If the target is to obtain lowest THD number, when a single frequency is measured by a conventional distortion analyzer, then there's specific sets of design rules, well documented by Self and Cordell.
If the target is to obtain lowest peak IM distortion, then according to this article by Nelson Pass (http://www.firstwatt.com/pdf/art_dist_fdbk.pdf), GNFB will likely makes things worse.
The real question to me is, whether it is valid to equate "audio transparency" simply to a low THD figure measured by a conventional distortion analyzer.
If the idea is to be accurate, and reduce output distortion THD to 0.002% or lower, why do tube amps sound wonderful with horrific levels at THD? Is THD important? Are all tube amps effect boxes, and why do people like them?
Hugh
Hugh
Er no. If you read the whole article it says that if you are going to use GNFB you need lots of it, not pussy footing around with a few dB.
What is 'transparent'? Well research says one thing and magazines say another!
Tube amps don't have to have high THD. I have one that will hit its 50th birthday next year and is 0.1%THD.
I did read the whole article (and many times 😉), and here is an extract at the last page:
"Negative feedback can reduce the total quantity of distortion, but it adds new components on its own, and tempts the designer to use more cascaded gain stages in search of better numbers, accompanied by greater feedback frequency stability issues.
The resulting complexity creates distortion which is unlike the simple harmonics associated with musical instruments, and we see that these complex waves can gather to create the occasional tsunami of distortion, peaking at values far above those imagined by the distortion specifications.
If you want the peak distortion of the circuit of figure 13 to remain below .1% with a complex signal, then you need to reduce it by a factor of about 3000. 70 dB of feedback would do it, but that does seems like a lot.
By contrast, it appears that if you can make a single stage operate at .01% 2nd harmonic with a single tone without feedback, you could also achieve the .1% peak in the complex IM test."
BTW, could you point me to the researches that you're referring to?The resulting complexity creates distortion which is unlike the simple harmonics associated with musical instruments, and we see that these complex waves can gather to create the occasional tsunami of distortion, peaking at values far above those imagined by the distortion specifications.
If you want the peak distortion of the circuit of figure 13 to remain below .1% with a complex signal, then you need to reduce it by a factor of about 3000. 70 dB of feedback would do it, but that does seems like a lot.
By contrast, it appears that if you can make a single stage operate at .01% 2nd harmonic with a single tone without feedback, you could also achieve the .1% peak in the complex IM test."
Most THD and all are measured with resistor load, speaker is reactive and interact with the amp. Even if you have absolute zero ohm output impedance, that still do not have a flat transfer of energy to the speaker. I don't think THD really means everything.
I remember two amps that impressed me the most, one was a Cary single ended tube amp with a nice Sonus Febous speaker, it was heavenly. Much much more impressive than Krell that I listened to, I think both cost just as much.
don't just listen to everyone, listen to it yourself and see what is important.
I really think there is a lot of truth that people listen to the distortion....or effect as some people put it. But do you care as long as you really like it? As I repeat over and over, speaker is the single piece that introduce the most distortion. Your sound is SET by the speaker you choose. Any amp is just to bring the best out of the speaker. It's all distortion, just what distortion you prefer.
Last edited:
Thanks DF96. I had to read what frequency compensation is. In layman's terms the circuit design has to be specific at each levels of feed back value. Like oscillations etc.
----
I guess what I wanted was all the benefits of all topology in one box. Out of advantages and disadvantages of any design one wants only advantages. be it advantage of low distortion, good power, no ills of feedback, efficiency etc. I think it would be difficult to achieve.
Regards.
----
I guess what I wanted was all the benefits of all topology in one box. Out of advantages and disadvantages of any design one wants only advantages. be it advantage of low distortion, good power, no ills of feedback, efficiency etc. I think it would be difficult to achieve.
Regards.
Hi Hugh,
One recurring point of confusion in these discussions that 'I like it' is equated with 'this is the best' and, to make matters worse, nobody has a clear definition of 'the best'.
Many people here who strive for what is loosely called 'transparency' equate that to 'input and output identical except for more level/power'. That is relatively easy to check with a battery of test equipment to our disposal (and please, don't think we only look at THD - that's sooo 20th century!).
It is then natural to say that the amp that has the least difference between input and output is 'the best'.
But maybe my preference is for another amp that is not so very transparent, and I am tempted to call that 'the best'. Bingo - comms breakdown, misunderstanding, endless discussions on diyaudio of people speaking two different languages.
I agree with SY that you should first say what you want, a technically 'perfect' amp with identical input and output, or an amp that you really like. Accept that these can be two different ones.
