Well looking at the post on giving an amplifier a better soundstage, I thought I would ask this question. There have been lots of posts about grounding schemes, power supply design, etc. I was looking for a more complete outline of what effects the noise floor of an amplifier. Every amplifier I have ever owned or built has had some bit of hum, maybe not enough to be obvious at the listening position, but some. My new amplifier, which I finished last night, has no hum at all. This is good, this was my first goal in designing the layout of this amplifier. I know the major components of Hum are grounding and power supply noise. What causes hiss though?
I've heard it mentioned before that Hiss comes from a lot of different sources. It can pick up noise along the way, my assumption is, EMI and RFI could be hiss then. Hiss can be caused by the power supply if its higher frequency noise? Hiss can be a result of the components used in the amplifier? Hiss can be caused by layout issues of the amplifier? Hiss can be caused by the grounding? Now I posted these all as questions because I dont know this to be true, but have assumed it. I was hoping that some people could post answers to these. First, is it true, and second, how do you fix it. For instance, will replacing the resistors in the signal path with lower noise higher quality resistors actually reduce the noise floor. I know there is a subjective belief that it does, but does it measurably as well?
Now for the power supply, I have read that smaller caps do a better job removing high frequency noise than do larger caps. My power supply is made up of multiple large caps, each rated at 22,000 uf's. By not having many smaller caps, am I letting higher frequency noise through, would adding paralleled caps of say 1000uf's, 100ufs, and say final bypassing with film caps further reduce the hiss? In what ways do grounding effect hiss? For instance, could isolating the signal ground with diodes reduce hiss, or does that only work on hum? What about speaker grounds? What about the semiconductors? I've heard Nelson Pass and others mention that its suprising more effort isn't put into experimenting with better transistors rather than tweaky caps and resistors. I can outline the specific ones I have, but when this is said, does he and others mean finding better manufactured ones, testing them for noise, what? Does matching the transistors better improve the noise floor? This may not be as easy as I'm hoping, but I would like an amplifier with a noise floor so low I can't even hear it when I'm right next to the speaker. I don't care if thats important or not, its a goal I would like to works toward. I figure anything that improves the noise floor probably wont hurt the sound, with just a couple of acceptions. I've noticed that some manufactures manage this such as Krell and Levinson. I've also noticed that even some well conceived tube amps achieve this, such as those from Response audio.
I've heard it mentioned before that Hiss comes from a lot of different sources. It can pick up noise along the way, my assumption is, EMI and RFI could be hiss then. Hiss can be caused by the power supply if its higher frequency noise? Hiss can be a result of the components used in the amplifier? Hiss can be caused by layout issues of the amplifier? Hiss can be caused by the grounding? Now I posted these all as questions because I dont know this to be true, but have assumed it. I was hoping that some people could post answers to these. First, is it true, and second, how do you fix it. For instance, will replacing the resistors in the signal path with lower noise higher quality resistors actually reduce the noise floor. I know there is a subjective belief that it does, but does it measurably as well?
Now for the power supply, I have read that smaller caps do a better job removing high frequency noise than do larger caps. My power supply is made up of multiple large caps, each rated at 22,000 uf's. By not having many smaller caps, am I letting higher frequency noise through, would adding paralleled caps of say 1000uf's, 100ufs, and say final bypassing with film caps further reduce the hiss? In what ways do grounding effect hiss? For instance, could isolating the signal ground with diodes reduce hiss, or does that only work on hum? What about speaker grounds? What about the semiconductors? I've heard Nelson Pass and others mention that its suprising more effort isn't put into experimenting with better transistors rather than tweaky caps and resistors. I can outline the specific ones I have, but when this is said, does he and others mean finding better manufactured ones, testing them for noise, what? Does matching the transistors better improve the noise floor? This may not be as easy as I'm hoping, but I would like an amplifier with a noise floor so low I can't even hear it when I'm right next to the speaker. I don't care if thats important or not, its a goal I would like to works toward. I figure anything that improves the noise floor probably wont hurt the sound, with just a couple of acceptions. I've noticed that some manufactures manage this such as Krell and Levinson. I've also noticed that even some well conceived tube amps achieve this, such as those from Response audio.
noise is also related to the feedback network value.
consider inverting amplifier, the 2 feedback resistor value can be high or low, if you make it high value, the noise will be higher, but if you make it too low, it can load the output (less likely) or the amplifier before it(more likely) because the input impedance will be lower.
also many other thing cause high noise floor, input stage topology, active component selection, global NFB factor.
edit: sometime one must sacrifice noise for lower THD
Hartono
consider inverting amplifier, the 2 feedback resistor value can be high or low, if you make it high value, the noise will be higher, but if you make it too low, it can load the output (less likely) or the amplifier before it(more likely) because the input impedance will be lower.
also many other thing cause high noise floor, input stage topology, active component selection, global NFB factor.
edit: sometime one must sacrifice noise for lower THD
Hartono
Yeah I know Feedback can be both a good thing and a bad thing. I wonder what I might do to experiment with that, as I'm sure I could change the value. I should talk with the amp designer about that issue.
