Guest richardmlea Posted August 13, 2011 Report Share Posted August 13, 2011 (edited) Hi all, I have just run some voltage tests on the speakers to see if they are up to the job and it doesn't look like they are. I opened the speaker cases to have a look at the back (nearly sliced off the end off my thumb with a Stanley blade opening one of them) Uploaded with ImageShack.us Remember! Always ask an adult for help if using scissors and knives, "we bleed so you don't have to" The back of the speaker says 4 ohms, 1.5v and the part number (forgot to take photo) After I stopped bleading I tested the speaker output. I used a synthesizer app to output a constant tone and YouTube app for tests. On corvus5 at full volume I got readings of 1.7v for the right speaker and 1.8v from the left (tested from the board with the speakers disconnected). So the voltage is unbalanced and over the rated voltage of the speakers. I also noticed the youtube app gave slightly higher readings than the synthesizer so I have quoted the reading from YouTube. On vegacomb test build 44 ( again at full volume) I got readings of 1.8v for the right speaker and a tiny bit under 2 volts for the left. I then tested at muted and at 0 Volume (here's where it gets really weird), I got 2v on the right and 2.2v on the left. ?????? As I turn the volume up the voltages gradually decrease until about half volume when it gets to 1.8 and 2 where it stays (roughly) until it gets to full volume. Strange! My meter is analog so the reading are accurate to .1V but this is accurratte enough for these tests. So to summaries, the voltage output is unbalanced and too high for the speakers. This is why the sound is so puny on our Vega's and why the increased volume on test 44 caused some speakers to blow. The volume has been limited by the manufacturer to prevent the under-rated speakers from blowing. Given a better pair of speakers the Vega's amp is capable of much higher volumes (as build 44 proves). Also muting the volume on Vegacomb test 44 probably wont save your speaker, if its going to burn out it would have done already (don't take this as Gospel, its just educated guess from what I have observed on mine). I am going to buy a big pile of used laptop speakers off ebay (£5 for about 20 assorted speakers) to see if I get lucky and find some better ones. I still love the Vega, its little flaws just make it more endearing to me. If everything worked properly I'd be bored. Edited August 13, 2011 by richardmlea Link to comment Share on other sites More sharing options...
Guest lesd Posted August 13, 2011 Report Share Posted August 13, 2011 Hi all, I have just run some voltage tests on the speakers to see if they are up to the job and it doesn't look like they are. I opened the speaker cases to have a look at the back (nearly sliced off the end off my thumb with a Stanley blade opening one of them) Uploaded with ImageShack.us Remember! Always ask an adult for help if using scissors and knives, "we bleed so you don't have to" The back of the speaker says 4 ohms, 1.5v and the part number (forgot to take photo) After I stopped bleading I tested the speaker output. I used a synthesizer app to output a constant tone and YouTube app for tests. On corvus5 at full volume I got readings of 1.7v for the right speaker and 1.8v from the left (tested from the board with the speakers disconnected). So the voltage is unbalanced and over the rated voltage of the speakers. I also noticed the youtube app gave slightly higher readings than the synthesizer so I have quoted the reading from YouTube. On vegacomb test build 44 ( again at full volume) I got readings of 1.8v for the right speaker and a tiny bit under 2 volts for the left. I then tested at muted and at 0 Volume (here's where it gets really weird), I got 2v on the right and 2.2v on the left. ?????? As I turn the volume up the voltages gradually decrease until about half volume when it gets to 1.8 and 2 where it stays (roughly) until it gets to full volume. Strange! My meter is analog so the reading are accurate to .1V but this is accurratte enough for these tests. So to summaries, the voltage output is unbalanced and too high for the speakers. This is why the sound is so puny on our Vega's and why the increased volume on test 44 caused some speakers to blow. The volume has been limited by the manufacturer to prevent the under-rated speakers from blowing. Given a better pair of speakers the Vega's amp is capable of much higher volumes (as build 44 proves). Also muting the volume on Vegacomb test 44 probably wont save your speaker, if its going to burn out it would have done already (don't take this as Gospel, its just educated guess from what I have observed on mine). I am going to buy a big pile of used laptop speakers off ebay (£5 for about 20 assorted speakers) to see if I get lucky and find some better ones. I still love the Vega, its little flaws just make it more endearing to me. If everything worked properly I'd be bored. Speaker impedence is normally quoted ar 1KHz. Your analogue meter can't measure much above 100Hz. When you measure the speaker voltage, are you measuring WRT battery negative? You would be measuring the DC bias of the amplifier on that case. You have to measure across the two speaker terminals, preferably with a scope. Link to comment Share on other sites More sharing options...
