Support my business? Do you think that I need to contribute here to continue to be successful in audio design? Perhaps you, PMA should look to your input. All you have to do is to surpass my efforts on a consistent basis and I will be forced to retire. Please NOTE, this thread has my name on it. This is why I choose to contribute here, rather than some other forum.
Gerhard, keep to positive input, please. Just because I do audio for a living, doesn't mean that I do not try to contribute here in a useful way. I certainly don't expect you or anyone else to buy one of the products that I have made over the decades. Support your own colleagues.
Same gain at same frequency of course does have the same reduction in distortion. Higher open loop bandwidth probably has more phase margin. Are there other differences?
The assumption that what is perceived is imagined or only due to THD may not be valid.
The assumption that what is perceived is imagined or only due to THD may not be valid.
It would be interesting to find what my critics recommend as far as IC op amps. Any favorites? No bias now! Just because you work for the company. I might have missed a part or two, but then I am only beginning using IC's for my audio designs, except for servos. I will still generally design with discrete parts for as long as I can, because that is what I do, but knowing which IC's out there are REALLY exceptional can be useful, if only to compare them to discrete designs, and note any difference. I have not done that for 35 years. It might be about time to do so again.
Whilst I'd not number myself as amongst your critics, I'll put in a shout for AD815. One attractive feature for me is its ability to run from a very low supply (down to 4V or thereabouts) and it has lower 1/f noise than the more recent ADSL drivers. Its also very cheap here in China though for production I doubt you'd get anywhere near the low price here. Like any opamp though its very sensitive to the power supply impedance. At higher supplies I'd not be inclined to use it as its supply current goes up substantially with voltage (and the increased temperature that results), so much so that the part is dissipating well over 1W with a 35V supply.
I will try it but maybe too late. I dont have a system to listen with at the moment ... I return to a good system on the 14th.
Re my question.... my question was answered re harmonic structure and order being affected by the OL BW and slope to my satisfaction. I am also aware that given enough feedback the distortion -regardless of structure - could be reduced to extreamly low levels.
The answer is useful to me when I look at circuits which do Not use enough gnfb to reduce distortion to unheard levels. And, there are many such circuits sold in audio and they get regularly reviewed and measured.... and yes they are not what can be done/ is often done. It does better in helping to predict how those data will sound (again correlating data with listening).
THx-RNMarsh
Mooly, I listened to the first round of tests and thought I could tell the difference but I didn't bother saying so because I didn't see the point. The second round with LT1 and LT2, I couldn't hear any difference that I was sure of.
Of course right now my sound system is heavily improvised. It's a Denon AVR588 hooked up to a coaxial "hi-fi" car speaker in an old Boston A40 box, pointed towards the wall on the other side of the room so the sound bounces back at me. That works surprisingly well. Half of the sound actually seems to be projecting from the wall rather than the speaker.
Of course right now my sound system is heavily improvised. It's a Denon AVR588 hooked up to a coaxial "hi-fi" car speaker in an old Boston A40 box, pointed towards the wall on the other side of the room so the sound bounces back at me. That works surprisingly well. Half of the sound actually seems to be projecting from the wall rather than the speaker.
We haven't had the large numbers trying this test as we did the others and so I'm going to reveal the results today. I'll leave the Dropbox files up for a while though.
Thanks for giving them a listen. Both devices in this later test are considered "good" by audio standards. Good, but very different 🙂
Also Mooly, this is the link you should put in your signature so that people with different forum settings won't get the wrong page:
www.diyaudio.com/forums/lounge/146693-john-curls-blowtorch-preamplifier-part-ii.html#post4105741
www.diyaudio.com/forums/lounge/146693-john-curls-blowtorch-preamplifier-part-ii.html#post4105741
Educate me 🙂 I take it you are ending up somewhere else with the link. The way I do the link is by clicking the post number I want and copying the URL. I can see the difference in yours in that it is missing the 2872 part and is wrapped in brackets. You are using the "ins...owtorch-preamplifier-part-ii.html#post4105741
Thanks as I am always curious about any potential sound differences. Esp at these lower distortion levels.... requiring a better sound system than a PC. Even if it is HP 'beatsaudio' [🙂]
If it is heard by >50%, then what? Could this be an indicator that the levels we have been told for so long are no longer valid? Like - you cant hear less than 1% thd (carried over from the dark ages). ?? But also, if the amps are not over-loaded and used within mfr speced limits..... and others can still hear a difference.... then what does that say about the amps used. And what does it say about our hearing.... more tests will have to be done. I'm all ears.
