Actually, I don't either, for spectra which are below the threshold of hearing. At the risk of repeating what others have said, the graph matters, not just a single number. Therefore the single THD number is only useful for identifying really bad amps as bad, say 1% THD. Useful as a basic go-no-go QC check.
So, in my view, if the IMD spectra's "grass" is at 110dB or lower across the frequency spectrum, two amps with different IMDs but at those low levels won't sound different.
It only helps when the design is sound. Poor design will always perform badly no matter how much the OLG is and resulting high feedback is used.
Given a good design and feedback that doesn't result in instability - yes. The higher the amount of feedback, the lower the distortion will be.
Given a good design and feedback that doesn't result in instability - yes. The higher the amount of feedback, the lower the distortion will be.
As long as it doesn’t jump up to -60dBc below 1 watt, where you will be able to hear it.
Yes, this is obvious, but the point of my question was: what has allowed higher OLG now than 25 years ago?
Faster drivers and outputs with better beta vs Ic characteristics. Better circuit design, better PCB layout.
Everything has progressed, components and our knowledge. You are seeing improvements across the board as a result.
Everything has progressed, components and our knowledge. You are seeing improvements across the board as a result.
Speaking of frequency response anomolies, there is an audio store that allows you to take home a switch-box that connects variant networks in series with the output terminals of your amplifier, that when you determine the switch selection you prefer, some manufacturer he is affiliated with will make up a couple of fixed networks. This was demonstrated, and expectedly there was a selection "preferred".
There can be more than just frequency response variations occuring though. This can drastically change damping in the speakers and feedback/stability margins that can alter the sonics for better or worse. Reactive speaker loads with drastic impedance dips can alter sonics dramatically as well. These can cause current limiting and instabilities that lesser amps can't always handle, particularly with high feedback arrangements looking to make the numbers look good.
There can be more than just frequency response variations occuring though. This can drastically change damping in the speakers and feedback/stability margins that can alter the sonics for better or worse. Reactive speaker loads with drastic impedance dips can alter sonics dramatically as well. These can cause current limiting and instabilities that lesser amps can't always handle, particularly with high feedback arrangements looking to make the numbers look good.
??
Why would you do what they are suggesting? This makes zero sense.
If you have a substandard amplifier or electronics, correct that situation. If your speakers are not up to snuff (no matter what reviewers say), replace them. I'm sorry, but to put anything in series with your speakers can only ever add impedance and lose power. Some people in the world exist by selling nonsense to others who don't understand how stuff works. Doesn't matter - medical, cars or anything. Audio is especially ripe for this kind of consumer abuse.
There is so much related to how a speaker performs. Room acoustics play a large part too. I designed speaker systems for the first decade of my career. You assume low source impedance #1. Then you try to keep frequency response as flat as you can, and same for your impedance curve and phase angles. To do otherwise is not intelligent. One reason is that as you increase source resistance, the frequency response more follows the impedance curve. Is that what you want? Heck NO!
Everyone wants to be special. Everyone wants to "discover" something. Ego is number one with some folks, and they are not hard to find when you read marketing material. They sell a story, not substance.
You can't overlook that adding a network to an unknown amplifier can't be improved by a network that should have been included of advantage by proper design.
Secondly, all that ultimately matters is the sonics. What people hear, and get off on, can't be concluded as necessarily wrong to get off on, regardless of being more or less distorted in some way. My niece seems enjoying listening to music over her cell phone, seemingly as much as I do over my exotic system in the comparison. Neither is wrong, nor should she be taught not to enjoy what she listens too necessarily.
By the way... it seems nothing more than a patch that I wouldn't engage in adding
Secondly, all that ultimately matters is the sonics. What people hear, and get off on, can't be concluded as necessarily wrong to get off on, regardless of being more or less distorted in some way. My niece seems enjoying listening to music over her cell phone, seemingly as much as I do over my exotic system in the comparison. Neither is wrong, nor should she be taught not to enjoy what she listens too necessarily.
By the way... it seems nothing more than a patch that I wouldn't engage in adding
Hi Hierfi,
Nothing can be improved that way. You can't fix anything by adding impairments. Folks go crazy trying to fix the sound of defective products by mating equipment, speakers with complimentary problems. The least expensive and most satisfying way to get good sound is to simply do it right. Get products that really perform very well, and that isn't by opinions from reviewers.
