. . . And this is especially so in those designs that use lead comp across the feedback resistor.
Last edited:
We often forget that simulators can be used to look at the effect of driving a signal into the output of an amplifier. Of course, with a bit more difficulty we can do this with the actual equipment as well.
Good point Brad. Lead comp can confer a good 15-20 degrees of phase margin if selected carefully but this has to be weighed up against the potential RF ingress.
we've had more than a few decades now of the recommendation to run audio power amp diff pair much hotter in bias current than the noise optimum, degenerate with emitter resistors
I have some really cheap discrete audio power amp that don't give obvious problem even in my 3rd floor wood frame building in a dense city
you can find extreme problems say across the street from a commercial AM broadcast antenna but I don't think the issue is that big for most
I have some really cheap discrete audio power amp that don't give obvious problem even in my 3rd floor wood frame building in a dense city
you can find extreme problems say across the street from a commercial AM broadcast antenna but I don't think the issue is that big for most
No need to apologize John.
I too have De Palma paper for years, and I've learned good deal about symmetrical semiconductors amps from it.
BR Damir
Having red Critical Damping article that Banat linked I must confirm that, in my opinion, it is not out of date! It will have relevance as long as dynamic loudspeakers are used, and that is probably forever. It's a pity that loudspeaker manufacturers don't quote optimal values for amp output impedance.
I have a copy of "Handbuch der Lautsprechertechnik" by Friedemann Hausdorf where he was explicit that "high Bl factor drivers (those with big strong magnets) can give low LF output". I now know why it is so.
I have a copy of "Handbuch der Lautsprechertechnik" by Friedemann Hausdorf where he was explicit that "high Bl factor drivers (those with big strong magnets) can give low LF output". I now know why it is so.
So true. But remember that it also reduces the circuit gain. I always use low pass on the input and zobel on the output and try to avoid lead comp if possible and if I use it, I adjust it empirically to mitigate unwanted phenomena like this.
I think that this mysterious "paradigm" is just alluding to how different audio was in its infancy. Amplifiers were nothing like the "ultralinear" amplifier that appeared on the hi fi scene in the 50s, and certainly nothing like the solid state amplifiers that came after that. These early amplifiers were almost always no global feedback, often interstage transformer coupled, amplifiers with much higher output impedance and nonlinear frequency vs output impedance characteristics, designed to drive high Q, high impedance, high efficiency speakers. Some old movie theatre amplifiers were only around 25 watts, and the speakers were huge freaking horns. I saw old "hi fi's" from the early-mid 30s (still worked in the late 60s!) that employed speakers as high as 600 ohms- they were direct coupled (no output transformer!) to the push-pull output tubes.
This is of course very unconventional by today's standards. Everything is standardized today; amplifiers are designed to drive 4 or 8 ohm speakers and everything is supposed to be nice and compatible. Even when I blew the full range driver in my dad's mid 50s hi fi I was able to go to Radio Shack and buy an 8", 8 ohm full range speaker and it sounded way better than the original - it worked without a hitch until he got rid of it and he never said a word. You couldn't do that with an early 1920s or 1930s piece; repairing it would be a real science project.
This is of course very unconventional by today's standards. Everything is standardized today; amplifiers are designed to drive 4 or 8 ohm speakers and everything is supposed to be nice and compatible. Even when I blew the full range driver in my dad's mid 50s hi fi I was able to go to Radio Shack and buy an 8", 8 ohm full range speaker and it sounded way better than the original - it worked without a hitch until he got rid of it and he never said a word. You couldn't do that with an early 1920s or 1930s piece; repairing it would be a real science project.
For owners of high efficiency loudspeakers, adding high power series resistor at amp output would not be that big power loss problem, at least for home listening, and it could provide stability without use of output inductor. But, of course, one must have data for the specific loudspeaker (presuming we use SS amp with low output impedance) to achieve right result.
There is some anecdotal evidence floating about where when damping factor is decreased using a 1 ohm or so resistor, in series with the amplifier output the sound is more pleasant. Someone might want to do a search on the forum for this. It seems high damping factor is not a prerequisite for good sonics.
So you happen to be traveling from point A to B on a road, some places are muddy others are rocky etc...now whether to use suspension or not or hard or soft tires or to use global feedback or not well.....
I never said that you made it up, so sorry if my comments seem to imply that. I assume you picked it up from somewhere. You are claiming something which violates physics, because you are saying that there are two definitions of output impedance: the 'voltage paradigm' one and the 'power paradigm' one. There is only one, and it applies whatever paradigm you think you are using provided only that the system is sufficiently linear that output impedance is defined. For more nonlinear systems there is no output impedance at all, as impedance is a linear concept.atmasphere said:
They can't be. These are defined by the science of electronics, not the way we choose to think about a circuit. Whether I am designing a current source, a voltage source or anything else which is sufficiently linear there is only one definition of output impedance. The value may vary from near zero to near infinity but the definition remains unchanged. As I said, I think you are confusing driving ability with output impedance, or maybe trying to redefine it as V/I when it should be dV/dI - a common newbie mistake.
We thought we were helping you by giving a bit of science.
Not at all. I have merely been pointing out the consequences of your own words. You can't keep asserting something which is untrue in a forum where the majority of people know it is untrue and then claim you are not looking for an argument but merely trying to explain your alternative paradigm.
It is classic trolling behaviour to latch onto an unimportant word ("new" in this case) and then seek to use it to avoid answering the main point. Just pretend that SY meant to say 'alternative' instead of "new". You are claiming an alternative paradigm: 'power' instead of 'voltage', in which terms (such as 'output impedance') have different meanings.
I acknowledge that my posts fall on deaf ears, but I have heard nary a correlation with sound "quality" uttered in these posts, perhaps we are missing the point?. Measurements are great, yes, but at the end of the day what does gnfb offer and or do to a detriment and or benefit to the final sound?
Colin
Colin
GNFB, like all NFB, reduces distortion and flattens frequency response. In both cases this means that the amp output is more like the amp input. If you are interested in sound reproduction then sufficiently low distortion and sufficiently flat frequency response are necessary; people can argue about whether they are sufficient. This remains true whether you actually measure it or not.vynuhl.addict said:
Yes but that was last century ;-)
My very first audio DIY project was a circlotron with a pair of 807's driving Philips 9710's, with 800 ohms voice coils. The idea was that with these speakers, you wouldn't need an output xformer. Worked beautifully. As often happened and happens, market forces insisted we should all go to 8 ohms...
Jan