I'm still wondering if the rest of the stuff I said does or doesn't effect noisefloor. ALso, like I said, is it in a measurable objective way, or more in your head subjective way. I know sometimes two amplifiers that measure near identically in the noise department can still sound subjectivly very different, I've experienced it myself to a point.
I'm still wondering if the rest of the stuff I said does or doesn't effect noisefloor. ALso, like I said, is it in a measurable objective way, or more in your head subjective way. I know sometimes two amplifiers that measure near identically in the noise department can still sound subjectivly very different, I've experienced it myself to a point.
You folks are way out of your knowledge base.
To establish your noise floor, you must first SHORT OUT the input of your power amplifier by using a shorting plug. This can be easily made with a piece of wire and an RCA connector, and soldering the wire in such a way to short the inside pin with the outside shell. What you hear in this condition IS the effective noise of the amplifer.
IF you want to make the amplifier quieter, then you have to either lower the effective voltage gain of the amplifier or reduce the effective noise of the input stage of the amplifier.
The input stage of the amplifier generates noise essentially from the input device(s) that it uses and how they are operated and interconnected. Any additional resistance in SERIES with the input device(s) will further add noise. It usually takes an engineer with some knowledge of low noise design to make these decisions.
However, it is possible that your noise floor is NOT determined by your power amplifier at all in practice. This is because the PREAMP is usually the primary noise source for your system. This is because the self noise of the VOLUME CONTROL is amplifed by the line stage of the preamp and this is usually the primary noise contributor.
Let us say that you have a 50K ohm volume control. The worst case noise would be 1/4 of 50K ohms, and this would only be true if you were playing the music very loud.
However this is about 10nV/rt Hz. Now multiply this by 20dB for the gain of the line stage and you get 100nV/rt Hz or the effective noise of a 1.25 million ohm resistor at the input of your amplifier. This is a lot of noise, especially if you have a high gain modern solid state amp with a gain of 25 or more. This is the LOWEST worst case noise in this example. It could be worse with faulty design.
Interference suppression and bypassing won't help much.
To establish your noise floor, you must first SHORT OUT the input of your power amplifier by using a shorting plug. This can be easily made with a piece of wire and an RCA connector, and soldering the wire in such a way to short the inside pin with the outside shell. What you hear in this condition IS the effective noise of the amplifer.
IF you want to make the amplifier quieter, then you have to either lower the effective voltage gain of the amplifier or reduce the effective noise of the input stage of the amplifier.
The input stage of the amplifier generates noise essentially from the input device(s) that it uses and how they are operated and interconnected. Any additional resistance in SERIES with the input device(s) will further add noise. It usually takes an engineer with some knowledge of low noise design to make these decisions.
However, it is possible that your noise floor is NOT determined by your power amplifier at all in practice. This is because the PREAMP is usually the primary noise source for your system. This is because the self noise of the VOLUME CONTROL is amplifed by the line stage of the preamp and this is usually the primary noise contributor.
Let us say that you have a 50K ohm volume control. The worst case noise would be 1/4 of 50K ohms, and this would only be true if you were playing the music very loud.
However this is about 10nV/rt Hz. Now multiply this by 20dB for the gain of the line stage and you get 100nV/rt Hz or the effective noise of a 1.25 million ohm resistor at the input of your amplifier. This is a lot of noise, especially if you have a high gain modern solid state amp with a gain of 25 or more. This is the LOWEST worst case noise in this example. It could be worse with faulty design.
Interference suppression and bypassing won't help much.
well I'm quite aware that this is outside my knowledge base, that is why I asked. I tried to put out my misconceptions, as it seems, to learn. There is so much information from so many sources, its often hard to tell what is legit and what isn't. This is one area that has confused me for some time.
Now please tell me if I'm incorrect on this, but, if the noise is the same regardless of the preamp's volume position, does that imply that its not the preamp. I mean, typically I would agree with you, and in this case, its an HT processor, all the noisier I suppose compared with a well designed audiophile preamp, but I thought that the noise would change depending on volume control position.
Now again, if the noise is mostly dictated by the input devices, is there more to it than just layout. I mean, can matching the devices better, or finding a better manufacturer of that specific part make a difference. You mentioned the resistance after that, do better resistors play a role, or is it solely the parts value?
The reason I mentioned bypassing is that, in the past, when I've asked what bypassing does, thats what they would tell me. They would say, bypassing lowers the noise floor. I didn't know if it was true, so I was asking here.
Now please tell me if I'm incorrect on this, but, if the noise is the same regardless of the preamp's volume position, does that imply that its not the preamp. I mean, typically I would agree with you, and in this case, its an HT processor, all the noisier I suppose compared with a well designed audiophile preamp, but I thought that the noise would change depending on volume control position.
Now again, if the noise is mostly dictated by the input devices, is there more to it than just layout. I mean, can matching the devices better, or finding a better manufacturer of that specific part make a difference. You mentioned the resistance after that, do better resistors play a role, or is it solely the parts value?