Guest richardmlea Posted August 13, 2011 Report Share Posted August 13, 2011 (edited) Speaker impedence is normally quoted ar 1KHz. Your analogue meter can't measure much above 100Hz. When you measure the speaker voltage, are you measuring WRT battery negative? You would be measuring the DC bias of the amplifier on that case. You have to measure across the two speaker terminals, preferably with a scope. Do these reading mean nothing then or does the fact that the voltage is higher than the speakers rating. Does the voltage increases on vegacomb test 44 not suggest that speakers are getting to much power. The speakers blowing being the result and ultimate proof that they are getting to much voltage, the speaker coil acting like a heating element? Edit. wait a minute, I am not measuring the speaker impedance. The speaker aren't connected Surly speakers are rated in ohms Edited August 13, 2011 by richardmlea Link to comment Share on other sites More sharing options...
Guest lesd Posted August 13, 2011 Report Share Posted August 13, 2011 (edited) Do these reading mean nothing then or does the fact that the voltage is higher than the speakers rating. Does the voltage increases on vegacomb test 44 not suggest that speakers are getting to much power. The speakers blowing being the result and ultimate proof that they are getting to much voltage, the speaker coil acting like a heating element? Edit. wait a minute, I am not measuring the speaker impedance. The speaker aren't connected Surly speakers are rated in ohms Yes. speaker impedence is rated in Ohms, in this case 4 ohms at 1KHz. At 2KHz it would be 8ohms, and if you measure the speaker with an ohmmeter it is zero. Speakers are also normally rated for power, not voltage. You can work out one from the other using the formula POWER = V^2 / R I don't think you're able to get any sensible readings of audio voltage using a standard analogue meter. So the only comment I can make about the different voltages with Vega comb is that the audio driver software is different. It may be that it doesn't match the audio driver circuit and is, as you suggest, heating the speakers up. EDIT. If you can measure a DC voltage across the speaker terminals with the speaker disconnected, then the driver software is at fault. Edited August 13, 2011 by lesd Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 13, 2011 Report Share Posted August 13, 2011 (edited) Here ya go! A little light reading about resonant circuits and the effects of inductive and capacative reactance in a tuned circuit: Wikipedia: LC Circuits Something to help you sleep at night dude. ;) Actually, that was more to do with what Les was saying about frequency. Unfortunately, speakers can burn out because an amplifier is UNDER powered. What happens is, the amplifier is over-driven causing clipping. Under such conditions pure dc is applied to the speech coils causing massive heat increase because of the resistive element. The rest is unfortunately new speakers! Edited August 13, 2011 by TimboDavis Link to comment Share on other sites More sharing options...
Guest jesterzoo Posted August 13, 2011 Report Share Posted August 13, 2011 You guys make me feel dumb :blink: Link to comment Share on other sites More sharing options...
