They sort properly in any system as the most significant numbers are to the left like in text. There are also no formats starting with the year that have the day in the middle. I always store dates in text files in this format because they will always be imported correctly by a spreadsheet application without day month ambiguity.
Thanks 👍and thanks for the date format info 🙂
Amazing skills Mick, great video!
Thanks Vince 👍yes a little bit out of the norm this one 🙂
I love your calm and collected approach when repairing something. Like AvE said...if it's efft is not like you can eff it even more :)
Thanks 👍 and exactly right 😂😂
Amplifier Classes:
Class A - Amplifies the whole wave cycle, it needs a lot of bias to do this and is conducting all the time. Great sound quality, but around 25% efficient, 75% dissipated as heat.
Class B - Amplifies half of the wave. Not used singly in audio amplifiers, but used in RF amplifiers. Efficiency in the 70%+ region.
Class AB - Probably the most common audio amplifier configuration. It has the efficiency of a class B, but two active devices are used, one for the positive cycle, and one for the negative, so both halves of the wave cycle are amplified like in a class A. They have to be biased to reduce crossover distortion to a minimum, this happens when one device switches off and the other switches on.
Class C - Amplifies less than half of the wave cycle. Not used in audio amplifiers, but is used in RF amplifiers. Technically 100% efficient because there is no bias and the device is off when inactive and not passing a signal, in practice they are about the same efficiency as a class B. Produces a lot of distortion, but this is filtered out by the tuned filters of the RF output stage of the transmitter.
Class D - Uses PWM (Pulse Width Modulation) where the active devices are either fully on or fully off, i.e. the device is switching. In this configuration, the active devices only dissipate power during the transitions from on to off and vice versa like digital circuits do, which increases efficiency. Efficiencies of over 90% are common with class D amplifiers, and amplifiers of hundreds of Watts can be squeezed into small enclosures that you can pick up with one hand. The downside is that they are more complex, but their power to weight and size factor negates this.
Thanks for that 👍
Class D are very efficient at the rated power but at low powers surprisingly inefficient. Most of the inefficiency being lost in the output filters as heat, as they still need to remove the switching frequency, even with no output.
I am sticking to my speakers that each contain one 15" woofer, 2 dome mid-range drivers, and 4 dome tweeters. Plus my Heathkit AA-1506 audio amplifier is still good enough for me. The audio amp is from 1976, and I built the speakers myself back in 1976 as well. Yep, things from the past last a lot longer than today's stuff.
Agree 👍
great fix....i bet the bloke who has the job lot / container load has now set up a repair production line to fix the lot ..following ur youtube video....
I just messaged him earlier, and he's done one so far but not tested it yet 😂😂😂
Class C amplifiers are for RF transmitters. They run in cutoff at idle, and produce positive “peaks” (when RF drive signal is present) intended for resonant filters and antenna.
Someone mentioned that in another comment but thanks for taking the time to also comment and let me know 👍
I love seeing the fault fining process and learning new methods, this a well earned SUB, cheers!
Thanks for the sub!
I would love to have seen the result using low resistance inductors . Ie: also higher quality inductors. Higher saturation current. loved the video. thanks for all the effort.
Thanks 👍
A beautiful mind. I have no other words. Just so impressive. Well done!
Thank you kindly Matt 😊👍
Brilliant detective work,very impressed with it -and the outcome. Well done, subbed.
Much appreciated!
That street fighter 2 arcade machine must be epic to own! Nice video as always.
@@perkulant4629 Thanks, it's pretty awesome 👍I did a video on repairing that too if you didn't see it...
Nice Job Good bit of detective work, Im amazed they dont put at least a small heat sink on that amp chip great video
Thanks 👍
How you do all of this thinking and research in 40 minutes I will never know ! so have you 24 more to fix now ? or half a cargo ship !! great stuff .....cheers.
Thanks Andymouse 👍. Unfortunately a lot longer than 40 mins... I actually did a lot more research than is on the video including downloading a tool from Texas Instruments to work out the values for the filters, but I didn't want to make the video an hour long. Then the time to edit it, as I had 5 different camera sources on this video. (An iPhone for the start and finish (as I didn't want to unmount my overhead camera), the thermal camera, microscope camera, camera on the scope and the usual overhead camer). It took me a good couple of days to edit and watch back, re-edit etc. The guy who has the other 24 just wanted to know how to fix them. I think he's already done 1 as we chatted over whatsapp 🙂👍
Yeah I can believe it ! thanks once again for your monumental efforts...cheers. @@BuyitFixit
Amazing fault finding skills, what do you do for your 'day job'? A man of many talents..
