Dead quiet. Now this is an intersting question.
My two yen on the subject:
As other memebers have expressed the total amount of of perceptible noise is very dependent on your local listening environment.
So somebody may say -60 db is enough, other may insist in excess of -80 db.
However....
It is not only the total RMS noise that matters, but also its spectral distribution. If my memory serves me correctly, thermal noise should have a "white" spectra. At least in theory. Practically, the circuit's stray behavior may reduce and/or amplify certain frequencies of this spectra.
For me at least, noise in the 3 to 10 khz band is the most annoying. Don't get me wrong, all noise is annoying, but this range drives me nuts.
My two yen on the subject:
As other memebers have expressed the total amount of of perceptible noise is very dependent on your local listening environment.
So somebody may say -60 db is enough, other may insist in excess of -80 db.
However....
It is not only the total RMS noise that matters, but also its spectral distribution. If my memory serves me correctly, thermal noise should have a "white" spectra. At least in theory. Practically, the circuit's stray behavior may reduce and/or amplify certain frequencies of this spectra.
For me at least, noise in the 3 to 10 khz band is the most annoying. Don't get me wrong, all noise is annoying, but this range drives me nuts.
Carbon films are a great compromise between carbon comp and metal. They don't have the clinical harshness of the metal, none of the noise (maybe a little) and drift of the carbon. BTW, I believe my best sounding amps exhibit a small amount of hiss and hum, but the ones that are "dead quiet" don't sound as good. Maybe it's just in my head. 
Morgan Jones 3rd edit book -> Valve amps on p.445 quotesfernando_g said:
<< 100dB down below 1W isn't unreasonable >>.
I couldn't disagree more. My cat can't hear this,. so if humans haven't a chance of hearing this, what is the point of such a low figure ? Most made tube amps had -75-85dB down. This usually implies a massive 30dB++ global nfb to get this level, and time and experience says that amps which have around 15-20dB global nfb are better sounding. So which way do we want it ?
I'm sticking to -75dB down with 91dB/W sens LS. I like to hear a touch hiss when my ear is up to the tweeter.
I'm not a fan of the 3rd edition h/bk. anyone else ?
richj
fernando_g said:
If you have an AES membership you might want to pull down Roy Dolby's paper "CCIR/ARM: A Practical Noise Measurement Method" -- or ITU Recommendation J.16: "Measurement of weighted Noise in Sound-Programme Circuits" -- the latter is free but Dolby's does a better job of explaining and lays out a useful CCIR/ARM filter.
Noise from 2kHz to 10kHz is the most annoying according to the Dolby paper.
You'd be surprised at the inadequacies of RMS noise measurement with conventional DVM's -- even the HP3478A and 3456A falls down in comparison to the Fluke 8920A or HP3403C.
SY said:
....the only commercial tube amp design with such a low noise level, (at hand) was the simple 3 stage design of the Radford STA100 at -95dB down at full output. With 36dB global nfb it had a reputation as up-tight sounding.
Perhaps I'm getting jinkst-up with scales reading dB/ relative to 1W as compared to full output which I use more often. A refresher is needed.
richj
jackinnj said:
Good! I'm in the ballpark! Does that qualify me in the "golden ears" club?
Naaaaah, I don't think so. For me, a low noise environment means the interior of a diesel truck doing 70 MPH on a rural road.I remember long ago reading an article that appeared in (if my memory serves me correctly) the HP Journal, about the difficulties in calculating an accurate RMS value for wideband noise. As usual in the Journal's articles, it was quite mathematically involved, and can't claim that I fully understood everything that I read. But it had to do, among other things, with crest factors, integration periods, and windowing.
richwalters said:
I know I've enjoyed many an amp that has had a little hiss and I wonder why this is. Maybe noise isn't everything.
Maybe this is a trait of amps that haven't needed too much NFB?
Maybe a little noise makes music a little easier to listen to, by masking low level issues?
Maybe a little balanced noise is the signature of a resistor that some enjoy the warmth of?
The O/P may be assuming that dead quiet is an important goal in building an amp, because it happens to be an important goal from an engineering point of view.
Hee hee, I probably used the subject line, but it sure attracted some attention
Quite frankly, what I really want to learn are the primary causes of excessive hiss and how to reduce that. I am still considered myself new to electronics, but with literatures like MJ's books in resource like this forum. I am able to make progress to reduce hum noise, caused by either inadequate filtering or wiring. But I don't seem to have any grab to deal with hiss, kind of like, once it hisses, it simply persists.
For the time being, I am using some fairly tube unfriendly commercial speakers, read inefficient. I would like to try some more efficient ones. When I hear the hiss from 2 feet away from a 86dB ribbon, I just couldn't resist to imagine how bad that would be at 95dB; hence, the posting.
So far, there have been very good discussion and suggestion!!
Quite frankly, what I really want to learn are the primary causes of excessive hiss and how to reduce that. I am still considered myself new to electronics, but with literatures like MJ's books in resource like this forum. I am able to make progress to reduce hum noise, caused by either inadequate filtering or wiring. But I don't seem to have any grab to deal with hiss, kind of like, once it hisses, it simply persists.
For the time being, I am using some fairly tube unfriendly commercial speakers, read inefficient. I would like to try some more efficient ones. When I hear the hiss from 2 feet away from a 86dB ribbon, I just couldn't resist to imagine how bad that would be at 95dB; hence, the posting.
So far, there have been very good discussion and suggestion!!
pchw said:
That would make me up-end the chassis and cover off ...with a cue to a fault and looking at the circuit design. There are many causes of hiss, including the obvious, a dying tube.
Generalising, assuming an amp circuit isn't oscillating (i.e creating hiss), it is well mentioned in many handbooks that to keep noise down to low levels, a triode is better than a pentode, or use pentodes strapped as a triodes which is my favourite tack for both preamp and power amp designs.
Morgan J comments alot on PSU's...as a source of noise, (I don't wholly agree as I use SMPS and noise floor is ident with standard cap input filtering) but put the blame on a poor earthing circuit bus and circuit layout. Many proven amp designs in the past used carbon resistors thoughout or cracked H/S carbons in first stage positions with a good layout, so to blame the passive components isn't wholly justified. The conditions of how the tubes are used accomplishes the noise. Some tubes are notoriously born noisy, and as the world supply gradually diminishes, more and more odd ones are pressed into designs.
Many amp builders often mention the EF86 isn't the quiet pentode as it is often made out to be.
In keeping to tradition, a tube amp measured by "A" weighting noise floor giving -75dB/W will satisfy nearly all listeners with 93dB/1W sensitivity speakers and I see no point being perfectionist and engineering further. I don't. The background noise from a CD player motor or an equipment fan is often far worse and the noise from broadcast programsis equally poor.
richj
Stuart, dB is always a referenced (comparative) measurement. I use dBV because it's an unambiguous (relative to one volt) measurement. In a power amp, would dBW be referenced to one watt into eight ohms, or 2.83V? I'm unclear on this. It seems to me a proportion in which the referenced quantity would change in (potentially) every circumstance.
Mahalo,
Poinz
Mahalo,
Poinz
The hiss is definitely not buzz and no cracking. I will say it is like waterfall or gas leak. It is a rather constant ssssssssssss. Based on what I have learned, the most likely cause is from excessive gain which probably can be considered as design flaw?? The 50's-60's vintage gears seem to like to use 12AX7 big time!!jnb said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.