Not sure if this is the right forum - please feel free to move it to a more appropriate place.
I'm currently doing a lot of reading on amplifier design (phono and pre) but I need to read up on some basic principles. One of them is noise - johnson noise, white noise, shot noise, etc. It's value is usually stated in -dBu. And that's where the confusion starts for me. I do understand that noise in this context is an unwanted disturbance of the electrical signal. But is that also a kind of noise that is actually audible from the speakers?
Say, if i turn up my amp to full volume (no music on) and towards the end of the amp some hiss starts to appear - is that the sum of electric noise from the circuit?
Can anyone point me in the right direction of beginner friendly reading on that subject?
I'm currently doing a lot of reading on amplifier design (phono and pre) but I need to read up on some basic principles. One of them is noise - johnson noise, white noise, shot noise, etc. It's value is usually stated in -dBu. And that's where the confusion starts for me. I do understand that noise in this context is an unwanted disturbance of the electrical signal. But is that also a kind of noise that is actually audible from the speakers?
Say, if i turn up my amp to full volume (no music on) and towards the end of the amp some hiss starts to appear - is that the sum of electric noise from the circuit?
Can anyone point me in the right direction of beginner friendly reading on that subject?
Yes, and you can hear the circuit's residual noise in many preamplifiers, and sometimes in power amplifiers. It depends a lot on the amount of gain in the circuit. More gain, more noise.
If the noise varies with the volume control, it's from the phono section's input stage (or from whatever source that is selected). Sometimes there is circuitry between the source and the volume control, and this circuit can be a source of noise as well. If the noise does not vary with the volume control, it's from the line section's input stage.
That applies to wideband noise, and hum is a different matter. https://pdfs.semanticscholar.org/d770/d93b92abf067aab6e469ba816da0d205777e.pdf http://www.ti.com/lit/an/slva043b/slva043b.pdf
Last edited:
If there is enough of it, yes. These days audio electronics can have really good signal to noise ratios, and the noise may be too low to hear in most circumstances.
But not so long ago, audible amounts of noise mixed in with your music was a daily event - if you listened to music on cassette tapes, or from FM radio, noise was always waiting to be heard during quiet passages. Some was electronic noise, the rest was from the source medium itself (randomly arranged oxide particles on tape, RF interference, etc.)
With the arrival of the audio Compact Disc in the 1980s, we could have for the first time a consumer audio reproduction chain where the noise was so low that it couldn't be heard at all, at realistic listening levels.
Nowadays you are probably most likely to encounter audible amounts of electronic hiss in microphone preamplifiers connected to a loud P.A. system. Also from high-gain guitar amplifiers turned up loud.
Yes! Usually, the input stage of an amplifier dominates the noise performance of the whole circuit (because that is where the signal is smallest, since it hasn't been amplified yet.)
As you pointed out, these days the noise is usually so low that you have to turn up the amp to full volume to even be able to hear noise at all. This is the result of decades of hard work and improvements by the physicists and engineers who create our audio devices, and design and build audio circuits with them.
Wikipedia isn't bad: Noise (electronics - Wikipedia)
A little understanding of thermal and shot noise will cover a lot of ground. Those two types of noise are dictated by the physics of the natural world, and we have no choice about them - we cannot eliminate them, only manage them as best we can.
Flicker noise can become an issue in some circumstances, though improved manufacturing techniques have made semiconductor devices much better in this regard in recent decades. Often flicker noise is only relevant at the low end of the audio frequency range, though some devices (eg. some valves used for audio) may actually have enough flicker noise through the entire audio range for this to be the dominant type of noise.
-Gnobuddy
I would say that what you are hearing is the result of ground loops and loop area in a circuit.
Ground loop (electricity) - Wikipedia
https://www.emcstandards.co.uk/files/designing_ios_for_ground_loops_emcj_issue_83_july_2009_1.pdf
EMI and How to Deal With It
Spatial conduction and its countermeasures | Murata Manufacturing Co., Ltd.
Ground loop (electricity) - Wikipedia
https://www.emcstandards.co.uk/files/designing_ios_for_ground_loops_emcj_issue_83_july_2009_1.pdf
EMI and How to Deal With It
Spatial conduction and its countermeasures | Murata Manufacturing Co., Ltd.
Sorry, but this is quite incorrect. Johnson noise exists in every resistor, unless that resistor is cooled to absolute zero (-273.15 degrees Celsius, or 0 Kelvin.) You cannot get rid of Johnson noise - there is no cure for it - because it is dictated by the laws of physics.