Thanks Jonny. I'm kind of semi retired. Although we run a small farm with around 70 sheep, and I also do TH-cam a bit although both are more like a hobby than a job.
Great faul diagnosis you need to now give them the fix and take a commission on what they sell lol you the man. 👍
Thanks 👍😂😂😂
Heya, nice design fault wel no problem just let re-design it as the manufactor ic shematic tell's us and wolla it's fixed pretty pretty cleaver of you well done
Thanks 🙂👍
Well done for finding they messed up in thier design 😮😮.
And with your mod made it nearer to what it should have been.
I would buy a modified one off from you.
Thanks 👍Honestly I wouldn't bother (not that I have any to sell) but they not very good. I'd save your money and buy something decent.
@@BuyitFixit Thank you for being truthful.
No worries 👍
Hi, sound man, sound! Well done as usual. Many thanks, great the way you explain every step. From Nr Chester.
😂😂😂Thanks 🙂👍
Good analyses. Looks to be a Friday night design. The electrolytic at 12:54 is above a ceramic cap and an-off-the-top-of-my-head calc shows 1.5A into the speakers at 20W through what looks to be 1A capable inductors. Oops. Your gonna need a bigger mat.
The data sheet showed 8ohm speakers..how would them using 4ohm speakers effect things?
@@markturner7459 My approx calcs were for 4 Ohm, with 8 Ohm you halve the current (sort of). But, just shoving in a hybrid filter to then trim the top end increases the current (more damping) and that's probably why the resistors burnt. The datasheet looks odd, it seems to show the same filter components irrespective of impedance. I would be very surprised if Philips designed that (unless a few pints of Jenever were involved).
The recommended filter page on the datasheet showed the same values for 8 ohm and 4 ohm speakers 🤷♂️
by the first look it likes part off the crossover to speakers
@@BuyitFixit Indeed, though the impedance will make a marginal difference, the cutoff seems to have been changed from 44kHz (the 33kHz in the datasheet is wrong, that would be true for only one inductor) to 36KHz (if the ones on the PCB are 10uH), closer to what the datasheet was aiming for. Pure guesswork on my part but I suspect they intended 8 Ohm speakers then fitted 4 Ohm without compensating for the increased current, frequency response etc. Your workaround should be ok, but if your subscriber has all the time in the world then consider doubling the power handling capability of all the components in the output circuits, as well as investigating that 'flying electrolytic'.
Brilliant Mick I knew you would get the problem sorted out, there no end to your talents 😊
P.s haven’t forgotten be in the post in the next day or so thanks Mick 😊
Thanks Gary 👍
@@BuyitFixit no thank you ☺️
Great video. Highly educational. Learnt a heap from you. Thanks from. Australia.
Thanks 👍 Glad you liked it. Hopefully you'll like some of my other videos too 🙂
I have been loving them! Appreciate the time you take to explain the circuits. A lot of repair vids seem to go straight to the fault or diagnose mostly of camera. It's great to see you work it out methodically even if it might be slower to repair.
I love watching your channel, you are so knowledgeable and logical in your approach to your repairs, my only comments are: when you put text up on screen I have trouble finishing reading it, could you leave it up a little longer?, my second thing is: you find and fix all of these things, why haven't you found a nice de-soldering station? I have a Pace micro portable which I love but there are some much less expensive ones that I think would really help you in your work. Keep going with the wonderful videos , I sure do enjoy watching you tackle all of these items, you get right into the component level which is so helpful. Oh yes, and I forgot, maybe a nice cordless screwdriver (does Milwaukee make one that you can repair?).
Thanks 👍I've been on the lookout for a decent rework station. The soldering iron I'm using is ok (although I may get some better tips). The rework station is just a cheap 858D which is not really up to the job on some things. I'm not a fan of cordless drivers, I've used them before and find it hard not to strip screws especially in plastic. I always like to give them a turn manually the opposite way until they click before screwing them in so they don't cut more threads in the plastic. Thanks for commenting 🙂👍
A cordless screwdriver with a good clutch mechanism, not one that is really a drill.
