Yes, all other drivers have varying impedance with frequency, they are mass loaded spring systems with an inductive drive. Not even close to anything possible to dive with a current source.
🙂 I don't think I ever had any doubts of that, nor was that ever part of the conversation!!
No, you were outlining your own private version of physics. That we grasp only too well. What you don't seem to grasp is that you can't have your own version of physics. We have tried to explain this to you in various ways, but you don't get it. For some reason you seem to want to contradict EE, but then deny that you are contradicting it by 'explaining' your own 'paradigm'.atmasphere said:
If you contradict known physics you must expect someone to point this out - at least on this forum. By re-asserting your error you appear to be welcoming an argument, even when you deny that this is the case. It is trolls who deny physics, not those who seek to correct them.
Ralph, could I ask a few questions to try to clarify (for me, at least) what you're claiming as a new paradigm?
1. Is the amplifier a two terminal box from the viewpoint of the speaker?
2. Is the Thevenin or Norton theorem incorrect? If so, in what way and how can that be demonstrated?
3. If Thevenin and Norton are correct, will you agree that the amplifier, as a two terminal box, can be represented by a Thevenin or Norton equivalent? If not, why not, and how can that be demonstrated? (Assume we don't run into voltage swing or current supply limitations)
4. If the amplifier can be represented as a Thevenin or Norton equivalent, how does one achieve constant power into a load that has impedance variations of 5:1 or more over the audio band?
1. Is the amplifier a two terminal box from the viewpoint of the speaker?
2. Is the Thevenin or Norton theorem incorrect? If so, in what way and how can that be demonstrated?
3. If Thevenin and Norton are correct, will you agree that the amplifier, as a two terminal box, can be represented by a Thevenin or Norton equivalent? If not, why not, and how can that be demonstrated? (Assume we don't run into voltage swing or current supply limitations)
4. If the amplifier can be represented as a Thevenin or Norton equivalent, how does one achieve constant power into a load that has impedance variations of 5:1 or more over the audio band?
Member
Joined 2009
Paid Member
Absolutely a huge benefit - when used well it makes the design of good sounding affordable and reliable stereo amps much easier.
I'm not claiming its a new paradigm! Its what was around prior to the development of the voltage rules, so its foundation goes back to the 1920s. And furthermore, it was not really very organized. I came to call it that because I watch the high end industry as part of my job and saw that some of the concepts were still in use- and for that matter, evolved. I got curious and set about outlining what the principles seemed to be, sort of as a hobby. One thing is certain: I can easily prove (and already have) that I did not make it up nor am I claiming something that violates physics as DF96 is trying to paste on me. The Fisher amp I mentioned is an example of what I am talking about. If physics does not allow that why does that amp (and others like it) exist?
I call it a paradigm because the way some things were handled (and still are in some circles) is quite a bit different from how things are done using voltage rules. People that are too young and steeped in the voltage rules see it as woo; doesn't exist, violates physical laws, etc., none of which is true. Terms and definitions are different, but they don't violate any physical law. If you can't think out of the box though you won't see it that way.
1) An example of such an amplifier is an SET. It has two speaker terminals- common and whatever tap you are using.
2) As far as I know, Thevenin and Norton apply. How would they not??
Cutting to the last, (4) I don't think any amplifier that falls into this category is truly a power source. I did mention that earlier!! (its why many older speakers had level controls for their drivers as the power response of the amp was not a given). It will just do the best it can within the limitations of Ohm's law. I bet that Fisher was probably pretty good at it as its constant power characteristic was based on voltage and current feedback balanced against each other.
(Sorry if I am getting a little sensitive about this! DF96 has been putting words in my mouth and as usual when that sort of strawman logical fallacy occurs (as he has been using), they don't taste very good. You try to help a person out by giving a bit of history and they try to shoot you. People don't like it when they find out Henry Ford was supporting Hitler either. I didn't create history...)
Since the Power Paradigm is all about zero feedback and several of our amps are zero feedback, I've measured some of them against this idea to see how it stacked up. Our MA-1 did a pretty good job being off by only 1/2 db or so across a 5:1 variance in the impedance. Its a property of OTLs though that their power does not decrease very quickly into higher impedances so if the impedance range is right they do pretty well. This seems to be one reason that OTLs and ESLs can often work (counter-intuitively) quite well together. I have mentioned all of this in previous posts.
I think the reason the Power Paradigm is still around is pretty simple. People have noticed that sometimes an amplifier (usually a tube amp, mind you) with good linearity can sound better sans feedback. So how do you get flat frequency response? Voilà.
Last edited:
I don't mean to be rude, but could you answer my questions? They're pretty straightforward technical inquiries.
Nor do I mean to be rude Sy but I did do that in my post, except perhaps point 3, the answer to which I would think is 'yes'. Do you need me to go through and number things or is it pretty obvious?
You have to also take into account that your premise was erroneous- I am not claiming a new paradigm.
edit: I went back and numbered them for you.
You have to also take into account that your premise was erroneous- I am not claiming a new paradigm.
edit: I went back and numbered them for you.
Last edited:
Perhaps you haven't seen the distortion specs for the Hypex class d ncore....distortion is below the noise floor out to several watts as I recall...
The Aliens have abducted SY and replaced him...
If I read correctly what is being described is an amplifier with very poor current capabilities and huge voltage swings into a no load condition?
