Reply to thread

Re: Thoughts about amplifiers at 3/4 volume.


You guys are confusing me.


If the noise floor is referenced to the output, turning down the input will not change that output noise, but will reduce the level of the input, making the signal part of S/N less. This would logically reduce S/N. 


If the noise floor is dominated by noise in the wiring/interface, or noise floor of the source, padding down the amp input would improve the combined result.


The best case for a noisy amp (with noise referenced to it's output) is to operate WFO, but as noted most amps are designed with adequate S/N to not be the limiting factor in a system noise floor. 


 

In a typical power amp the balanced input stage will be running at a noise gain of only 2x or so, while the amp proper will be running at tens of dB voltage gain. ASSuming similar low noise components the amplifier proper should dominate the amp's internal noise.


The next consideration is how the amp trim is performed. Your advice makes sense if the amp trim is actually a fully variable gain controlling the actual amplifier stage, but in practice most amps use a variable pad placed in front of the fixed gain amp. With a fully variable gain amp, you could turn down the amp's own noise while you turned down it's gain, but for stability reasons amps are typically operated at fixed gain** with a variable pad in front of the amp, between amp stage and input receiver/buffer.


To the operator (and readers here) this is TMI and inside baseball since the net effect of how the trim is performed inside the amp is transparent to the end user, when looking at the amp as a black box. That said it does make a difference for this esoteric noise analysis. In my experience most amplifiers internal noise will be output referenced, but this noise level is not large enough to be a concern, so debating this is not really productive.


JR



** some amps may have two position switched gain, but not fully variable. Note: Low gain is the most difficult to stabilize and prevent oscillation with reactive loads when using negative feedback.