Nice bit of deduction! One thing good about the 'net, looking up data sheets!
Thanks Tom, yes the internet is usually good for things like that 👍
Great job mate. I enjoy your videos and learn from each
Awesome, thank you!
Great forensic work as usual, thanks Mick 😀
Thanks Mike 👍
I always knew you were a sound guy Mick.
😂😂😂😂😂😂👍
Such amazing talents. Thanks for sharing.
Thanks 👍
Even the cat dislikes the "lift musicke" 😆. Excellent video!
😂😂😂Thanks 👍
Looks like you could be onto making a few bob fixing that container load.👍 Good on ya😊
They're not mine mate. I was just fixing this one to show the guy how to do it 🙂
Hey , good Video 👍👍👍👍👍🍻
The heatspot isnt a faulty design from the Ingenieurs.
Smaler components which a little bit heat radiate are a sign of power🤣🤪🤪
Circuits that work at border area are cost effective and guarantee a faster failure after warranty.
No mistake, a good obeselezens calculation.
Thanks 👍😂😂😂
I had a great 5.1 Philips CSS5530B soundbar for some years (out of warranty now), that suddenly refused to turn on (with a light smell of "too hot"). After viewing your video I decided to open it, but I couldn't find any sign of a component having melted or burned...
I have a multimeter, but much to basic skills to use it smartly like you do.
So the mystery remains unsolved.
I am pretty upset with Philips, who sells over 500€ soundbars, which after a few years are good to throw away... My Samsung TV is much older than my Philips soundbar, but still works perfectly well...
If it "smells too hot" and you don't have a thermal camera, then you could try using some isopropyl alcohol (brush it on parts of the PCB and then power it on then see where it evaporates quickly, switch off and repeat until you see what is getting hot).
Does it have any power indicator? If it lights / doesn't light it may give some idea of where to start looking.
@@BuyitFixit Thanks a lot for your reply !
The device does not light on (no more red power indicator).
By luck I do have a thermal camera (that I put on my phone) : I powered the soundbar on, then some spots are rather warm (let say around 45° celsius), but 2 spots are REALLY hot :
- a very large square chip rises to 95° celsius,
- and a tiny chip on the other side rises to more than 110° celsius !
I don't know if that means something
Sounds like that might be the problem. If I'd perhaps look up the chip numbers on the internet to see what the chips actually are and do.
@@BuyitFixit Sorry for the late answer, I think I won't be able to give you any code or figure about the small hot chip (it is so tiny I think there is no figure on it)... Thanks a lot anyway ! 👍
Great job Class D amplifiers are high frequency switching devices and must have very good sharp filters on their output to reject the switching frequency I'm afraid like a lot of stuff these days poor quality control and design errors yet they get past QC and are for sale
Thanks Paul 👍
@@BuyitFixitgreat job pal
Well technically as the switching frequency should be far above any audio frequency you could run a class D amplifier without an output filter. Only then the switching frequency noise would be dissipated as heat in the voice coil of the loudspeaker instead of in the filter. That and speaker wire becomes an antenna radiating the switching frequency as rf noise.
The fixer of Everything!
Well...I do try 🙂👍
Calling a PWM amp a digital amp is not totally wrong. In the end, it just flips a switch very quickly and this switch only knows two states: On an off. And that's in fact a digital behavior. The analog wave forms are not generated by the amplifier itself but by the LC resonator (capacitor and inductor) following it and how the wave form looks like is controlled by varying the switching times of that digital switch. And the switching time is controlled by the analog input signal.
This is BTW exactly how a 1-bit DAC (digital to analog converter) works, which is the most common form of DAC in use today. People think that DACs must be 16, 20 or 24 bits, as their audio recordings are in that format but instead of fabricating a 16 bit DAC that can produce 65536 different output levels, it's much easier to produce a DAC that only knows 2 output levels and then convert the x-bit value into a switch time that controls how long this 1-bit DAC will remain in one state before switching to the other one. The result at the final output will look the same either way when done correctly. There may be a tiny ripple with a 1-bit DAC but it's at such a high frequency, way higher than what a speaker can produce (so it would never transform into an actual audio signal) and also easy to filter out using an active or even passive low-pass filter, that cuts off after 20 kHz.
