Tube amps have a softer, more gradual overdrive "Transfer Function". Transistor amps clip abruptly and much more audibly.
Assuming the amps are not over driven, the difference is usually very small, but generally, the transistor amps will usually generate more odd harmonic distortion, and/or musically unrelated distortion (crossover distortion in push-pull out put stages), but at very low levels, and the tube amps will usually generate more even order harmonic distortion products at arguably slightly audible levels.
The 2nd harmonic distortion product, which is an even harmonic, is the same note as is at the input of the amp, but one octave higher. It enhances the sound in that way, but even harmonic distortion is created by the same non-linearity that creates Intermodulation Distortion (IM). So if you're listening to a solo guitarist /singer, a little bit of even order distortion can slightly enhance, or "warm" the music. But if you're listening to a full orchestra, piano or a chorus, the IM distortion generated by many tube amps will cause significant generation of sum and difference frequency distortion products, and the difference frequencies will often be just a few HZ, so you might hear an audible amplitude modulation effect, that is not Hi-Fi.
Many guitar amp designs allow for a certain amount of non-linearity in order to generate some even harmonic distortion, part for the enhancement effect of that same note an octave higher, and also for the IM that comes with it, for effect. It's blatantly audible in the lead guitar solo of ZZ Top doing Well Dressed Man, when he strums two strings, two notes apart, back and forth without muting either. I like how it works there. Personally I'd rather go with the lower distortion of a well designed transistor amp for my Hi-Fi playback system, and just make sure it's never over driven.
On the issue of amplifier output impedance, output Z is reduced by negative feedback. Transistor amps usually have much more negative feedback for various reasons, so a lower output Z. Lower output Z means that the amp is more capable of forcing the reactive speaker to follow the amps input signal. When a low feedback tube amp is used (which the hard core tube amp addicts seem to love), the reactance of the speaker is significantly less controlled due to the amps higher output impedance, which can be anywhere from about 1 ohm to 8 ohms from my experience. The other problem with an amp that has an output impedance more than a small fraction of an ohm is that it will un-calibrate a passive crossover network in a typical speaker system. All passive crossovers are designed assuming the amp output Z is essentially zero. The typical transistor power amp has an output Z that's less than a tenth of an ohm, due to it's high distortion correcting negative feedback. There's more to the story, a bunch of variables, but this is sort of the bottom line in most cases.
The trick is to get a transistor poweramp that has very low output stage crossover distortion, and good phase margin (so it never oscillates spuriously or continuously). I think most of today's transistor amps are pretty good in this way, although now the switch mode class D poweramps are taking over the world for financial reasons, and some appear to be good, but many aren't. It's still a relatively young technology in some ways. I hope this isn't too confusing.
Hi-Fi tube amps are cool looking art objects, or certainly can be. Also nostalgic for older folks. But they're large, heavy, low efficiency, and generate a lot of heat. For a guitar amp, I prefer tubes and no global negative feedback. For Hi-Fi, I prefer transistors, and use of active crossovers ahead of the poweramps. Tri-amping with tubes takes a LOT of space, generates a LOT of heat, and may actually blow the breakers in your fuse box for your house or apartment. Tube amps are much less efficient.
Assuming the amps are not over driven, the difference is usually very small, but generally, the transistor amps will usually generate more odd harmonic distortion, and/or musically unrelated distortion (crossover distortion in push-pull out put stages), but at very low levels, and the tube amps will usually generate more even order harmonic distortion products at arguably slightly audible levels.
The 2nd harmonic distortion product, which is an even harmonic, is the same note as is at the input of the amp, but one octave higher. It enhances the sound in that way, but even harmonic distortion is created by the same non-linearity that creates Intermodulation Distortion (IM). So if you're listening to a solo guitarist /singer, a little bit of even order distortion can slightly enhance, or "warm" the music. But if you're listening to a full orchestra, piano or a chorus, the IM distortion generated by many tube amps will cause significant generation of sum and difference frequency distortion products, and the difference frequencies will often be just a few HZ, so you might hear an audible amplitude modulation effect, that is not Hi-Fi.