The reason I mentioned bypassing is that, in the past, when I've asked what bypassing does, thats what they would tell me. They would say, bypassing lowers the noise floor. I didn't know if it was true, so I was asking here.
As I tried to say, it is a serious business to understand noise. You have to know how noise is generated and what devices are optimum for your circuit.
This is like trying to improve the performance of an auto without being a mechanic.
If you put the schematic up, it would be easy to tell you the situation, almost by inspection.
As far as 50K looking like 12.5K worst case: Think that both ends of the pots are effectively grounded, because the source is usually low impedance when it is connected to the pot. Think of it being folded back upon itself. An OPEN pot connection would be 50K or 2 times as noisy, but who listens like that?
This is like trying to improve the performance of an auto without being a mechanic.
If you put the schematic up, it would be easy to tell you the situation, almost by inspection.
As far as 50K looking like 12.5K worst case: Think that both ends of the pots are effectively grounded, because the source is usually low impedance when it is connected to the pot. Think of it being folded back upon itself. An OPEN pot connection would be 50K or 2 times as noisy, but who listens like that?
Hi everyone, something my practice has thought me is beeing aware of gain, why do you think national almost always make their designs with a tipical gain of 20. The gain of the stage is important thoug, the more the gain the more noises will be amplificated, so it's better to amplify the signal little by little for example usig 2 opamp's with a gain of 20 plus the gain of the amp you would have an overall gain of 60 plus little noise.
Regards.
Regards.
National is just not paying attention. The gain standard for THX today is 28.3 in order that .1V in = 1W out. Having more stages will not fix this, but using more of the actual input signal will. Therefore, for a given number of gain stages, you want to have them each to have as low gain as is practical so that you don't throw away most of the input signal across the volume control.
Well John, it sounds like what your telling me is that I shouldn't worry much about the noise floor. The amp really does sound good as is, I probably should let it finish breaking in and worry about these little things later. I mean, the Central Air, water pump, traffic outside are all louder than the noise floor of my amp, so I suppose its not a big issue.
I have to say John, I know you are aware of who you are, but I'm impressed to have gotten such good information from someone like you. Your designs have always impressed me, and the newer Parasound JC-1 is one of my favorite amps on the market right now. If I could afford a pair, it would be high on my list of amps to own.
I have to say John, I know you are aware of who you are, but I'm impressed to have gotten such good information from someone like you. Your designs have always impressed me, and the newer Parasound JC-1 is one of my favorite amps on the market right now. If I could afford a pair, it would be high on my list of amps to own.
Pjpoes, please do not give up, just give me more info on your amp design. To make things more clear to you, it would be advisable that you do make a shorting plug for testing both your amp and even your preamp. Later, just put it away for further use in future. This way, you can ISOLATE where the hiss is mostly coming from. I don't like hiss either, and it can be minimized, but it must be faced directly, and not with just tweaks and mods.
sometimes (and this isn't as common as it was years ago) you can have the same part-numbered transistors, but from different manufacturers have vastly different amounts of noise. a good example of this is the ubiquitous 2N2222 (not that i would want to use them in an amplifier diff amp), which are second-sourced by just about every semiconductor manufacturer in the world. even today with modern, somewhat standardized fabrication methods, it would be difficult to get 2222's from two different manufacturers that were very close to each other in noise performance (such as within 2%). when building amps that use differential inputs it's best to noise test several different types that are close to the specs you are looking for, and if neccesary test samples from several different manufacturers to find devices with the best noise performance.
Hey John, I sent you an email and was wondering if you got it. If not, let me know. The designer of the amp has asked that I don't put the schematic up on the web, but said I could share it with you, as long as you promised not to spread it around. He mentioned its nothing that special, but would rather keep it private for now.
The last post before mine was put up before I had a chance to response. I wanted to say, I have no Hum or Buzz in my amp, and the main issue I have with most amps I have heard is not hum or buzz, rather its hiss. According to Anthony Holton, the designer of my amp, what I am hearing is Shot Noise. He says that the Shot noise in the input stage is whats causing the hiss. I know very little about this, but from his comment, my assumption is, yes we can hear the shot noise of transistors.
I don't have an answer, like I said, thats what Anthony said to me. I would assume still the answer is yes you can hear device noise since it appears that both Mr. Curl and Anthony have said the most common source of hiss is the input stage, within the amplifier that is. I guess that doesn't guarantee its device noise, but I assumed thats what was being implied.
Is your thought that device noise should be so low you think it unlikely that one could hear that, and that its something else? My experience with grounding schemes is that it effects buzz/Hum, not hiss, which is why I'm inclinded to not think its that. None the less, that was one of my questions, does grounding effect just buzz/hum, or hiss as well.
Is your thought that device noise should be so low you think it unlikely that one could hear that, and that its something else? My experience with grounding schemes is that it effects buzz/Hum, not hiss, which is why I'm inclinded to not think its that. None the less, that was one of my questions, does grounding effect just buzz/hum, or hiss as well.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.