Thanks 👍yes was thinking that but when I looked on the internet it said it wasn't digital as that's what I was calling it at first 🤷♂️
You wondered what a class C amplifier was. Class A amps the active element conducts all of the time, Class B uses a pair of active elements with each element conducting on either the positive or negative half cycle of the waveform (lots of distortion around the point where elements turn on due to non linearity), Class AB is similar to Class B except that both elements are biased to have a small current flowing when then they would otherwise be non conducting in a Class B system (better compromise between distortion and power efficiency. Class C amplifiers have the active elements conducting for much less than half of the output cycle, higher efficiency, considerable distortion of the waveform. The were primarily used for RF circuits where there is very significant narrow band tuned circuit filters to eliminate the distortion. However the narrow band tuned circuits severely limit frequency response etc, hence the limitation to RF circuits rather than audio circuits. These 4 classes of amp have existed since the vacuum tube triode was invented and are applicable to any active element family (vacuum tube, bipolar transistor or FET). Class D really only became viable as a design option when FET and particularly MOSFETs become commonly available as bipolar transistors generally either lacked the current capability but had the speed, or had the current capability but significant turn on/turn off delays.
Thank you for taking the time with your detailed explanation 👍
I had an issue with a Vizio sound bar I got on clearance years ago and it worked great until it didn't, I would turn on my TV which would turn on my soundbar and it didn't turn on as usual. So I hit the power buttons on the remote and on the bar itself and still no life, I tried power cycling it and it still didn't work. I figured it was dead, so I left it unplugged for about a day and then it worked. I began to suspect that the built in power supply was faulty and was going to change it but Vizio doesn't make parts available for that model so while it still worked I got the power output of that power supply (12v) and just removed that junk and put an external 12v, 5 amp power supply and it worked great for years after that until I upgraded to a newer Vizio soundbar.
Nice workaround 👍I did something similar to a radio someone brought into the repair cafe where I volunteer at. The internal transformer had failed that produced the 9V DC it used. I just found an old external 9V DC adapter and wired that to the board inside 🙂
Excellent work again.
Thank you! Cheers!
Very interesting video - learnt a lot!
Glad you enjoyed it!🙂👍
Nice to show how we all like to learn from doing.
Thanks 👍
hi,i`m giantkey,your video amazing,and it helps me to gain more knowledge to fix an item.
Thanks 👍glad it helped 🙂
you picked that "subwoofer" up way to easily, this seems like a very good quality product /sarcasm
😂😂😂Check out the "Fantastic" reviews it gets on amazon 😂
Makes you wonder if they're skipping out on proper QC testing somewhere with these things if enough get made with the same problem? Also you'd think for environmental testing they'd consider some people like the heat on rather than having the AC set to ice-box room temperatures, and they should use both ends of the range while also pushing the device to its limits. If not that, then somewhere in the pipeline people are signing off on defective stuff because of things like production delays. Wouldn't be too surprised if that happens too.
Could well be the case. Thanks for commenting 👍
It's everything I have several BlackVue dash cams I find 2-5 have issues that require replacement. QC is probably the majority of the issue with modern electronics. That build them so fast.
Could well be. Thanks for commenting 👍
Nice fix ! thanks for sharing👍
Thanks 👍and thanks for letting me know 🙂
Your videos make me glad I can convert C to F easily LOL
😂😂😂Yes we always use C over here. I've always got to look up the formula or use an online converter 😂😂😂👍
I expect the subscriber is going to want you to do this modification on his other soundbars so he can sell them on... Great work, and nice explanation of audio amplifier design. Class D amps are everywhere these days - it is how all smartphones can produce so much audio in a small form factor, efficiently.
Thanks 👍 He's quite capable of doing it himself as he repairs bluetooth speaker. This one he was just having trouble working out what the cause of the problem was 🙂👍
"it is how all smartphones can produce so much audio in a small form factor, efficiently"
Have to give credit to Digital Signal Processing too. Without a DSP, phones, TVs and even sound bars would sound much worse. The DSP lets manufacturers improve the sound of a speaker that's either constrained by space, size or price.