THx-RNMarsh
Last edited:
Mooly, what I did was to just remove the page number, which is that number just before the .html. DIYAudio will take you to the page number rather than the post number, if the page number is specified. I'm surprised this hasn't been fixed yet.
I've read the last page here to see what I missed.
I said the harmonics behaved AS IF they went through a HP filter. I wasn't describing exactly what happened, I was describing how it LOOKED to someone investigating with an FFT or spectrum analyzer. I thought that saying it this way would be accessible to more people.
Now I'm not sure if I am really wrong about this or if people just got the wrong idea. So I'll try to put this to rest.
If your OLG is falling at 12db/oct, then won't the residual harmonics have an upward slope of 12db/oct relative to the error signal (regardless of absolute level)? After all a harmonic at 4KHz is divided by the open-loop gain at 4KHz, no matter where the fundamental is? And so OLG is not the same for each harmonic, when the signal falls within the OLG slope region?
I said the harmonics behaved AS IF they went through a HP filter. I wasn't describing exactly what happened, I was describing how it LOOKED to someone investigating with an FFT or spectrum analyzer. I thought that saying it this way would be accessible to more people.
Now I'm not sure if I am really wrong about this or if people just got the wrong idea. So I'll try to put this to rest.
If your OLG is falling at 12db/oct, then won't the residual harmonics have an upward slope of 12db/oct relative to the error signal (regardless of absolute level)? After all a harmonic at 4KHz is divided by the open-loop gain at 4KHz, no matter where the fundamental is? And so OLG is not the same for each harmonic, when the signal falls within the OLG slope region?
Last edited:
My favorites are
AD797 - very quiet, very low distortion. Sets the benchmark.
LM4562 (dual) a good modern opamp with low distortion
LME49990 - low noise, low distortion. Needs Zobrl from OP to 0V
AD4898 - very quiet, very low distortion
NE5532/34 - old warhorse from the late 70 's that still cuts it from the noise perspective between Rsource of 2.5k to about 15k. Not bad sounding either if you take good care.
I have not used any CFA's for small signal - but they aren't going to better the VFA types above on noise, but they win big on bandwidth and rise/fall time.
There are a few very fine JFET input opamps as well (OP637) but they are expensive.
In all cases, keep feedback junctions close to the inverting pin; keep capacitance from the input pins to surrounding circuitry low ( layout is important for good results); decouple well and follow data sheet recommendations.
I like to bias my opamps into class A - personal preference. Try it and see - benefits accrue from light loading of the OPS send low order only harmonics on the supply rails.
AD797 - very quiet, very low distortion. Sets the benchmark.
LM4562 (dual) a good modern opamp with low distortion
LME49990 - low noise, low distortion. Needs Zobrl from OP to 0V
AD4898 - very quiet, very low distortion
NE5532/34 - old warhorse from the late 70 's that still cuts it from the noise perspective between Rsource of 2.5k to about 15k. Not bad sounding either if you take good care.
I have not used any CFA's for small signal - but they aren't going to better the VFA types above on noise, but they win big on bandwidth and rise/fall time.
There are a few very fine JFET input opamps as well (OP637) but they are expensive.
In all cases, keep feedback junctions close to the inverting pin; keep capacitance from the input pins to surrounding circuitry low ( layout is important for good results); decouple well and follow data sheet recommendations.
I like to bias my opamps into class A - personal preference. Try it and see - benefits accrue from light loading of the OPS send low order only harmonics on the supply rails.
Last edited:
- Status
- Not open for further replies.