| Electronics > Beginners |
| TL071 distortion - bad amplifier design? |
| << < (18/26) > >> |
| AngraMelo:
Hey Mark, Keep in mind that Im a noobie at all this (electronics). So the amplifier's resistors type would have no impact on the sound? I cant talk at all about the impact a very specific part of the design of an amplifier, what I can do is this: As a musician that has been for 20 years listening to a very specific recording, one can get extremely familiar with the particulars of how it sounds and be able to perceive very small nuances. How scientifically small? I dont know. It was initially a song that I really liked nowadays I use it just to test speakers and amps. But I sometimes can hear the difference from the same model of amp/speaker if they are in the same room and position. Look, most engineers would laugh (as they have done to me) at you if you say that if you wanted to buy a guitar amp, and had all the money in the world, a tube one would be brought home. They know it has more THD than solid state but the sound at the end is what matters. I get that it must be frustrating to listen to the enormous amount of bullshit that surrounds this subject but is the example you gave on the link measurable? Correct me if Im wrong but if you wanted to see if a pot at the beginning of the chain affected the frequency response wouldnt that be easy to measure? |
| Kleinstein:
The resistor types can have some effect, but usually only very little and the loudspeaker will have a much larger effect as it's resistance is made of copper. So at the very high end (not this circuit) things like self heating of some resistors (e.g. the gain setting ones) can have a little effect and cause distortion. Another possible effect could be noise from some resistors that seen an sizable DC bias. However a good circuit often avoids this problem by not having a DC bias at noise critical resistors. Without a bias the noise of a carbon resistor is not higher than an expensive metal foil resistor. So there is quite some audio-foolery going on. Similar with capacitors - usually no need to absolutely have PP caps in an audio circuit. The cheaper polyster caps add some small phase shift - but this is more like equivalent to having the speaker moved a fraction of a millimeter. It a common misunderstanding to think that dielectric absorption would cause distortion - it's a linear effect and only gives a minor effect on phase and amplitude, usually no harmonics. It may be a little overreacting, but if one sees specs for PP caps in an amplifier, expect over-specified parts. :popcorn: |
| Mark Hennessy:
--- Quote from: AngraMelo on December 13, 2018, 04:56:19 pm ---Hey Mark, Keep in mind that Im a noobie at all this (electronics). So the amplifier's resistors type would have no impact on the sound? I cant talk at all about the impact a very specific part of the design of an amplifier, what I can do is this: As a musician that has been for 20 years listening to a very specific recording, one can get extremely familiar with the particulars of how it sounds and be able to perceive very small nuances. How scientifically small? I dont know. It was initially a song that I really liked nowadays I use it just to test speakers and amps. But I sometimes can hear the difference from the same model of amp/speaker if they are in the same room and position. Look, most engineers would laugh (as they have done to me) at you if you say that if you wanted to buy a guitar amp, and had all the money in the world, a tube one would be brought home. They know it has more THD than solid state but the sound at the end is what matters. I get that it must be frustrating to listen to the enormous amount of bullshit that surrounds this subject but is the example you gave on the link measurable? Correct me if Im wrong but if you wanted to see if a pot at the beginning of the chain affected the frequency response wouldnt that be easy to measure? --- End quote --- Broadly, there are two schools of thought in amplifier design: 1. Design an amplifier that has medium to high amounts of harmonic distortion, but play around until the circuit produces a particular spread of harmonics that is subjectively "nice". 2. Design an amplifier that has low to very low amounts of harmonic distortion, and then you don't have to worry about how the distortion sounds because it's utterly inaudible. Perhaps unexpectedly, option 2 is much easier. Option 1 is for guitar amplifier designers. And similar things used in the creative process - for example, a recording engineer might favour a particular vintage valve/tube microphone preamp because of what the distortion adds to female vocals. Or perhaps part or all of a mix might get "bounced" via a 1/4" tape machine to add some "analogue warmth". And I have no problem with any of that - it's all part of the performance. However, when it comes to hi-fi, some might favour valve amps, or perhaps analogue replay media like vinyl, because of the subjective change in sound quality that these impose on the signal. OK, that's a personal choice, but my view is that I prefer to hear what the recording engineer wanted, without adding my own flavour of "niceness" to it. In terms of how much distortion is too much for a home hi-fi, that's a difficult question to answer because not all distortion is the same. There are many different distortion mechanisms in an amplifier, and they result in different characters. But any well-designed solid-state amplifier will have less distortion than the loudspeakers, and probably less than the microphones too. So it's difficult to know what you've reached the "good enough" point. Better than 0.1% is probably a good rule of thumb - that means that the harmonics will be less than 60dB below the level of the fundamental, and if you research how the human hearing system works, you'll see why that seems like a reasonable number. Also, it's spectacularly easy to come in at well below that target. I mentioned 0.003% from an LM3886 earlier - that's 90dB of