Your diagnostic skills are sound
😂😂😂Thanks 👍
I saw someone use an E-cigarette and fill it with rosin to create a white coating on the circuit bord and them powered up the bord to find the hot componnents that wet out the rosin vapor film. Neat trick when you don’t have a thermal camera.
Yes I bought one of those a while back too, although I've not used it yet. At a push you can also use IPA and see where it evaporates first when power is applied 👍
@@BuyitFixit nice trick, I’ll try it. I really like your videos there aren’t a lot of electronic diagnostic videos out there and you have a really simple methodology for fixing the problem.
Thanks 👍Much appreciated 🙂
🎉20K🎉 well deserved.
Thanks 👍
Can we talk about that irregular box shape? What a nightmare for shipping!
I think they tessalate like on the thumbnail picture, just like Tetris 😂😂😂😂
Excellent video, and superb diagnostic skills.. Just out of interest, could you put thermal pads on top of the inductors to bring the heat dissipation to a level that suited the schematic?.. thanks and all the best, kimbo
I think they just screwed up the design on the output, either that or a wrong value component was used on all of them. It was just easier to alter it to the chip manufacturers recommended design instead of trying to fix their mess.
@@BuyitFixit sounds good to me
Philips hasn't made any consumer stuff since 1999, as its sub contracted out under brand franchising.
2022 Philips completely sold its consumer compaines. Philips only does a military and medical client base now.
I suspect its a class N amp (digital ic amp), and using a trick to decrease audio bandwidth to achieve a louder audio output.
Other trick I seen them use 2ohm or even 1ohm deep throw speakers, on a 8 ohm nominal outputing amp, and they just use these hybrid crossover filters for control.
Creative Lab has a simular problem with the 1st and 2nd revision main pcbs on the flagship Gigaworks T40 series 1 and series 2.
Just like SMPSU's, the boards with N class (one chip simple design digital amps) ic amps tend to blow throu, as the fault stress of one going faulty component stresses out multiple compoments making them a different value, making the system to snowball sprial to another fault shortly after some time of a working fix. The T40 pcb are unfixable, thus why so many 2ndhand T40 's on sale on ebay.
In my opinion its these cheap psu plugs and smpsu chocoboxes that allow mains spikes throu, and so on the limit for wattage to save money that the psu output regulation wobbles and usually to a higher value for a moment taking or weakening these all in one amp ics.
These pcb designs also drives these amp ics hard, to get the maximum audio loudness out. So these ics are run hot with no heatsink and in a non vented enclosed box.
They are simply made to cheap.
Yes, cheap and not very good judging by the Amazon reviews..
Yep, all the AV ops were sold to Funai in 2013..
The inductor measurement probably seemed off due to the limitations of your tester. I'm guessing that it only goes down to 0.01mH otherwise it should have autoranged to µH at that point (I'm guessing it can't measure that low). And it probably isn't calibrated well enough so what should have shown as 0.01mH showed as 0.02mH. Does the tester have a specification sheet of its range capability?
Yes I was thinking the same, it's only a cheap tester which is why I assumed the inductors were probably ok, but it's the only tool I have that measures inductors currently.
😁"Better than Factory" (as Alex would say) ---- a lot better, as it now actually works. Could be a wrong value capacitor landing in the assembly line component carrier.
Thanks 👍Yes better than factory 😂😂😂😂
Sorry for my reaction i am a fan of this channel and still new to it. I know its not you but that board looks like they have used very cheep nasty ass components as well.
Yes, it does look cheaply made 👍
I used to work for Combined Electronic Services way back in 70’s ! C.E.S. Was the service side of Philips at that time! The factory was in Holland! C.E.S. Covered all the Philips and Pye equipment. TV, audio, comes, were all covered ! I worked on audio, radio’s etc! They had their finger in almost everything like washing machines, microwave s and they had a huge market in office equipment! Dictating machines , language labs, voice recognition etc ! They took out the micro cassette for dictating m/c’s and pocket recorders ( tiny gears , pulleys, etc,,.) that lawyers, police, business, people used! I think you can still get them but all digital now! Many people carried that pocket m/c’s on their hip pocket so many ended ended in the toilet!😂 Fraser
😂😂Thanks for sharing 👍
You should build your own products mate atleast people could rely on its good quality build and performance if was designed by you as someone who loves fixing faulty electronics myself i really enjoy your content keep the videos coming mate amazing work and logic 👍
Thanks 👍I'm no expert. I haven't really designed much apart from an internet connected incubator for hens eggs a while back, and a couple of arduino button bashing games and a shooting gallery type game.