The nasty truth is this. What is accurate is what 99% of people prefer long term. Different tends to be a fad, and many get stuck by initially liking something different.
People listen to what they prefer, and much may have to do with convenience. Some folks honestly don't care about sound quality - that's cool. But you can't use them as an example. By definition they don't care at all and are not impacted by quality. However, people who are not into audio do identify and prefer systems that perform well. As long as they are easy to use.
Much of snake oil sellers throw out is there to simply muddy the waters. This is where they do business. Truth is not something they want to know about.
Nothing can be improved that way. You can't fix anything by adding impairments. Folks go crazy trying to fix the sound of defective products by mating equipment, speakers with complimentary problems. The least expensive and most satisfying way to get good sound is to simply do it right. Get products that really perform very well, and that isn't by opinions from reviewers.
The nasty truth is this. What is accurate is what 99% of people prefer long term. Different tends to be a fad, and many get stuck by initially liking something different.
People listen to what they prefer, and much may have to do with convenience. Some folks honestly don't care about sound quality - that's cool. But you can't use them as an example. By definition they don't care at all and are not impacted by quality. However, people who are not into audio do identify and prefer systems that perform well. As long as they are easy to use.
Much of snake oil sellers throw out is there to simply muddy the waters. This is where they do business. Truth is not something they want to know about.
I may be all wrong, but...
I think people got less scared of conditional stability in the control-theoretical sense of the word. Due to undesired phase shifts (among other things), there are practical limitations to how high you can make the frequency where the magnitude of the loop gain passes through unity. Suppose it is 1 MHz and you let the loop gain drop at a first-order rate, like was traditionally done. The magnitude of the loop gain at 20 kHz can then at most be 1 MHz/20 kHz = 50. With a higher-order roll-off, the loop gain at 20 kHz can be much higher, but when the order exceeds two, it becomes conditionally stable. For further explsnation and a third-order audio amplifier, see
This opamp must be the all-time record holder for high gain! It's meant for dc to 2kHz use in instrumentation etc. http://www.cirrus.com/en/products/pro/detail/P1010.html
Check those dB's!!
Check those dB's!!
Another thing that happened is that Edward M. Cherry invented nested differentiating loops, a simplified version of which is known as nested Miller compensation. With that you can apply more effective local feedback at frequencies where the overall loop doesn't do much anymore. Very useful when you don't want conditional stability.
Benchmark uses a combination of feedback and error feedforward that Acoustical/QUAD invented halfway the 1970's and dubbed current dumping. Acoustical used it to get rid of bias potmeters, Benchmark to get very low distortion figures. The silly thing is that Benchmark has a patent on their version, while it is a trivial variation on Acoustical/QUAD's invention as far as I can tell.
Benchmark's patents explicitly cite as prior art, US Patent 3,970,953 invented by Peter J Walker and assigned to Quad Electroacoustics Limited. The patent examiner(s) announced to the world: we knew about Quad Current Dumping when evaluating the Benchmark application. Evidently they found Benchmark's was not a trivial variation.
I know, and I disagree with those patent examiners. In fact, I'm often amazed how trivial granted patents are. 'A person having ordinary skill in the art' must be a complete moron.
Last edited:
Marcel, don't expect much from people. It is a very low bar. What we call "ordinary skill in the art" is rare today.
I prefer TMC over TPC. I found TMC is more forgiving regarding conditional stability.
Here is an example of TMC with more than 2 poles.
But you can improve things by making "improvements". Amplifiers unknown of substance are just black boxes, hence adding components that ought to have been included in a properly designed amplifier in the first place can "improve" them to become that of a properly designed amplifier. This can include Zobel networks, inductors perhaps with resistors across them. And there is also nothing necessarily of sonic "impairment" in adding resistance to manipulate the interface to a loudspeaker. Having a stranglehold on a speaker driver isn't always sonically better.
Why is it necessary to purchase a more costly "properly designed amplifier" containing those internal components that could be added externally to an improperly designed one? Further, no "properly designed amplifier" can be guaranteed to work in every application, not withstanding how much money you throw at it. Why be in search for the holy grail if you already possess that which is gratifying, to whatever extent a person wants or needs that to be. There is no measure of gratification that can be identified as inferior by someone listening to music over whatever medium including a cell phone, as seems you are suggesting.
