None I would say considering how many PA's are simply op-amps on steroids and how many discrete op-amps are just the same topologies realized with discrete devices. YMMV of course?
The most useful responses are buried, but mostly surround the risk factors of bare foot/sock-only stepping on a PDIP package buried in your carpet. 😀
Otherwise, they tend to have really variable behavior when you drive them out of their rated specifications. Mind your input and output impedances and read the datasheet carefully. Ask for help if you get stuck. 🙂
And binomial statistics are evidently hard.
Last edited:
How's that relevant, most PA's I've heard are deficient even on the "speech inteligibility" scale, aka completely out of question to be adequate for "hi fi music playback"?
try not posting for a while, like a week - this is like running a FFT multiple times, it lowers the noise floor and increases the value of the data.
...frankly you're all pretty much acting like children, regardless of your job/retired from job/money/degrees, or anything else...
this is a waste of my time... later...
...frankly you're all pretty much acting like children, regardless of your job/retired from job/money/degrees, or anything else...
this is a waste of my time... later...
Really PA = power amplifier (all power amplifiers) I think you made a wrong assumption.
Its an interesting question, agreed but in my estimation not quite an apposite one. Better would be 'What's wrong with many commercial implementations of opamps that gives rise to such sucky sound?'.
How about you take the time to be less ambigous, as PA also stands for "Public Address" systems. Whoever makes them these days must be some SSB radio nostalgic.
'Good' is rather a subjective word, I've got an answer which has worked for me. By which I mean that having identified the problem and then corrected it the sucky sound went away.
The identified problem in most cases was to do with the power supplies and the associated decoupling. To cut a long story short the output stages of opamps generate noise due to running classAB and this noise appears as signal-correlated ripple on the power supply. It often couples from the supply to the 0V signal reference. There are several ways to attack this, I can spell them out in more detail if you're interested?
Guys lets try not to get personal here!
None of the commercial pieces of kit using opamps that I've examined has made a distinction between 'signal ground' and 'power ground' in the layout. So first of all its important to separate these - signal ground can be considered to be 'clean' and hence no power decoupling must be returned via it. Groundfill is useful for power decoupling but less so for signal.
Secondly the feedback networks of opamps are often loading the opamps outputs unnecessarily. When JFET input opamps are used, these networks can be increased in impedance. Driving an easier load makes for lower PSU noise.
Third - output stage current source biassing can be employed in cases where feedback impedances have to be low because of noise concerns. Linear Technology shows this technique in at least one of its datasheets.
Secondly the feedback networks of opamps are often loading the opamps outputs unnecessarily. When JFET input opamps are used, these networks can be increased in impedance. Driving an easier load makes for lower PSU noise.
Third - output stage current source biassing can be employed in cases where feedback impedances have to be low because of noise concerns. Linear Technology shows this technique in at least one of its datasheets.
So you've measured the impact and you know for a fact that the return currents are flowing where they can disturb the circuit's operation?
A separate ground done improperly can be a dangerous thing and a good way to make yourself susceptible to RF and fail EMC testing.
Nope, its obvious from the topology of shared impedances that the return currents are going to be imposed on the signal ground. No need for measurements.
Yes, I agree.
"output stages of opamps generate noise due to running classAB"
Abrax, can't the opamp be kept in the class A range and stay out of the B stage or is that not possible with the opamps?
Abrax, can't the opamp be kept in the class A range and stay out of the B stage or is that not possible with the opamps?
Yes it most certainly can, though it pays to watch where the currents are flowing when the opamps are running on balanced supplies. It tends to be easier to work out what happens with only a single ended supply. Here's what Linear Technology show in one of their DSs - see the JFET on the LT1115 output.
Attachments
- Status
- Not open for further replies.