Genius, great work.
Thank you very much👍
Are you at MyMateVince's house? you've got his carpet! 🤣
😂😂😂😂 No lol. I've not noticed that, you've sparked my curiosity now 👍
Haha, you're not an expert you say in the description. I know many "experts" with much less skills. Great approach to get to a possible root cause. And great to experience you can think out of the box (a design fault is not a common path to follow)
Thanks 👍🙂
The filter has to dissipate everything above 20.000 Hz. With a 48kHz carrier pulse width modulated signal that's a lot of energy that has to be dissipated by the filter coil. Of course it heats up. That's the whole idea.
I thought one of the benefits of the class D was lower energy consumption. Seems awfully lossy to just dissipate it as heat..
Great stuff as always 😎
Thanks mate 👍
That was very impressive
Thanks 👍🙂
Nicely done!!
Thanks 👍
Nice shot !
But in the end, I don't understand what's the fix exactly ?
Rewiring the circuit to match the chip manufacturers recommended circuit, instead of the output circuit that Philips had used.
@@BuyitFixit Thank you, that's what I think I understood. But what did you do exactly ? I'm a little bit confused with that part.
I didn't buy any Philips labeled stuff since over 20 years by now. The reason for it is that I remarked how much of their stuff does just survive the warranty.
Just started watching, good content interesting. Can i ask you put a link to your phone usb camera , yiu used for the heat detection. Is it android? Regards
Thanks, yes it's android. There's a link in the video description. They do an iphone version too.
I wonder if Phillips and Polk sound bars are made in the same facility? I have a Polk with Alexa that came in the exact same box. The remote even looks the same.
Could well be. Sounds like it from your description 👍
That thermal camera is very useful i think i need one lol
Yes they are very handy! I put some links in the description of the tools that I use and also did a video review on that camera VS the FLIR thermal camera that I repaired in a previous video.
I remember fixing several Phillips portable radios in the '70's, the volume control was mounted on the edge of the circuit board and the corner of the board and control used to snap off. Awful design even in them days.
Interesting, Thanks for commenting 👍
I wonder why they even attempted to create their own filter. It usually is no bad idea to stay close to what the chip manufacturer suggests.
Yes agree 👍 perhaps they were trying to get more power? Or improve the audio quality?
My guess is that they were either finding problems with EMI/RFI interfering with the wireless stuff in the box or they were failing to meet CE or FCC or other standards for emissions.
2:20 bro my dearest blessings how long were you having to blast this full vol omg
😂😂😂😂😂👍
The values of the inductor and capacitors in the filter on the datasheet are different then what you measure on the board because the datasheet schematic is using an 8 ohm speaker and the speaker bar is using 4 ohm speakers. The excessive heat might be due to the inductors and resistors in the filter circuit being under sized. As the inductors heat up their current carrying capacity drops, their inductance value changes and being in a small enclosed space probably forces a thermal runaway that eventually damages the inductors and resistors.
Also, in the datasheet there is a ground in the filter design but there doesn't have to be. The filter can be designed just between the two outputs of the H-bridge.
Edit: But you need to use higher voltage caps and higher watt resistors as the caps are seeing a higher AC voltage. The filters are difficult to get right and those original capacitors in the original configuration probably didn't have a high enough AC voltage rating to be in that particular filter, they looked really undersized for that even at ~20-40w because of the switching frequency of the amp.
Interesting. On the datasheet it shows the 4 ohm filter just under the 8 ohm one. However both show the same value. I did download a tool from TI to work out the LC filter and it said that a 1.2uF would be better. A 1uF gave a slighty higher cut off frequency. The calculations for all of it were way over my head 😂 luckily they had that spreadsheet that worked it all out for you.