This is not to suggest I haven't designed and modified countless amplifiers of every technology in the manner you suggest, and simply accept the negative sonic consequences that often comes from exposing limitations and faults in other components in chain. Yet I often pause for extended periods of time when a synergy exists amongst those components. When I get bored... and always have done... move on in attempting to improve a perceived weak link. That is what audiophile designers do in my opinion.
Hi Hierfi,
Yes, improvements can be made, measureable and by the judgement of others. Sometimes you can't, depends on the amplifier design and PCB layout, plus wiring in the amplifier. Maybe I am lucky, but improvements made that improved measured performance always "sounded better". I never assess my own work, this is from clients and others I trust that never know what they will be listening to. They never know if a unit has been repaired, improved or in as-found condition. Keeps everyone honest.
You are not buying something with more expensive components. Most often the better parts cost less. Hmmm, go figure. You are not buying a box of parts. You are buying the expertise of the designer and manufacturing process. Same when you pay a technician for servicing or (truly) upgrading something. You aren't buying "x" hours on the bench. What you are paying for is decades of experience, knowledge (it costs money to go to university or school), good judgement and honesty. Never mind the equipment that person bought and learned how to use, and the parts stock. So why is it necessary? Because if you want something good, you have to pay for it, or work for it.
This discussion hasn't anything to do with people who like listening on their cell phone or factory car radio. WHat we are discussing is the person who enjoys and appreciates music. What you are dragging in is just confusion. I know people who don't care, listen to bluetooth speakers or their cells. I don't give them grief, they are happy. So let's limit the discussion to people who do care and want better.
If you modify a piece of equipment for improved performance, (and do it correctly), it does sound better. Period. If it doesn't, you didn't improve it.
synergy. Allowing a defect in one piece compliment the other so it either hides the defect, or sort of compliments it. Really?? This makes any sense to you? Improve each piece, the performance will converge towards excellent every single time. "Synergy" is used to sell audio equipment in the continuous "upgrade" game, or the person who is never happy. Connect people with good equipment and they don't feel the need to change it. That isn't good for the people trying to sell substandard equipment. They will fight tooth and nail over these concepts. Not backed up by physics or truth, they forward a religion.
Yes, improvements can be made, measureable and by the judgement of others. Sometimes you can't, depends on the amplifier design and PCB layout, plus wiring in the amplifier. Maybe I am lucky, but improvements made that improved measured performance always "sounded better". I never assess my own work, this is from clients and others I trust that never know what they will be listening to. They never know if a unit has been repaired, improved or in as-found condition. Keeps everyone honest.
You are not buying something with more expensive components. Most often the better parts cost less. Hmmm, go figure. You are not buying a box of parts. You are buying the expertise of the designer and manufacturing process. Same when you pay a technician for servicing or (truly) upgrading something. You aren't buying "x" hours on the bench. What you are paying for is decades of experience, knowledge (it costs money to go to university or school), good judgement and honesty. Never mind the equipment that person bought and learned how to use, and the parts stock. So why is it necessary? Because if you want something good, you have to pay for it, or work for it.
This discussion hasn't anything to do with people who like listening on their cell phone or factory car radio. WHat we are discussing is the person who enjoys and appreciates music. What you are dragging in is just confusion. I know people who don't care, listen to bluetooth speakers or their cells. I don't give them grief, they are happy. So let's limit the discussion to people who do care and want better.
If you modify a piece of equipment for improved performance, (and do it correctly), it does sound better. Period. If it doesn't, you didn't improve it.
synergy. Allowing a defect in one piece compliment the other so it either hides the defect, or sort of compliments it. Really?? This makes any sense to you? Improve each piece, the performance will converge towards excellent every single time. "Synergy" is used to sell audio equipment in the continuous "upgrade" game, or the person who is never happy. Connect people with good equipment and they don't feel the need to change it. That isn't good for the people trying to sell substandard equipment. They will fight tooth and nail over these concepts. Not backed up by physics or truth, they forward a religion.
Last edited:
Thanks a lot ... I had better read up about nested Miller compensation, feedforward, and the Topping NFCA. Wish these were in books.
You had better also study the Wolverine (which you praised in post #1), whose full and complete schematic is available here on diyAudio, and which doesn't use nested Miller compensation, or feedforward, or Topping NFCA. You've already got Wolverine's "book".
Last edited: