SET sound question

[sumotan]

Question for you Hellokitty, if you amp sounds that good than why are you still looking for an alternative ? My answer, it sound so good that it’s clinical. Everything is there but the soul isn’t there in the music. I that in order for you to find what your looking for, you’ll have to change your mind set. Audio is such a mystery that at times all the simulation & design skill one has cannot explain
why by just changing a resistor type affects the sound. Like to you in your mind an OPT is an evil part in the equation, why not take a different view, embrace it & capitalize on its short comings. Speaking of SE, the magic that one hears can only come from well design DHT tube amps & not any other form of tubes.

Cheers

 

[gabdx]

Nope, on an SET thread, you shouldn't say that. SETs are definitely not low (none) distortion and noise amps, but quite the contrary, I'll repeat: SETs are a high distortion amps. People buy or build them exactly for their distortion and noise patterns. If you want an absolutely transparent amp, Class D is unbeatable - it is a perfectly transparent amplifying wire (regardless of what others might think about this).
Good luck playing low quality MP3s through a class D amp though. Perhaps you know, one way the MP3s get so small is by erasing as much of the naturally occuring harmonics as they can get away with, presuming that people can't hear them. That's why solo cellos sound so bad in 128kbps MP3s. Playing the same MP3 through an SET, and you'll get back at least some these harmonics and overtones, and that distortion WILL make it sound better, warmer, sexier, etc..

This is true. Unless your talking about a class B driven power tube, most SET are lacking with a 'natural pre-feedback' distortion of around 0.5 - 1% at one watt. However most of it is 2nd and has no impact on the sound.

Class D sound very good from the low cost designs to the high cost designs. They are especially 'smooth' and plenty of details for the high cost ones.

I doubt compressing a sound in mp3 removes 'naturally occurring harmonics' as the process is an undiscriminating one. mp3 have a lack of detail/profoundness and a permanent glare/halo of distortion, something lacks to the hear, but they still can have overtones and plenty of enjoyment for good recordings.

Designing a true to source SET must be extremely hard. they all add much to the music for the good or for the worse.

 

[Bigun]

Sumotan - A valid reason seek out 'tube sound' or 'SET sound' with SS amplifiers (which have the benefit of providing decent power levels) is in order to be able to better sell your products.

Nelson Pass designs SS amps that people love. However, I think he once said on this forum, if you want it to sound like a tube, use a tube. He does very well without them. He also uses feedback and knows how to use it whilst still earning acolades.

 

[sumotan]

Sorry Bigun I dont manufacture or sell any products. Like you Im just a diyer. Isn’t audio like food, everyone have different liking & taste. Like it or not my comment still stands as with regards to DHT SET sound. I myself am listening to SS but inside me I know I miss that magic sound that I’ve been exposed too but at the same time looking at those high voltage on 845 amps for example, it kind of scares me to diy .

Cheers

 

[DF96]

If I recall, you need to use at least 20db of NFB to eliminate the increase in high order harmonics that it creates.

That depends on how much distortion is in the forward path, and the pattern of that distortion. In most cases the re-entrant distortion created by NFB is smaller in amplitude than the high order distortion already present due to the active devices.

So either no NFB or a lot of it. This is a measurable proven fact of NFB.

There is an element of truth in this. About the worst level of NFB to use is 3-6dB: enough to add distortion; too little to reduce distortion by a significant amount. However, if the device is sufficiently linear then you can get away with a small amount of feedback for other reasons e.g. to lower output impedance.

There are other ways to achieve as good or better results than NFB.

It is difficult to get good results by other methods, as feedforward, balancing and pre-distortion all need accurate matching which is fine in SpiceWorld but less so on Earth.

I'm soon going to build a WCF buffer with 6c33c tubes.

I thought you never use feedback? Any cathode follower uses massive feedback.

 

[hellokitty123]

Question for you Hellokitty, if you amp sounds that good than why are you still looking for an alternative ?

Because there is a stark difference between euphonic and Hi-Fi sound. We've been over this

why not take a different view, embrace it & capitalize on its short comings.

Because I can't test everything, I need to specialize to be efficient. We've been over this too.
Output transformers have too many short comings to be a top consideration when I've heard the "magic" from OTL amps, which has none of the issues transformers have.

Speaking of SE, the magic that one hears can only come from well design DHT tube amps & not any other form of tubes.

The most orgasmic sound I ever heard came out of an OTL IHT amp.
We've also been over this.

It is difficult to get good results by other methods, as feedforward, balancing and pre-distortion all need accurate matching which is fine in SpiceWorld but less so on Earth.

I know, I usually avoid methods that rely heavily on matching, but there is a big sea of possibilities and I've achieved very good real world results.

I thought you never use feedback? Any cathode follower uses massive feedback.

I said I restrict myself from adding NFB. Cathode degeneration is impossible to avoid and necessary for the tube to function at all. I'm not that gung-ho against NFB.

 

[DF96]

Degeneration is negative feedback.

 

[hellokitty123]

I know. You're not reading my words You are taking what I said to the extreme.

 

[DF96]

Ok, so when you say you avoid feedback what you actually mean is that you avoid some feedback and enthusiastically embrace other feedback?

 

[traderbam]

Degeneration is negative feedback.

Only if you also define a resistive potentiometer as a NFB device.
I don't so degeneration is not NFB in my world.

 

[DF96]

I think we had this argument before?

Degeneration has all the effects of feedback: modifies gain, frequency response, stability, distortion, input and output impedance. Now of course it is possible to analyse any feedback situation using straightforward circuit theory while not noticing that it is feedback. This doesn't stop it being feedback.

 

[hellokitty123]

DF96, I specifically stated that I don't use Gnfb and I avoid local NFB when possible but I'm not entirely against local NFB when it is required. Basically the shorter the feedback loop the more I'm okay with it.
I'm beginning to have to repeat myself a lot here. No need to incite a fight about NFB.

 

[DF96]

Using NFB trades lower order harmonics for higher order ones until a certain point.
If I recall, you need to use at least 20db of NFB to eliminate the increase in high order harmonics that it creates. So either no NFB or a lot of it. This is a measurable proven fact of NFB. One of the many reasons I don't use it.

Sorry, when you said "NFB" I thought you meant NFB. Thanks for clarifying that you meant GNFB. You are of course aware that GNFB and NFB (i.e. local NFB) are mathematically the same, so both cause re-entrant distortion so both have to be applied either 'not at all' or 'a lot'. Fortunately most degeneration is 'a lot' of feedback so works OK.

 

[Bigun]

I don't think anybody is looking for a fight 'hellokitty' but you have set a number of rules which appear to many to fly in the face of good engineering practice, yet you have also said you are desirable of good engineering practice - it seems like making up reasons to justify what you want to do rather than just saying 'hell, this is how I like to do it and that's all there is' and at the same time asking for input from people - so of course you are going to be picked at over that

Local vs Global feedback - both are use of the same tool. It's much easier to do local because the short loop is often much easier to stabilize. Of course, with enough open loop gain any local loop can become unstable which is why high gm tubes need stoppers on every pin. It takes more skill to apply the tool for global use, but the rewards for doing so are there.

As we all know, even a triode has local feedback. The anode resistor degenerates the triode because the anode is a signal input (sensitive to ~1/mu). You can avoid degenerating the anode by cascoding it.

 

[hellokitty123]

The main reason I don't use GNFB is because I've never heard a design that relies upon GNFB for linearity that I liked. Indeed most of them I distinctly disliked. I'm not saying using NFB is definitively bogus or it can't sound good, I'm just saying I'd rather not use it because of personal experiences.
As I've said I get perfectly good results without relying on NFB as the main linearizing factor, I don't see a need to shift to NFB based design when my current design curiosities lead elsewhere.

 

[traderbam]

I think we had this argument before?

Degeneration has all the effects of feedback: modifies gain, frequency response, stability, distortion, input and output impedance. Now of course it is possible to analyse any feedback situation using straightforward circuit theory while not noticing that it is feedback. This doesn't stop it being feedback.

I disagree with you. You are erroneously confounding two engineering definitions into one.

I only intervened because you seemed to be nick-picking Hellokitty to death which is a behavior I've noticed from you in the past. Instead of nit-picking, which only benefits you, why not add something positive to the discussion or don't say anything?

 

[Bigun]

The main reason I don't use GNFB is because I've never heard a design that relies upon GNFB for linearity that I liked. Indeed most of them I distinctly disliked. I'm not saying using NFB is definitively bogus or it can't sound good, I'm just saying I'd rather not use it because of personal experiences.
As I've said I get perfectly good results withing relying on NFB as the main linearizing factor, I don't see a need to shift to NFB based design when my current design curiosities lead elsewhere.

I think NFB is a red herring here anyhow - the question your raised is how to get the SET sound. And you quickly followed up by saying you hope not to have to use a transformer because of it's imperfections. It maybe, that the SET sound is about having imperfections. These imperfections maybe what provides the 'magic' for people who like SET amplifiers. As you already have current designs that deliver 'perfectly good results' you may have to take steps in a different direction where perfection is defined by the listener not the designer ?

 

[acg]

hellokitty123, have you thought forward to the market for your SET amps?

The reason that I ask is that they really should only be used on a relatively small portion of speakers out there that have relatively benign impedance loads and at least 100dB/w/m sensitivity, preferably more. That may only be 1% of the speaker market, and in general would probably be heavily biased toward DIY speakers...horns in particular...and those DIYers may prefer to build their own amplifier rather than buy one ready made.

I love SET amps when they are used in a situation that suits them, there is nothing better that can be built in my opinion, but in most systems these days they are a flawed topology, and the market is niche at best.

 

[traderbam]

As for "good engineering practice" I must express my praise for hellokitty at attempting just that. Good practice is not to keep trying the same tired and populist ideas in spite of failure to give improved sonic results. Good engineering is doing something that gets results.

In this community, some members get highly fixated on overly simplistic theories and naive applications of textbook models and all too often shoot down people trying to break out of the resulting mental prison. An example of this is the ludicrous stand-offs between THD and the human auditory system, with the human usually being dismissed as unreliable.

So I say to hellokitty, ignore anything that sounds familiar advice and do something different, with my blessings.

 

[hellokitty123]

I think NFB is a red herring here anyhow - the question your raised is how to get the SET sound. And you quickly followed up by saying you hope not to have to use a transformer because of it's imperfections. It maybe, that the SET sound is about having imperfections. These imperfections maybe what provides the 'magic' for people who like SET amplifiers. As you already have current designs that deliver 'perfectly good results' you may have to take steps in a different direction where perfection is defined by the listener not the designer ?

Oh I'm fully aware that the transformer is a big part of it, but the transformer has the least imperfections in the midband, which is curious since SETs are known for their midrange "magic".
I've also mentioned that I've heard magic from OTL tube designs that don't have high and low frequency issues like transformers have. This is why I am currently wondering what else is at play.