What makes an amplifier have a lower noise floor?

[sam9]

In my opinion J. Curl's most important comment was regarding the shorting plug. I say that because I you can measure the problem correctly there's not much hope for solving it.

The second most important from a DIY perspective is noting the noise sources external the amplifier circuitry. Preamp, grounding issues, power supplies. And then there are noisy transformers, especially if on the same circuit as a triac (includes light dimmers).

My point is not to make light of noise in the amplifier itself, but that there are so many sources one needs to isolate and/or remove each in turn so that you can measure what is going on.

[AMV8]

Hi

John; thanks for your information on source of noise in amplifiers.

I take your point about keeping gain in each stage low to try and aviod the use of a volume control. However in practice I am still going to need some form of gain attenuation. Is it better if I attenuate the gain slightly at the input to the preamp and then use a low value, say 10k potentiometet at the output. Or am I better to use a 20k/50k potentiometer at the input to the pre amp.

Don

[pjpoes]

Sam thanks for the comment. For the record, I did test the amplifier with shorting plugs, I guess I didn't make that clear. I heard a specific type of hiss, and what I noticed was that, when I plugged my preamp into it, the hiss stayed and didn't really change in nature. I believe that the amplifier has a noise floor higher than my preamp, but I could be wrong. Mind you, both of them are rated at over 110 db's S/N, though I'm sure thats with some sort of weighting, and not something I have measured myself. Unfortunatly my computer measuring rig can't do any better than 95db's itself, so almost everything I measure comes out like that. I know my noise is better than that I guess, but not exactly how low it is.

Also, as I said, the noise is very low, and not something that bothers me, I just was hoping it could be improved upon. If anyone, like John, has input on that, it would be appreciated.

[sam9]

just to add to your worries... some computer soundcards don't do too well driving relatively low impedance loads. In this case "low impedance can mean anything less than about 30k ohm. Thus if the input impedance of the amplifier is 10-12k (normally considered acceptable for nearly any audio preamp, the soundcard might be generating considerable THD+N that obscures your measurement.

This isn't true of all soundcards, and testing a bunch of soundcards is not something I care to spend my time on so I can't give advice on specific cards. Just be aware of it. If your amplifier has a nominal input impedance of 47k (a common value), I doubt you will have a problem. You can detect the problem just by temporally sticking in 30K resistor in series - if the TDH+N figure goes down chances are the soundcard has issues. If that is the case it is just an opportunity for another small DIY project - a buffer.

[john curl]

Pjpoes, the input noise is most probably shot noise from the input transistors and resistor noise as well. All amplifiers have some, of course. IF the gain of your amplifier is too high, and your loudspeakers are very efficient (this is very important) then you will always hear some hiss.
If you have normal direct radiator loudspeakers with 90 or so dB of sensitivity, then the hiss should be very low, except if you put your ear up close to the mid or tweeter speaker. Then you still could hear some in every practical design.
Since you can't post the schematic of the amp, it is difficult to say if you are in the optimum range of operation.
If I had a amp design as an example, posted here, then everybody could learn from it. However, I understand that the designer of the amp is reluctant.
If you email the schematic to me, I will tell you what I find, and share the general answer to your question with everyone here, however I won't divulge the schematic, if that is not wanted.
When it comes to power amps, the overall gain of the amp is something that will add noise, all else being equal. Even though the input devices usually make significant noise, it is usually the parts IN SERIES with the input devices that make even more noise.
This can be feedback resistors, input pots, or RFI stopper resistors. Sometimes, the second stage can contribute as much noise as the input stage, depending on the design.
Differences in individual parts such as a 2N2222, is not too important here. Going to a HIGH BETA part can be an advantage, however.

[pjpoes]

Can I have your Email john, you can send it privately if you prefer. I'm not sure how to attach a schematic using the email system built into this forum. Thanks.

[john curl]

It is j_curl at earthlink dot net

[pjpoes]

Hey John, thanks again, this is very appreciated. Your input I hope can be beneficial for not only myself, but the DIY community as a whole. I hope you got my email ok with the file attached. Please let me know if you didn't.