@@BuyitFixit Yeah, I got that spreadsheet too. Class D LC Filter design tool. The "differential mode" design is usually used for BTL/PBTL modes because it can be cheaper with fewer components but it can be dangerous if you don't choose the right components. The capacitor in the middle can look like a dead short at higher frequencies. The "hybrid mode" is pointless to me as it is just as dangerous as the "differential mode" version and doesn't give much more benefits. With the "common mode" filter design is best as the capacitors are not stressed as much and don't need 2 to 3 times the AC voltage rating of the output.
An inductor running at 50 °C is perfectly OK unless the core material is something very strange with an exceptionally low Curie temperature. Any change in inductance due to temperature would be of no consequence in a circuit like this.
@@d614gakadoug9 Yes, the main thing is the AC rating of the capacitors at the output frequency of the class D amp. The filter needs to filter out the PWM frequency and using cheaper capacitors with a series resistor does not help the capacitor handle the AC voltage or frequency of the circuit. Might help with DC voltage but not really AC voltage. The capacitor can still look like a short circuit at the higher frequencies so its a bad idea. The next problem is the current capacity of the inductors but really only when the capacitors are underrated. I've had to replace inductors before because underrated capacitors caused excessive current and heating in the inductors.
There was a Corsair 2.1 speaker set that was bi-amped stereo + a subwoofer, three chips total. Two chips in BTL stereo mode and one in PBTL for subwoofer. Hi+Mid amp for left and another Hi+Mid amp for left. It was notorious for power supply failures but the main cause was the design of the Class-D output filters (aside from other issues with the heatsink). They chose to use underrated caps in the output and put resistors inline thinking it would help. The burning of the adhesive they threw around the output filters showed that it was literally burning itself to death at regular output levels. The filter components should never be getting that hot. The solution was to redesign the output filters with new caps and also new inductors for the subwoofer section as there was usually at least one bad inductor in the subwoofer section. The subwoofer section was PBTL mode using a hybrid filter that meant a capacitor directly across the + and - outputs and, of course, it was not rated with a high enough AC voltage for that circuit and would explode or meltdown and then kill an inductor or two. The inductors would have been fine had they chosen a capacitor with the correct AC voltage rating for the hybrid mode design. And even TI doesn't have any design notes for us on how to choose a proper rating for those filter capacitors. :/
I forget which exact capacitor I went with for the hybrid filter center cap in the subwoofer section, but I remember the one that didn't heat up or cause the inductors to heat in testing was a film cap rated something like DC 150v while the power supply for the amp was only DC 21-24v output. That is how I discovered the AC voltage rating was related to the amount of heating I observed. Test two different brand caps with the same DC voltage rating but with different AC voltage ratings, the one with the lower AC voltage specs would always heat up more.
put a scope on the filter network to see if its suffering from a parasitic oscillation, you may be able to add extra LCR that would prevent it.
Looks like they were attempting to use the more complex filter to filter out the remaining switching noise but got the layout wrong. They obviously didn't do any long term testing before sending them out the door.
Yes, I think that too 👍
Or a batch was made with the wrong component values or flakey components. Post-assembly inspection of boards with surface mount parts can be a pain because some components, notably almost all ceramic capacitors and many resistors have no markings of any sort. SM ceramic caps have "never" been marked. Resistors larger than 0402 used to be, but lots of manufactures quit marking them years ago.
I've seen quite a lot of stuff where there was evidence of good engineering that has been overridden by some other consideration. In most cases the override appeared to be motivated by the desire to save a few pennies.
I've had idiots who thought they knew better and didn't follow instructions make a complete mess of building products I've designed. One was doing fine in the hands of one company. Its parent company acquired a competitor and decided the manufacturing should be done by the competitor company because they did more manufacturing. But they were too stupid to follow some instructions for a method that had been carefully worked and made a mess of things. It killed the product.
I had another design for which a small batch of prototypes were to be built. My client had picked an enclosure they wanted to use and it presented some challenges. I carefully worked out a method for handling the challenges. I wrote it all up in substantial detail. The job of actually doing the work was handed off to someone known to my client but not to me. The fool complained that I must not be used to working with professionals, then botched the job because he wouldn't let anyone tell HIM how to do anything.
I had a client who manufactured a product of my design in-house. It contained several capacitors that were critical to performance and longevity. Some rep from some component vendor offered them something he claimed was a direct equivalent. They were CRAPacitors.