Guest lesd Posted August 13, 2011 Report Share Posted August 13, 2011 Here ya go! A little light reading about resonant circuits and the effects of inductive and capacative reactance in a tuned circuit: Wikipedia: LC Circuits Something to help you sleep at night dude. ;) Actually, that was more to do with what Les was saying about frequency. Unfortunately, speakers can burn out because an amplifier is UNDER powered. What happens is, the amplifier is over-driven causing clipping. Under such conditions pure dc is applied to the speech coils causing massive heat increase because of the resistive element. The rest is unfortunately new speakers! Ah! The LC tuned circuit... I remember it well. (Nostalgic sigh) I was going to recommend the Wikipedia article on inductors, but really would be a cure for insomnia. Anyway, it's not actually possible for me to say what Richard is measuring because of all the different types of amplifier available. My guess is a C class H bridge, and that Richard is measuring the average voltage on one side of the bridge. If there is an audio signal there, this should be half the voltage supply to the amplifier. If I was to make a guess at the problem, I'd say the Vegacomb audio driver software is putting the wrong bias on the audio signal, causing clipping. Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 13, 2011 Report Share Posted August 13, 2011 (edited) If I was to make a guess at the problem, I'd say the Vegacomb audio driver software is putting the wrong bias on the audio signal, causing clipping. I agree, it seems to fit the bill here for me. Nostalgia for me too Les. I don't remember too much about the amps tbh. Only the A,B, AB and C class amps but not in detail. If you asked me to calculate the values for the bias networks I'd be well screwed! LoL Are you still in the electronics game Les?? Edited August 13, 2011 by TimboDavis Link to comment Share on other sites More sharing options...
Guest lesd Posted August 13, 2011 Report Share Posted August 13, 2011 I agree, it seems to fit the bill here for me. Nostalgia for me too Les. I don't remember too much about the amps tbh. Only the A,B, AB and C class amps but not in detail. If you asked me to calculate the values for the bias networks I'd be well screwed! LoL Are you still in the electronics game Les?? I teach it at FE level. Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 13, 2011 Report Share Posted August 13, 2011 I teach it at FE level. That's good! It never occurred to me that circuit theory was still taught!! I always assumed that these days you'd learn a more modular approach - in other words, card and box changing. Link to comment Share on other sites More sharing options...
Guest richardmlea Posted August 13, 2011 Report Share Posted August 13, 2011 Thanks for your analysis guys. I bow to your superior knowledge. Thanks for the reading timbo I will have a look now. What are your thoughts about replacing speakers and improving the vegas sound/volume in general. Test 44 didn't blow all the Vega speakers it came in contact with so surly some Vegas are capable of better sound? Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 14, 2011 Report Share Posted August 14, 2011 Thanks for your analysis guys. I bow to your superior knowledge. Thanks for the reading timbo I will have a look now. What are your thoughts about replacing speakers and improving the vegas sound/volume in general. Test 44 didn't blow all the Vega speakers it came in contact with so surly some Vegas are capable of better sound? Well, it's not superior knowledge on my part Richie. In fact, more of memories from 30 years ago! Funnily enough, I don't think replacing the speakers on their own would help. If the amplifier stages are overdriving to cause clipping then, as Les says, the bias voltage on the amplifier has to be changed to enable the signal equal displacement either side of the carrier. It still wouldn't prevent clipping if the signal level exceeds the cut-off level. That's determined by the circuit design itself. All this ONLY if my memory serves! Please correct me if I've got it all wrong Les! For me, the solution would be to use a tiny amp and install it between the existing amplifier and the audio output. Then there would also be something to gain from replacing the speakers. Now that's just my thoughts and I have no idea if it would work tbh Richie. The man to ask if and how it can be done would be Les. He's probably forgotten more than I was ever taught! ;) By the way, I take my hat off to you for looking in to this bud! It's people like yourself, opening up discussions (AND the Vega!), that make modaco what it is! A great place to learn from everybody elses hard efforts! Reschpekt dude! B) Link to comment Share on other sites More sharing options...
