If the distortion is caused by to heavy loading of the VAS it does. Some older studio inputs are still 600 ohms input impedance. Adding 4 opamps as outputstage reduced the distortion significant.
With kind regards,
Bas
True, but if you make the distortion versus freq flat by increasing the distortion at the lower frequencies (which is what loading the Vas can do), does that make it sound better in your opinion?
jd
No. But the musical instruments does sound as it should (if the distortion is not too high, of course). I really can not see any reason to make an amplifier change the sound of an instruments or voice. Low distortion is not the only goal in amplifier design, but a low distortion amplifier should have an open loop bandwidth of 6-10 kHz at least, not 10 Hz as in OPAMP's.
ha! ha! ha! ha!
[FX: mad scientists voice]
Actually I regularly login from two places with 2 different IPs - never realised that the cookies were different!
sometimes it is the other way around.
A hypothesis followed by a long time experimenting to find a technique that exposes the evidence and finally confirms the hypothesis.
It becomes a theory.
paralleling (=halving the loading) removes the 6dB/octave rising distortion with increase in frequency. Does it?
Agreed, but I don't see what it has to do with OL bandwidth. Where does that requirement come from?
I can take that opamp with a -3dB roll off point at 10kHz and increase the DC/LF gain by say, 20dB which moves the rolloff to 1kHz. Or increase the DC/LF gain by 40dB and which moves the rollof to 100Hz.
jd
Last edited:
I prefer 6K and above open loop bandwidth. In fact, the best sounding IC that I use has a 10+KHz open loop bandwidth. It is about phase modulation.
May we be permitted to ask why? Are you a believer in the 'TIM fallacy'? Or do you believe that the change in THD between 100Hz and 1kHz is audible in, say, an LM4562 (which appears to have an OL bandwidth below 10Hz)?
I agree with that, but only if full instrumentation grounding is used, with poured copper grounds under the Op Amps. As provided in most commercial audio equipment, I find them sorely lacking in completeness of musical signals. But then I find point to point / discreet schemes and PCB / discreet schemes lacking in that same category, without proper ground planes, too
Bud.
Curious to learn more- these 'proper ground planes' are being used as sinks for decoupling only? Or as signal grounds, or as a combination of both?
Signal grounds. Though poured ground has so many benefits in so many categories that not using them, when Express PCB does not charge for their application, is almost criminal, from my tiny perspective.
Bud
Bud
I'll don my asbestos jacket here and suggest that you've been listening to less-than-competent designs. I can see no good engineering reason for putting signal grounds onto a groundplane in a piece of audio equipment. The purpose of a groundplane is to provide low inductance up to RF frequencies. But signal grounds are generally highish impedance (greater than a few hundred ohms) - they're connected to shunt resistors, feedback resistor networks, opamp inputs and the like, none of which require low inductance. Making the signal ground a large area fill just increases its capacitance to noise sources and its potential to act as an antenna to RF.
Yes, I've heard this claim from you before John. But where's the evidence for it?
I got into a kind of spat about decoupling and signal grounds with another member here a while back, so I was hinting at that. Don't take it too literally😀
If something sounds good (or bad) to a person, I don't take that as their opinion. I take it as their perception, another thing entirely.
Walt Jung has shown a way to increase the open loop bandwidth by using a compound amplifier. One with voltage feedback and one with current feedback.
Here is an example to increse the open loop bandwidth :
http://waltjung.org/PDFs/WTnT_Op_Amp_Audio_4.pdf
http://waltjung.org/PDFs/WTnT_Op_Amp_Audio_4.pdf
Okay, you say that as if "everone knows" that, yet THD was used earlier in this thread as it historically has been to compare tube and transistor amps with no mention or clue that some harmonics are more offensive than others.
I've also heard the mistaken assumption that THD is an always-valid measurement regarding speakers (which give mostly low-order harmonic distortion): "How can anyone hear 0.1 percent distortion in transistor amps through speakers that give 2 percent distortion?" when the distortions are distinct and to an extent orthogonal.
I of course can't speak for Mr. Geddes but I see it as an attempt to supersede and provide a more useful alternative to the still-widely-used THD figure, which is also a static measurement, and as discussed, can easily be misleading. I don't have any of Mr. Self's books though I've read a few pages on his website, especially "Distortion in Power Amplifiers." I'm not sure what dynamic distortion you're talking about, though I did just see his page on thermal-based dynamic distortion in chip amps.
You're claiming Geddes' measurement is not the be-all and end-all to distortion measurements, and while I may agree, it's a substantial step forward from the still-too-widely-used THD, a step that arguably should have been made decades ago. As far back as the '60's (when transistors were first applied to hifi audio) I could imagine someone inventing a distortion measurement consisting of two figures, one for low-order harmonics and one for high-order harmonics.
Thanks, that's interesting.
I hadn't heard of Mr. Putzeys before (I haven't looked into Class D amps, for one thing), it was interesting to read the IEEE article on him.
- Status
- Not open for further replies.