Many guitar amp designs allow for a certain amount of non-linearity in order to generate some even harmonic distortion, part for the enhancement effect of that same note an octave higher, and also for the IM that comes with it, for effect. It's blatantly audible in the lead guitar solo of ZZ Top doing Well Dressed Man, when he strums two strings, two notes apart, back and forth without muting either. I like how it works there. Personally I'd rather go with the lower distortion of a well designed transistor amp for my Hi-Fi playback system, and just make sure it's never over driven.
On the issue of amplifier output impedance, output Z is reduced by negative feedback. Transistor amps usually have much more negative feedback for various reasons, so a lower output Z. Lower output Z means that the amp is more capable of forcing the reactive speaker to follow the amps input signal. When a low feedback tube amp is used (which the hard core tube amp addicts seem to love), the reactance of the speaker is significantly less controlled due to the amps higher output impedance, which can be anywhere from about 1 ohm to 8 ohms from my experience. The other problem with an amp that has an output impedance more than a small fraction of an ohm is that it will un-calibrate a passive crossover network in a typical speaker system. All passive crossovers are designed assuming the amp output Z is essentially zero. The typical transistor power amp has an output Z that's less than a tenth of an ohm, due to it's high distortion correcting negative feedback. There's more to the story, a bunch of variables, but this is sort of the bottom line in most cases.
The trick is to get a transistor poweramp that has very low output stage crossover distortion, and good phase margin (so it never oscillates spuriously or continuously). I think most of today's transistor amps are pretty good in this way, although now the switch mode class D poweramps are taking over the world for financial reasons, and some appear to be good, but many aren't. It's still a relatively young technology in some ways. I hope this isn't too confusing.
Hi-Fi tube amps are cool looking art objects, or certainly can be. Also nostalgic for older folks. But they're large, heavy, low efficiency, and generate a lot of heat. For a guitar amp, I prefer tubes and no global negative feedback. For Hi-Fi, I prefer transistors, and use of active crossovers ahead of the poweramps. Tri-amping with tubes takes a LOT of space, generates a LOT of heat, and may actually blow the breakers in your fuse box for your house or apartment. Tube amps are much less efficient.
Last edited:
This is one of the things that annoys me with mankind as it is at the moment - surely we have enough know-how and technology to bring everything into one place, so we can sort the wheat from the chaff and arrive at some best practices. It doesn't just apply to the audio world of course - the hours I've lost programming where in order to do 'a' you need to learn 'b', and to do 'b' you need to learn 'c', even though you have no interest in how 'b' or 'c' works, and 'b' and 'c' were never written with ease-of-use in mind.
OK, I'll get off the soapbox now...
Ancient Babylonians developed cuneiform script and left a trove of valuable information on clay tablets designed to last for millenia. Now only PhD historians are capable of retrieving this information. Tube circuits are like Babylon from the point of modern technology.
To the OP, if you are a noobie at electronics as well as diyaudio, I'd suggest you pick up a couple of tubelab pcbs (sse and spp) and get to work. That's essentially how I started a dozen amps ago and I can tell you that sometimes I prefer single-ended and sometimes I prefer push-pull. Make an amp or two and you might either enjoy it or hate it, but you'd be in a great position to determine what floats your boat. You always need a second amp for the kitchen, right?
Second the importance of basics, but I wouldn't disparage the forums, especially this one. I learned a lot of important things here that are not in the books. Often diyAudio is on top of google search for difficult to find answers. Of course, there is a dose of snake oil, this is internet. Some background and common sense is helpful in sorting grains from chaff.
Not something I would definitely agree on, especially on the part were SE gets congested. It does if it goes into heavy distortion. But a well designed SE, where critical attention to every part is given, especially to the power supply capacitors, chokes and output transformer does great with dynamic music and can achieve great levels of speed, chest punching bass and hair rising attacks with the right speakers. I can clearly distinguish multiple, tens of different tenor or baritone voices with my SE during a choir music recording.
I admit I haven't heard all PP amps in the world, but the ones I did seem to lack something I hear in well designed SE every time - naturalness and intimacy. Although I'd definitely prefer a better designed PP to a poorly made SE, especially if the later sounds more like a guitar effect pedal than a hi-fi amplifier.
Last edited:
No offense, but I consider this as a partial insult towards many wise people on this forum. The Morgan Jones book is great, I respect the author and I'd recommend the book to any beginner, but it doesn't say all about tube circuitry.