Guest lesd Posted August 14, 2011 Report Share Posted August 14, 2011 (edited) Well, it's not superior knowledge on my part Richie. In fact, more of memories from 30 years ago! Funnily enough, I don't think replacing the speakers on their own would help. If the amplifier stages are overdriving to cause clipping then, as Les says, the bias voltage on the amplifier has to be changed to enable the signal equal displacement either side of the carrier. It still wouldn't prevent clipping if the signal level exceeds the cut-off level. That's determined by the circuit design itself. All this ONLY if my memory serves! Please correct me if I've got it all wrong Les! For me, the solution would be to use a tiny amp and install it between the existing amplifier and the audio output. Then there would also be something to gain from replacing the speakers. Now that's just my thoughts and I have no idea if it would work tbh Richie. The man to ask if and how it can be done would be Les. He's probably forgotten more than I was ever taught! ;) By the way, I take my hat off to you for looking in to this bud! It's people like yourself, opening up discussions (AND the Vega!), that make modaco what it is! A great place to learn from everybody elses hard efforts! Reschpekt dude! B) Yeah, we still teach circuit theory. LC tuned circuits are a minor part of it now, because we tend to use active filters. We also teach our students to get a good night sleep ;) You are correct, replacing the speakers alone will not work if the problem is clipping. But I was confused about where this fault crops up. If the problem is only with test 44 software only, then the solution is probably to backoff whatever was done to cause the fault. One software error that would cause clipping is to forget to allow for the bias offset before and after scaling. For example, in a CD system, the top, bottom and crossover of a maximum volume sine wave are converted to 5V, 0V and 2.5V rspectively. Digitally, those points are represented by 2^16 -1, 0 and 2^15. A simple integer multiply to increase the volume would cause the zero point to shift. Instead you need to convert to standard integer representation (by adding 2^15) do the multiply, and convert back. Edited August 14, 2011 by lesd Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 14, 2011 Report Share Posted August 14, 2011 .... We also teach our students to get a good night sleep ;) LoL! Well, I blame the teachers. If they didn't impart the knowledge so damned well then people like me wouldn't have to work night shift! :P The maths is something that I have long forgotten Les but if I get time maybe I'll even go back through my old stuff and try to re-learn it all. I remember [fairly vaguely] changing the amp bias to vary the gradient in order to reduce clipping but that's really the limit of my memory. My suggestion about an additional amplifier stage was really an idea to improve the gain overall as opposed to the immediate problem of burning out speakers. Sorry about that. I lost sight of the actual intention of Richie's investigation in the first place. My apologies Rich! Thanks again Les! I think I've got a bit of reading-up to do! Link to comment Share on other sites More sharing options...
Guest richardmlea Posted August 15, 2011 Report Share Posted August 15, 2011 (edited) This has become a very inertesting, timbos idea of a seperate amp got my mind racing and I found this on ebay DIY audio for your MP3, MP4, Laptop or any application which USB power can support. Power Supply: DC 3.6v-5.5V, Output power: 3W +3 W SNR: 90dB Efficiency: >90% Size：24x15.5mm Hole Size: Pitch 20mm, Bore: 2mm What do you think? would it be ok wiring it to the headphone jack? What would be the advantage of wiring the new amp to the mainboard before the on-board amp? If its wired to the headphone jack the volume will be controlled by the vega. Its £4.30 inc. p+p i may buy 2 or 3 and have the loudest tablet in the world. :lol: Edited August 15, 2011 by richardmlea Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 15, 2011 Report Share Posted August 15, 2011 What do you think? would it be ok wiring it to the headphone jack? What would be the advantage of wiring the new amp to the mainboard before the on-board amp? If its wired to the headphone jack the volume will be controlled by the vega. Its £4.30 inc. p+p i may buy 2 or 3 and have the loudest tablet in the world. :lol: LoL! That's the kind of thing I had pictured in my mind. Would you have to limit the existing output stage though, now that it would effectively become an intermediate stage? In other words, the effect of keeping the same gain might lead to the NEW output stage being overdriven. Link to comment Share on other sites More sharing options...
Guest xathras Posted August 15, 2011 Report Share Posted August 15, 2011 Its £4.30 inc. p+p i may buy 2 or 3 and have the loudest tablet in the world. :lol: Not for long - thinking blown speakers here! Forgive me If I'm missing something, but I've never had a laptop, netbook or tablet where the sound from the built-in speakers has been anything other than awful. They are simply too small. I use a pair of X-miniIIs and a splitter cable for my netbook and tablets, and they sound great (for their size - fit in a pocket). Even so, they are the lowest quality of stereo I'm prepared to accept, and they have 40mm drivers. Link to comment Share on other sites More sharing options...