Similarly, shot noise exists everywhere there is current flow. It is caused by the fact that electric charge ultimately comes in small "lumps", i.e. electrons. As those lumps of charge flow through a circuit, they cause noise in the same way that tiny lumps of water (raindrops) cause noise when they fall on a metal roof. You cannot get rid of shot noise either. It too is dictated by the laws of physics.
Ground loops are something entirely different. They are not caused by fundamental laws of physics, they are caused only by bad engineering, and therefore, can be engineered out of existence (or at least, to the point where they are too small to matter.)
-Gnobuddy
Lots of the old Wireless World magazines had articles on noise. You might find some of them useful.
~ Scanned and cleaned up Wireless World Articles ~
~ Scanned and cleaned up Wireless World Articles ~
Thank you all for your replies. I've already read through some of the suggested material. The ESP site (as always) proves very helpful. Still, I'm confused with the the dB, dBu, dBv, etc units out there.
Maybe i should clarify the kind of noise I'm talking about. I have two different Denon amps. With no signal attached to their inputs there will be some audible noise towards the "full on" position of the pot. I have one LM3886 amp I built recently. Same as with the Denons - some noise towards the full on position of the pot. Again, no signal at the inputs. Things start getting a lot louder as soon as I hook up a phono pre to my LM3886. Currently in combination with a Bugle2 phono pre with 40db of gain:
- Nothing connected to the RCA inputs of the phono pre: Some very clean hiss that is audible 1m away from the speakers as soon as the pot is past 12 o'clock and gets clearly audible at the listening position in the last quarter of the pot (an area which honestly i never use).
- Record player connected to the RCA of the phono pre: The level of the noise stays about the same but it's quality changes from a clear hiss to something like a mixture of hiss and hum. Again, only audible once the pot is past it's 12 o'clock position.
Really most of the time the described noise will be below the background noise of my room. I'm just trying to understand where it comes from. Is there noise from the various resistors in the signal path even if no signal is applied (as discussed above)? What else contributes to the (audible noise)? Would a lower value pot reduce the noise as it's resistance would be lower?
Maybe i should clarify the kind of noise I'm talking about. I have two different Denon amps. With no signal attached to their inputs there will be some audible noise towards the "full on" position of the pot. I have one LM3886 amp I built recently. Same as with the Denons - some noise towards the full on position of the pot. Again, no signal at the inputs. Things start getting a lot louder as soon as I hook up a phono pre to my LM3886. Currently in combination with a Bugle2 phono pre with 40db of gain:
- Nothing connected to the RCA inputs of the phono pre: Some very clean hiss that is audible 1m away from the speakers as soon as the pot is past 12 o'clock and gets clearly audible at the listening position in the last quarter of the pot (an area which honestly i never use).
- Record player connected to the RCA of the phono pre: The level of the noise stays about the same but it's quality changes from a clear hiss to something like a mixture of hiss and hum. Again, only audible once the pot is past it's 12 o'clock position.
Really most of the time the described noise will be below the background noise of my room. I'm just trying to understand where it comes from. Is there noise from the various resistors in the signal path even if no signal is applied (as discussed above)? What else contributes to the (audible noise)? Would a lower value pot reduce the noise as it's resistance would be lower?
Unlikely. Such things create hum, buzz etc. not random noise.Mark Whitney said:
No, completely untrue. Johnson noise exists wherever there is impedance and non-zero temperature. There are noise sources which only exist when some sort of signal is present (including DC) but Johnson noise is not one of them.
Yes, all resistors contribute Johnson noise. Many resistors also have some form of excess noise, which may depend on current flow etc. Semiconductors also add noise. A lower value volume pot will reduce Johnson noise, but take it too far and it may start loading the source too much and increase distortion.Mixi said:
fantastic, very helpful, thanks!
now looking at an average stereo setup with say 88db/1w speakers, 26db gain in the amp and 40db gain in the phono stage. would it be possible to make the system completely quiet? meaning that there would be no unwanted noise audible at all when the volume is set to max and no signal applied? or is it just very common to hear some sort of noise at that setting?
i do understand that 26db plus 40db of gain equals and amplification of 4.000x so i assume even the tiniest signal (or noise) will eventually be audible.
now looking at an average stereo setup with say 88db/1w speakers, 26db gain in the amp and 40db gain in the phono stage. would it be possible to make the system completely quiet? meaning that there would be no unwanted noise audible at all when the volume is set to max and no signal applied? or is it just very common to hear some sort of noise at that setting?
i do understand that 26db plus 40db of gain equals and amplification of 4.000x so i assume even the tiniest signal (or noise) will eventually be audible.
This is Johnson noise from the resistance of the volume control potentiometer. The pot forms a voltage divider at the input, and the resistance between the wiper and ground changes as you adjust it.
