Not to mention that with no input, class AB amps are running class A, including output devices.
I surely do.
Getting as close to Johnson's noise as possible is my current goal. So far chipamps have come the closest. It's easier to get THD under 120db than noisefloor to -130.
Last edited:
I think a lot of the problem is more around system level setting than the noise performance of individual parts treated in isolation.
I like to run my power amps at much lower than normal gain, for example (typically 20dB rather than 30+dB). This has two effects - firstly it gives me more excess gain for correcting distortion, which is good, and secondly it means that I have to run the input signal higher than I otherwise would, so the noise contribution of the power amp is reduced. Couple that with 4 or 5 nV/√Hz, through sensible input stage design and keeping impedances as low as the preamp is able to supply, and you can be sure there's no discernible noise at the speaker.
It's possible as long as you pay attention to the preamp stages and use reasonable power supplies for these parts (+/-12 or 15V rather than +/-5V) to ensure you don't hit clipping at the preamp, and use opamps with good distortion specs all the way to clipping.
I like to run my power amps at much lower than normal gain, for example (typically 20dB rather than 30+dB). This has two effects - firstly it gives me more excess gain for correcting distortion, which is good, and secondly it means that I have to run the input signal higher than I otherwise would, so the noise contribution of the power amp is reduced. Couple that with 4 or 5 nV/√Hz, through sensible input stage design and keeping impedances as low as the preamp is able to supply, and you can be sure there's no discernible noise at the speaker.
It's possible as long as you pay attention to the preamp stages and use reasonable power supplies for these parts (+/-12 or 15V rather than +/-5V) to ensure you don't hit clipping at the preamp, and use opamps with good distortion specs all the way to clipping.
Last edited:
Yes, this is a key element to get low noise, <=20 db gain.
While the OP was thinking only on audible noise ear to speaker, I'm pursuing high resolution, minimizing noise in every piece of equipment, including power amps.
Skipping the preamp altogether (or eventually using a TVC) is one headache less. Still, most commercial power amps are too noisy (because most people don't care).
Many posters here on DIYaudio have pointed out that you can use 3W or 5W resistors in the feedback voltage divider, to ensure very low resistance and hence very low thermal noise. The other side ("left") of the IPS LTP differential amplifier is already taken care of by the shorted input (Rthermal == zero!) and the wisely chosen RFI filtration network.
Now you simply build the IPS from 2 x (4 parallel BF862s) and you're down below 2nV/rt.Hz with very little perspiration.
Now you simply build the IPS from 2 x (4 parallel BF862s) and you're down below 2nV/rt.Hz with very little perspiration.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.