Guest richardmlea Posted August 16, 2011 Report Share Posted August 16, 2011 LoL! That's the kind of thing I had pictured in my mind. Would you have to limit the existing output stage though, now that it would effectively become an intermediate stage? In other words, the effect of keeping the same gain might lead to the NEW output stage being overdriven. How would I go about attenuating the output? Resistors? Link to comment Share on other sites More sharing options...
Guest ejtagle Posted August 16, 2011 Report Share Posted August 16, 2011 Hi! It would be fair to tell, that, as the one who wrote the audio codec driver for the Vegacomb .36 version, i am also worried about this speaker blowing issues... As i have the equipment to do measures, i did a full test of the speaker amplifier in the audio codec. Using an analog meter, even using a digital meter, you wont get meaningful readings... The cause ? -- The power amplifier inside vega (and, just to mention, the power amplifier is integrated into the audio codec ALC5624) is a class D amplifier... Yes, it is a switching amplifier, a pulse width modulation (PWM) amplifier.. The way this works is strange, but much more efficient that traditional class AB amplifiers. Instead of sending an analog voltage, the amplifier sends a high speed square wave, whose duty cycle encodes the analog voltage to output. So, for example, the ALC class D amplifier sends an signal whose period is fixed (about 100khz), and that has 2 possible voltages: 0 and 5v. The amount of time that 5 v is output relative to the total period time gives the duty ratio. And, as the switching frecuency (the inverse of the period) is too high, the speaker reacts to the medium value, demodulating the encoded audio signal, finally outputting sound The problem is that that square wave fools the analog meter, even the digital meter, so you wont get valid readings from them. What i did was to hook a digital oscilloscope across the speaker terminals, to check for continuous voltage across the speaker terminals, and also to try to figure out the peak to peak voltage reaching the speakers. The good news is that there is not DC across the speaker terminals. But, if you put the volume to the maximum, about 10 v peak to peak could reach the speaker termnals (yes, as the speaker is tied between 2 amplifier outputs, that is the maximum voltage that can reach the speaker). This is way too much for those small speakers. The original .32 android2.2 limited the voltage reaching the speakers (according to my measurements) to about half that (5v peak to peak). I do even consider that too much for those speakers.... So, it is not a clipping issue. It is just overpowered speakers :( I hope to limit the power delivered to the speakers to the same level used by the 2.2 android kernel ... Even if i consider it too much. Eduardo Link to comment Share on other sites More sharing options...
Guest lesd Posted August 16, 2011 Report Share Posted August 16, 2011 (edited) Hi! Using an analog meter, even using a digital meter, you wont get meaningful readings... The cause ? -- The power amplifier inside vega (and, just to mention, the power amplifier is integrated into the audio codec ALC5624) is a class D amplifier... Yes, it is a switching amplifier, a pulse width modulation (PWM) amplifier.. The way this works is strange, but much more efficient that traditional class AB amplifiers. Instead of sending an analog voltage, the amplifier sends a high speed square wave, whose duty cycle encodes the analog voltage to output. So, for example, the ALC class D amplifier sends an signal whose period is fixed (about 100khz), and that has 2 possible voltages: 0 and 5v. The amount of time that 5 v is output relative to the total period time gives the duty ratio. And, as the switching frecuency (the inverse of the period) is too high, the speaker reacts to the medium value, demodulating the encoded audio signal, finally outputting sound The problem is that that square wave fools the analog meter, even the digital meter, so you wont get valid readings from them. What i did was to hook a digital oscilloscope across the speaker terminals, to check for continuous voltage across the speaker terminals, and also to try to figure out the peak to peak voltage reaching the speakers. The good news is that there is not DC across the speaker terminals. But, if you put the volume to the maximum, about 10 v peak to peak could reach the speaker termnals (yes, as the speaker is tied between 2 amplifier outputs, that is the maximum voltage that can reach the speaker). This is way too much for those small speakers. The original .32 android2.2 limited the voltage reaching the speakers (according to my measurements) to about half that (5v peak to peak). I do even consider that too much for those speakers.... So, it is not a clipping issue. It is just overpowered speakers :( I hope to limit the power delivered to the speakers to the same level used by the 2.2 android kernel ... Even if i consider it too much. Eduardo Hi Eduardo, It looks like you've solved this problem. I assume you are controlling volume using the CODEC hardware, and will simply reduce the upper limit by 3 or 6dB. The problem arises from the choice of speakers, which are 4ohms while the ALC5624 is designed for 8ohms. Therefore, the speaker take twice as much power as expected. Just to clarify, 5V peak to peak (1.77Vrms) is about what Richard reported as the rated voltage of the speakers. Also, for reference in future testing, the 100KHz square wave will always be 10Vpeak to peak, if the supply (VDDSPK) is 5V. This isn't a problem at 100KHz as the speaker impedence rises to 400 ohms at that frequency. You will only see a change in the maximum mark-space ratio. SECOND EDIT for anybody measuring the amplifier with a scope (CRO). The 10V square wave across the speaker is volltage only. It happens because of the inductive nature of the speaker, and the commutation diodes in the amplifier. To be clear, each of the four transistors in the H bridge has a schottkey diode between source and drain, facing the positive supply. When the transistors delivering the squarewave swirch off, the commutation diodes take over the current. If the top left and bottom right transistors were in use, then the bottom left and top right commutation diodes take over the current. The left end of the speaker is now clamped to 0V and the right end clamped to 5V. The polarity across the speaker is reversed, so a scope across the speaker terminals sees a 10V waveform. A scope between one terminal and GND would show the amplifier output switching between 0.3V and 5V when the top transistor of that output is operating, and between 0V and 4.7V when the top transistor of the other output is operating. @Also, while a digital meter will not reliably measure the DC component of a PWM output, an analogue meter will. Richard reports measurements of 1.8V to 2.2V and that is exactly what you would expect measuring between one amplifier terminal and GND. Measuring between the terminals you'd get 0V. Finally, many thanks for your work on Vegacomb. We are all very excited about the next release :) EDIT @TimboDavies I reckon you'll get some satisfaction from understanding this explanation. The audio amplifier is called an H bridge because it is arranged like a letter H, with the speaker forming the bridge of the H and each leg being a class B output stage, connected (in this case) to a 5V supply at the top, and 0V at the bottom. If you turn the top left and bottom right transistors on, you connect 5V from left to right across the speaker. Conversely, if you turn on the top right and bottom left transistors, you connect 5V from right to left across the speaker. If you take one speaker terminal as a reference point, the peak to peak voltage at the other terminal is 10V. Hence you can deliver four times the power that a class A amplifier can. The amplifier is class D, which means each transistor is operated as a switch, not in a linear mode. To simulate a sine wave, a series of variable width pulses switch the top left and bottom right transistors ON and OFF. The width of each pulse is proportionate to the instantaneous value of the sine wave at that point (Pulse Width Modulation). The same sequence is then applied to the other two transistors. The amplier has 'flywheel' diodes built in, so the inductor that is the speaker is able to smooth the pulsed current into a sinewave. Edited August 16, 2011 by lesd Link to comment Share on other sites More sharing options...
Guest lesd Posted August 16, 2011 Report Share Posted August 16, 2011 I've edited my previous post twice since I originally posted it, so I'm bumbing it in case any of the additional info is important to anyone, especially Eduardo. Link to comment Share on other sites More sharing options...
Guest ejtagle Posted August 16, 2011 Report Share Posted August 16, 2011 Hi Eduardo, It looks like you've solved this problem. I assume you are controlling volume using the CODEC hardware, and will simply reduce the upper limit by 3 or 6dB. The problem arises from the choice of speakers, which are 4ohms while the ALC5624 is designed for 8ohms. Therefore, the speaker take twice as much power as expected. Just to clarify, 5V peak to peak (1.77Vrms) is about what Richard reported as the rated voltage of the speakers. Also, for reference in future testing, the 100KHz square wave will always be 10Vpeak to peak, if the supply (VDDSPK) is 5V. This isn't a problem at 100KHz as the speaker impedence rises to 400 ohms at that frequency. You will only see a change in the maximum mark-space ratio. ... You are absolutely right. What i was talking about, is that under severe clipping, the amplifier can outout a square wave of +/- 5v... 10 v peak to peak. or 3.5v rms. Ahh.. class D was a passion of mine, i did several power amplifier designs a few years ago. Most of them were posted to the internet (for example, http://www.forosdeelectronica.com/f31/amplificador-hi-fi-200w-rms-dos-mosfet-16229/) .. And,also, now you know a little bit more of me.. i am an electronic engineer, rather than a software engineer ;) - but due to the world we live in, i do much more software than hardware designs, even if hw design is one of my passions ;) Regards, Eduardo Link to comment Share on other sites More sharing options...
Guest lesd Posted August 16, 2011 Report Share Posted August 16, 2011 You are absolutely right. What i was talking about, is that under severe clipping, the amplifier can outout a square wave of +/- 5v... 10 v peak to peak. or 3.5v rms. Ahh.. class D was a passion of mine, i did several power amplifier designs a few years ago. Most of them were posted to the internet (for example, http://www.forosdeelectronica.com/f31/amplificador-hi-fi-200w-rms-dos-mosfet-16229/) .. And,also, now you know a little bit more of me.. i am an electronic engineer, rather than a software engineer ;) - but due to the world we live in, i do much more software than hardware designs, even if hw design is one of my passions ;) Regards, Eduardo Well, if you've designed several D class amps, you dont need me to tell you what you're looking at on the scope. My apologies. I'll have a look at those designs later, but I note some of them are fairly high powered. Motor drive? I'm a motion control engineer by trade. I'm teaching now, but once we get past the basics, we have to teach software. I blame NASA. Link to comment Share on other sites More sharing options...
Guest TimboDavis Posted August 16, 2011 Report Share Posted August 16, 2011 (edited) WoW! Now there are 3 replies here: @Richie. Dude, you pretty much need to ignore everything I've said from the looks of it. My explanations don't fit this scenario at all and are largely irrelevant to be honest. Sorry about that! In answer to your question though, yes, in general resistors are used in a potential divider network to provide the bias on any transistor amplifier. The choice of values is part of the design stage. In addition to this, each amplifier output is matched to the input stage of the next amp usually using a CR network to buffer dc whilst allowing the information to pass through unhindered. Does that make sense?? I hope it's right or Les is gonna have me in detention!! ;) Here ya go RichWikipedia Biasing Link here @Eduardo. Thanks very much indeed for telling us about the way this works and how you've managed it. Although it's above my full understanding I think I understand at least the concept. Like Les, I too appreciate all of your efforts in getting HC up and running on the Vega. It really is incredible that you, along with the other guys, have managed this incredible achievement! @Les. I really appreciate you explaining this for me! I'm now about to put a schematic on a bit of paper and see if I can get it to work, if you know what I mean! Thanks! Funnily enough, I mentioned to you a couple of posts ago that I thought the modern way of industry was box-changing rather than discrete components etc, well, that unfortunately is the way my industry went. I now understand how 'de-skilled' I've become over the last few decades! Thanks guys! Edited August 16, 2011 by TimboDavis Link to comment Share on other sites More sharing options...
Guest lesd Posted August 16, 2011 Report Share Posted August 16, 2011 @Les. I really appreciate you explaining this for me! I'm now about to put a schematic on a bit of paper and see if I can get it to work, if you know what I mean! Thanks! Funnily enough, I mentioned to you a couple of posts ago that I thought the modern way of industry was box-changing rather than discrete components etc, well, that unfortunately is the way my industry went. I now understand how 'de-skilled' I've become over the last few decades! Thanks guys! Well, as Eduardo says; 'That's the world we live in'. Circuit densities have increased enormously of the the decades (Moore's law) so you normally get quite a lot on a chip. You still need the basic knowledge though. There are specialist applications where you need to use discrete components. I designed a 2.8KW D class H bridge once(but the bandwidth was only 2KHz); you can't buy chips like that. Also, at that power rating, even at 95% efficiency, heat is a problem. Cheers! Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now