This is amplified noise coming from the pre-amp. Pre-amps, of course, take very small signals and amplify them along with any input noise (including hum). The amplification is so great that noise from the pre-amplifier circuit itself is also audible.
Phono cartridges also have internal resistance and the Johnson noise associated with that. Moving magnet designs also require a 47,000 ohm load resistance, with its associated noise. Moving coil designs have lower resistance and less noise, but require greater amplification.
A well designed MM phono amp will add very little noise above that of the cartridge itself. MC phono amp noise is typically set by the internal resistances of the amplifier circuit.
Good components and good pcb layout is vital to get rid of noise. I use metal film resistors rather than carbon. I tend to use star grounding on pcb's to stop power supply noise getting into the audi ground. The input signal (especially if high impedance) should be as short as physically possible. Keep audio signals away from transformers and HVAC. Use well decoupled supplies to stop feedback from subsequent stages. I use copper pours connected to ground around sensitive signals.
Since it's kinda hard to describe noise. Here's a recording made with my iphone about a foot away from the speaker. record player connected to 40db gain phono stage, 26db gain amp and volume to 100%.
phono stage - Clyp
phono stage - Clyp
In addition to his explanation of the Decibel here , take a look around Henry Ott's site. There are various references to noise of types not discussed yet in this thread that whether or not they apply to your situation are interesting to take in. The first post in his FAQ section, for example.
The "volume to 100%" may be misleading. If you play a full-volume signal at that setting, the amplifier may clip, and/or the speakers may seriously distort. Level controls and gain settings are usually set up so that it can drive the next stage into distortion at full level, so that may not be a good test.
You probably would never listen to at this volume (except maybe a very quiet passage). More realistically, you could play some "loud" program material and set the volume at the highest you would normally listen at. Then check how much noise you get at this setting.
Rane is (was?) a really good company for explaining audio-related technical things (their note110 is infamous for cable connections). THis page discusses noise and (some of) the different dB things (dBSPL, dBm, dBu, dBV are the ones that easily come to mind).
Audio Specifications
There's also this (and it's often useful to look for "pro audio" info, vs. audiophile or consumer audio, where technical terms are more often misleading, undefined or nonexistent):
http://www.sabadosco.com/wp-content/uploads/2016/01/audio.pdf
This looks complete (overkill?) for dB definition:
The Decibel
ETA: I see the previous posted also posted that last link. Great minds think alike...
Listening to your recording, I hear hum (roughly 120Hz) as well as a little hiss/noise. This might be to ground loops and/or improper grounding (as already mentioned in this thread), but could also be from the design of the unit's power supply, or its power supply capacitors may have aged and have less capacitance than when they were first manufactured.
Getting better audio can be a deep rabbit hole ...
Last edited:
No.
There are several types of "unnecessary noises". The neighbor's lawnmower. My barky dog. Humm and buzzz. These can, in principle, be totally eliminated.
BUT the Universe jiggle randomly in every part and particle. (If it didn't, we could not be here: the Universe would be totally uniform and very thin. But random fluctuations in the early stages encouraged matter-lumps like stars to form.)
If you look close enough, you ALWAYS find random motion.
Your listening room has noise. (Even after you silence the lawnmower and dog.) The sound of faraway trucks. If you build heavy sealed walls, the air molecules still jiggle. And not *that* much below the limits of human hearing. However homes, concert halls, and recording studios have noise levels significantly above the air-molecule jitter (trucks, blowers, etc).
In Audio, at every stage of history, designers struggle with this. Bell worked to get telephone speech above room noise. The carbon-mike made loud speech but added another source of noise. OK for speech but marginal even for early "gee-wizz!" radio programs. The vacuum tube has its own noise, but less than carbon-mike. Dynamic and condenser microphone designers work to trade-off output level against bandwidth, and get "just enough" of both. Phono pickup designers beat the same path.
Well, the cheap-trick is a Noise Gate. If "signal" is not-much above the inherent hiss of the system, a tiny demon turns it off; if signal gets higher, demon turns it on.
"Volume at max" is a totally arbitrary choice. I can add 10X gain and always have high hiss; or take off 10X gain and never hear hiss (or very soft passages). Since you CAN turn it down, but might complain if you could not turn-up enough, and gain is *cheap*, designers tend to give "excess gain", more than any reasonable customer could want. If they do it right, the noise is never really higher than inevitable for that gain.
Do not run phono-amps with "no" source. Some will squeal and burn your tweeters. On many the self-hiss level is higher on "nothing" than with some nominal source impedance. And how often do you listen to empty jacks? (If you do that, for phono, try a 2K resistor film in a metal shield, a rough ballpark to the 500r-47K impedance of a cartridge.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.