I don't do any of this stuff anymore and I'm glad to be out of it.
your guy must be some kind of magician - changing the Ic without disturbing the silicon around it - just saying
It was a different sound bar he tried changing the IC on but it didn't make any difference. He sent me a fresh one that hadn't been messed about with 😉
It seems to be a common occurrence of products getting past r&d with stupid issues.
At my work we had to swap out 600 Ethernet switches due to two issues. A batch of bad diodes, and a chip burning up due to insufficient heat sinking. Major pain in the butt trying to get down time to swap all of these out. Someone in engineering with a thermal camera should have spotted this right away. I can't blame the manufacturer for a bad batch of diodes. That can happen from any supplier. Not taking a thermal shot of the board is inexcusable. 🤬
it doesn't sound fun having to swap out 600 network switches.
@@BuyitFixit it took a year and each switch had at least 50 ports and probably three quarters of them 100 ports. Loved the density of the 100 port switches. Unfortunately our supplier no longer makes them. 24 or 50 port switches only. 😡. Thank you Extreme Networks
Well done Sherlock, another awesome bit of detective work. I guess we need to look out for a flurry of Phillips Soundbars hitting eBay over Christmas now he knows how to fix them 😉
Thanks Marc, I think he bought these a while ago, but he got stuck on diagnosing the problem.
Well done great stuff
Thanks 👍
what kind of thermal camera are you using? It looks like it's got better resolution than the Fluke you repaired.
It has! 256x192 and 30fps I think it was. It's the P2 pro. There's a link in the video description along with some of the other tools I use 👍
I'm not too impressed with most of Philips' later tech. Although on TV they're still great. Bought a cheap soundbar, B5105, without sub for use in bedroom. Well.. Yes it is better than the TV speaker but even for the cheap price it's not really a good buy. I then gave a cheap Sony soundbar to my sisters kids, an HT-SF150, and that is worth every penny. Doesn't have deep bass but it has what you reasonably expect these days and great tonal balance - doesn't sound thin like the Philips. While it only has Dolby 5.1 (no Atmos) it is very convincing when you sit right and every streaming box has capability to down convert anyway. Tt was just pennies more in contrast to the Philips yet so much better.
I've got a Sony TV and sound bar and the sound is fairly good. The Philips one isn't that good at all. There's a number of reviews on Amazon that pretty much say that the TV speakers sounded better and it was more of a downgrade 😂
Amazing. Thank you!
👍🙂
Wow very nice video you are professional engineer master of the master 👏 I have philips pronto tsu9600 touch screen remote control can you please kindly do a video of it as its not working ie not powering up it was working before I changed new 18650 battery full changed still it doesn't switching on? If make a video of this it will be the best video on TH-cam please kindly help me on this issue 🙏 waiting for your reply Thankyou ps how can I send this to you for repair 🙏please
Thanks. There's an email address on the channel "about" page under "for business enquiries"
Someone had a container full of a single product all with the same fault. That's not janky at all. They already replaced it with the B5306 which works well.
It's a shame they didn't simply update the board and replace the faulty ones instead of designing and manufacturing an entirely new model while these were packed up as e-waste.
Well hopefully they learn from their mistakes...
Brilliant. Thanks
You're welcome!
There is something strange afoot there.
The over-temperature shutdown threshold for the amplifier IC is nominally 150 °C (die temperature). The top of the package will be cooler than the die, but not by a huge amount. I'd expect the top of the package to be in the range of 100 °C at shutdown. Since no testing to shutdown was done with board exposed we have no useful data.
My suspicion is that shutdown was occurring due to something in the filter network causing short circuiting when hot, which also causes shutdown.
I see design errors like this very often. I always ask myself why a chip manufacturer specifies certain component values and their arrangement and wiring in their data sheet if something different is installed in the end.
The probability that a chip will no longer work within the manufacturer's specifications due to different component values and wiring is very high.
Is this planned obsolescence, or simply a design flaw?
Are such devices no longer tested extensively as prototypes before they are mass-produced and sold?
Yes it is odd. I guess they were trying to improve the sound quality or get more out of the chip perhaps? Thanks for commenting and have a great Christmas 👍