If there's an amp you really like, who cares whether it has NFB, hot and cold running water or genetically manipulated caps?
But if you talk about transparency, that's a technical discussion where NFB has a place and we should not poison that discussion with 'but I like the other one'. Good for you - your search is over!
OTOH, it will mak this place pretty boring - like a pub where everyone agrees which girl is the hottest. Boring. 😎
Jan
on the basis of this review I am not going to bother with the effort of downloading it.
Or should I read it for a laugh?
edit
This has convinced me I made the right decision.
Last edited:
But there will probably be a large consensus as to which girl is hottest. Isn't that what a good amplifier design should aim for? The majority of listeners find the sound pleasing?
Really what we should aim for is a set of measurements that correlate well with aural pleasure - seems the existing ones in their present form are only good for checking an amp's working properly. But what would I know? Apparently I'm confused because I think amplifiers should sound good rather than just measure well...
I believe the biggest difference in amplifier is how they deal with multi-tone signals.
I recently made some simulations of CFA-VFA feedback amplifiers. (amplifiers that I have actually built and listened to). Amplifiers that are close to identical in specifications and distortion. amplifiers built on the same OPS and use the same PSU and output-board as base-unit.
Input stages and VAS are on separate PCBs that can simply be interchanged in a matter of seconds.
Sound experience with the two were completely different with a different focus on different elements in the music. one cold lean and a bit clinical, the other full warm slam with ease and naturalness.
Why puzzeled me quite a bit, initially I had my "hots" set on the CFA, as I believed that injecting feedback as current would be the shorter more direct and better way, one lees device cob and thus less phaseshift with the feedback signal. (CFA has better slevrate).
But I found that with two-tone testing (simulation) the CFA produces a lot more IMD. Maybe this was what I heard, and maybe also the reason (or one of the reasons) why normal distortion measurement is only one of the factors we can use to evaluate an amplifier design.
I recently made some simulations of CFA-VFA feedback amplifiers. (amplifiers that I have actually built and listened to). Amplifiers that are close to identical in specifications and distortion. amplifiers built on the same OPS and use the same PSU and output-board as base-unit.
Input stages and VAS are on separate PCBs that can simply be interchanged in a matter of seconds.
Sound experience with the two were completely different with a different focus on different elements in the music. one cold lean and a bit clinical, the other full warm slam with ease and naturalness.
Why puzzeled me quite a bit, initially I had my "hots" set on the CFA, as I believed that injecting feedback as current would be the shorter more direct and better way, one lees device cob and thus less phaseshift with the feedback signal. (CFA has better slevrate).
But I found that with two-tone testing (simulation) the CFA produces a lot more IMD. Maybe this was what I heard, and maybe also the reason (or one of the reasons) why normal distortion measurement is only one of the factors we can use to evaluate an amplifier design.
Last edited:
Just want to confirm you think the VFA is the better one.
To add to your statement, here is the schematic of the YBA amp that got me into designing my amp. Just look at it and you'll be laughing. How can this amp sound good?!!! But that's the amp that change my point of view of the importance of the amp.
Attachments
Last edited:
But if you read Nelson Pass article from the link given above, you can see that global negative feedback can lead us further away from that ideal.
MiiB, if you are talking about the slewmater amp and the different input stages, which one of the lot gave you the warm relaxed sound ?
No-one ever claimed that this was valid as the only parameter to be considered using one specific instrument- this is a fantasy of "audiophile" magazines, not anything that any engineer actually believes. Nor would they, since things like noise, frequency response, and overload recovery have been demonstrated to be audible.
But there will probably be a large consensus as to which girl is hottest. Isn't that what a good amplifier design should aim for? The majority of listeners find the sound pleasing?
.
Depends on which majority - if you are looking to sell your amp, designing and building it is easy. NOW you have to convince people to buy it, people that never have been close to it, let alone have heard it. The word 'marketing' comes to mind - which is a very powerful technique to convince people that indeed it sounds wonderful and they need it!
jan
Countless of blind tests teach us that great amounts of GNF can be perfectly transparent.
Problems can arise but that's about implementation not feedback loops as such.
Sure, objective mesurements of imaging may be dificult, but you can still set up listening tests of a DUT that either results in a positive or negative regarding detection of the DUT.
If a positive it becomes more tricky to handle the data, if negative it's (for that test) end of discussion sort of.
When it comes to driving speakers it's of course more tricky but simple tests with line level devices can teach us that feedback is just fine and that there's a lot of BS floating around in the audio biz (no news there).