separation. This is a good presentation from a guy from Audio Precision who (informally) demonstrates the audibility of distortion. It's very interesting. I said earlier that there are folk who believe that properly implemented passive components sound radically different. They do not. The key part of that sentence is "properly implemented". Douglas Self has written about this in his book "Small Signal Design". Improperly implemented passive components might result in higher than ideal levels of harmonic distortion, but whether that's genuinely audible in strict AB or ABX testing (with tightly matched signal levels) is up for debate. Unfortunately, people don't just listen with their ears :-+ If you're new to all this, the best thing you can do is immerse yourself in books from Self and other rational authors. My biggest regret is paying far too much attention to all the BS when I was very much younger. I became a teenager in the '80s, when the subjectivist movement was gathering serious momentum and every hi-fi magazine that I enthusiastically absorbed was spouting utter nonsense about cables and suchlike. As an engineer, I'm lucky to have some out the other side of all that - many people get trapped there for ever. |
| Zero999:
I've done a quick and dirty simulation of this circuit. The original version wouldn't work because the 220R resistors didn't give the transistors in the second stage enough voltage to turn on, so I increased them to 1k. I used a VBE multiplier, similar to that proposed by spec. In real life a potential divider would need to be used to adjust the operating point. [edit]I was wrong. It works when proper transistor models are used, rather than the generic ones. I think it's because the beta of the internal, default models was too low.[/edit] I've used generic models for the components, so the simulatio isn't very accurate, but it gives the general idea of how this is supposed to work. |
| spec:
--- Quote from: Kleinstein on December 13, 2018, 06:36:03 pm ---The resistor types can have some effect, but usually only very little and the loudspeaker will have a much larger effect as it's resistance is made of copper. So at the very high end (not this circuit) things like self heating of some resistors (e.g. the gain setting ones) can have a little effect and cause distortion. Another possible effect could be noise from some resistors that seen an sizable DC bias. However a good circuit often avoids this problem by not having a DC bias at noise critical resistors. Without a bias the noise of a carbon resistor is not higher than an expensive metal foil resistor. So there is quite some audio-foolery going on. Similar with capacitors - usually no need to absolutely have PP caps in an audio circuit. The cheaper polyster caps add some small phase shift - but this is more like equivalent to having the speaker moved a fraction of a millimeter. It a common misunderstanding to think that dielectric absorption would cause distortion - it's a linear effect and only gives a minor effect on phase and amplitude, usually no harmonics. It may be a little overreacting, but if one sees specs for PP caps in an amplifier, expect over-specified parts. :popcorn: --- End quote --- I notice that you were absent when the OP asked why his amplifier wasn't working. If you are such an expert, why didn't you point out the problem, which was glaringly obvious even from a cursory inspection of the circuit. And now that has been fixed you are, once again, on the band-wagon passing judgement from on high and, once again, in a negative and contradictory fashion. You are making some pretty sweeping statements there, and with what purpose. What does 'there is no need' mean And are you the arbiter of what is needed and what is not needed? You keep saying things are expensive like metal film resistors- this is not true, not in the order of things. All my low wattage resistors are metal film, and that is the complete range- they did not cost that much. About what you need and what you don't need. Well that all depends on your own personal preferences and perception. There are no hard and fast rules, like you are implying. One person may be happy with the sound of a TV and another may only be happy with the sound of the very best high end system. Also the sound of a TV may be acceptable for listening to the news, but not serious listening to music. And there are no morals about this, the important thing is to get a system that you like the sound of. And, in your case, if you are happy with carbon resistors and aluminum and polyester capacitors in the audio signal path good for you, but don't keep lecturing and forcing your view on others? By the way, what is your audio equipment. Can you describe it so we know what your level is. If you like music and have never listened to a high end system you are in for a shock, of course you may have cloth ears in which case it would be wasted. The other thing is that you need to match your system components. If you have cheap low-definition speakers, a cheap low-definition amplifier will be compatible and will produce a cheap low-definition sound. But if you have high-end speakers that same amplifier will sound awful. I have already established that the OP is a musician, has good sound perception, and has high-end speakers and other equipment. About distortion and fidelity. The human sensory system is not fully understood and is strange. For example second harmonic distortion can sound quite pleasant while just a sniff of third harmonic distortion will sound awful. And the distortion produced by amplifiers does sound bad. The odd thing is that speakers have a relatively high level of distortion, even the best, but they sound OK, and even the human ear distorts. And finally, if you are so worried about costs, the answer is to buy a ready-made amplifier from ebay for around £10UK or, failing that, buy one of the audio power amplifier chips for about £2.50. |
| Navigation |
| Message Index |
| Next page |
| Previous page |