If I read correctly what is being described is an amplifier with very poor current capabilities and huge voltage swings into a no load condition?
referring back to http://www.firstwatt.com/pdf/art_cs_amps.pdf if you inspect the impedance curve for a lowther DX55 in free air i you will see it reaches 100 Ohms at 100Hz. This reacts well to a transconductance amplifier. However if you compare to the mini-medallion enclosure to the 0.7cuft box you can also see that putting it in the wrong box makes all the difference.
on inspection, if you are going to spaff nearly $2k on a pair of a pair of lowthers, an F1 clone is the only sensible amp to build. Or be a heretic and use any good 50W+amp and miniDSP 😀
on inspection, if you are going to spaff nearly $2k on a pair of a pair of lowthers, an F1 clone is the only sensible amp to build. Or be a heretic and use any good 50W+amp and miniDSP 😀
Sadly these days lumps of wood get good reviews. Having said that, I did build an OTL on a bit of chipboard for a lol with a friend 25 years ago and it made music. Not hifi, but fun.
Danap, thanks for the De Palma article. I LEARNED from this paper, and I am impressed how far he actually took tube amps.
Thanks. So if an amplifier is a two terminal source and it follows the Thevenin and Norton theorems and it can be represented as such, then what's the "paradigm"? Any amp is just a voltage source with a series impedance or a current source with a parallel impedance. The impedances are measured in just the way Dave and Jan suggested, regardless of "paradigms." And you agree that there's no such thing as a "constant power" amplifier.
So what is it exactly that's being disputed?
Sorry again, it was BANAT who first contributed the paper to this thread (I'm pretty sure). Thanks again.
It seems that there is some confusion between output impedance and output drive capability here. (most) tube OTLs can approach such a thing using classic constructions and fairly low power (on the order of perhaps up to 100W). The reason is that the power loss in the amplification elements is hug compared to power delivered to the load, combined with an amplification element that employs internal feedback by virtue of the physics of it's operation. So, there is a rather high resistance (not very linear compared to a resistor but resistive nevertheless for frequencies of interest) on the way from the power supply to the load, masked into a much lower amplifier output resistance by local (and perhaps global) NPV. Unlike an idealized case where there is a variable current source in series, if the load and series resistances are around the same order of magnitude, you get a linear approximation of a parabola of maximum power and if you further limit the actual range of loads to ones often used and tailor the series resistance carefully, you get much more constant maximum delivered power into load than the usual doubling for constant voltage or halving for constant current output, with half of the load impedance attached.
This sort of optimization is actually well advised as it's actually attempting to calculate the minimum number of output tubes needed for a given power output, something which seriously impacts the price-performance ratio of an OTL amp. Too few tubes, the internal resistance (NOT output impedance!) is much higher than load impedance and the maximum power delivered approximates a current limited output stage. Too manz tubes, and the internal resistance approaches the load resistance, so it begins to approximate a voltage limited output stage. Just right and you get acceptably similar power delivery into impedances found in sensible speakers. But you pay for this with VERY high thermal losses - and, if your speaker relies on a low output impedance amp driving it, speaker impedance fluctuation dependant frequency response.
There are a couple more things one could go into into that may fall into this category, one of them being that output impedance for output stages employing low or no feedback and low Gm amplifying devices - and OTL output stages are almost the showcase for this, have an output impedance quite variable depending on output voltage - courtesy of the usual tubes being used and the fact t hat it is extremely price and energy prohibitive to operate them in class A - and let me clarify here that I am taking class A to mean 'biassed in the middle of the usable operating area'. Typical tubes used in OTL output stages have rather variable amplification factors which translates into a scenario where it's almost impossible to turn off the tube so a push-pull output stage really has no discernible crossover area, it operates in a mode very much like a hyperbolic conversion output stage, so it satisfies the class A definition where the requirement is for neither half to go into cut-off. Although such stages built around the usual tubes such as the 6AS7 are much more linear when properly biassed than one would at first glance expect, there are non-trivial variations in output impedance as the dynamic operating point changes. Depending on the construction, an ideal bias point may not be reachable before the tube power dissipation limit is violated (and it should be said that this is almost a rule for OTLs, but of course not statically), which will make it even more nonlinear.
This sort of optimization is actually well advised as it's actually attempting to calculate the minimum number of output tubes needed for a given power output, something which seriously impacts the price-performance ratio of an OTL amp. Too few tubes, the internal resistance (NOT output impedance!) is much higher than load impedance and the maximum power delivered approximates a current limited output stage. Too manz tubes, and the internal resistance approaches the load resistance, so it begins to approximate a voltage limited output stage. Just right and you get acceptably similar power delivery into impedances found in sensible speakers. But you pay for this with VERY high thermal losses - and, if your speaker relies on a low output impedance amp driving it, speaker impedance fluctuation dependant frequency response.
There are a couple more things one could go into into that may fall into this category, one of them being that output impedance for output stages employing low or no feedback and low Gm amplifying devices - and OTL output stages are almost the showcase for this, have an output impedance quite variable depending on output voltage - courtesy of the usual tubes being used and the fact t hat it is extremely price and energy prohibitive to operate them in class A - and let me clarify here that I am taking class A to mean 'biassed in the middle of the usable operating area'. Typical tubes used in OTL output stages have rather variable amplification factors which translates into a scenario where it's almost impossible to turn off the tube so a push-pull output stage really has no discernible crossover area, it operates in a mode very much like a hyperbolic conversion output stage, so it satisfies the class A definition where the requirement is for neither half to go into cut-off. Although such stages built around the usual tubes such as the 6AS7 are much more linear when properly biassed than one would at first glance expect, there are non-trivial variations in output impedance as the dynamic operating point changes. Depending on the construction, an ideal bias point may not be reachable before the tube power dissipation limit is violated (and it should be said that this is almost a rule for OTLs, but of course not statically), which will make it even more nonlinear.
Last edited:
Setting aside arguments about design philosophy and violations of physical
principles, 😱 , we have a pair of Atmasphere M-60's here as a reference
driving circa 1962 Tannoys.
They are subjectively top rank in sound quality.
😎