I'm not trying to offend anyone either. I fully respect the Morgan Jones book as well. I've read it a few times and I do not believe it will answer the subjective question of the original poster which is "(which topology) will provide 'thicker air' (between instruments)?".
I simply think this is the type of question that is best served by personal experimentation. In most cases, someone that has never built an amp before will probably be more likely to come to a reasonable conclusion if they get their hands dirty.
there are a lot of snake oils going around here as are useful information,
so that if you are not well grounded on the basics, you fall easy prey to
oil snake salesmen....
sure it will not, because the 'thicker air' can not be achieved without any speakers that connects to the amp...the OP needs to learn the basics and then go on building amps to gain actual knowledge, it simply can not be given to him on a silver platter...
I value your opinion a lot and you have my understanding, but I'm just used of trying to keep myself from posting subjective opinions just to keep the peace here. If you want my honest opinion, my audio devices are always "voiced" starting from tubes, capacitor brand, hook-up wire brand, connectors sand so on.
Even though there are conflicting views, I can read between the lines and take valid points from all respondees.
I agree that the way ahead should be to read the books (thanks for the recommendations) and get a grounding before trying to interpret threads on the forum.
Perhaps (as several have stated) I should try to use shops to at least get a vague idea of where my starting point (topologically) should be, although that's easier said than done where I am. I should also take note as to the speaker design being used with that particular amp type.
Then I need to get an appropriate speaker for the amp, build and experiment!
I agree that the way ahead should be to read the books (thanks for the recommendations) and get a grounding before trying to interpret threads on the forum.
Perhaps (as several have stated) I should try to use shops to at least get a vague idea of where my starting point (topologically) should be, although that's easier said than done where I am. I should also take note as to the speaker design being used with that particular amp type.
Then I need to get an appropriate speaker for the amp, build and experiment!
You might want to check out my hobby website for way more info than you wanted:
http://www.ephaseaudio.com/
http://www.ephaseaudio.com/
For me that's a bad idea. As I said earlier your room is the essential element. If you have a small-medium room just buy small-medium monitors that are suitable for valve amps and start from there. If you have a big room you can look for other speakers.
Control the room acoustics PASSIVELY as much as you can. Use a good source (as good as you can) and good music (the one that you like not crappy demonstration pieces). These are the things you need to learn first. You can use a standard valve amp that has good reputation. The amplifier is the strong link in the chain.....
If you start to use room equalizers (that do NOT change room acoustics but just pre-distort the signal to achieve an certain result), if you use active crossovers, if you are going to use sub-woofers in funny places to achieve some target frequency response then you are wasting your time in the search of a valve amp. All these things will dilute so much music that you won't hear substantial difference between amplifiers, regardless of their technology (valves, transistors, class A class AB, etc....)
Last edited:
I value your opinion a lot and you have my understanding, but I'm just used of trying to keep myself from posting subjective opinions just to keep the peace here. If you want my honest opinion, my audio devices are always "voiced" starting from tubes, capacitor brand, hook-up wire brand, connectors sand so on.
me too....and i also do not go around forums claiming this or that subjectives as gospel truth...i just build things from knowledge accumulated from previous builds...
and i also value feedback from those who got to listen to my amps, and since i built it, i would know what, why and how they come to like in what i built....
Once that Babylonian information is translated and published by "PhD historians" it is available.
Tube circuits are not that old and all the formula describing the behaviour of tube circuits and making it possible to engineer tube circuits can be understood and immediately used by contemporaries without translation or interpretational help.
Once you manage to properly simulate a tube circuit in Spice and then build it and (properly) measure it you will recognise this. There is no black art anywhere in this.
Nothing like that!
Attached there are the impulse responses (over precisely the same time window) of an affordable 8" full-range with whizzer cone (so no crossover) driven by low Zout amp (DF>100) and relatively high Zout amp (DF<2, typical valve amp with no feedback). The only thing that the low Zout amp does is damping the initial overshoot (left image). For the other amp this damping effect is visibly less. However for the rest of it the high Zout amp is the clear winner as the decay is cleaner and faster! I would say, having listenened to it, the high Zout amp has more control as the intial overshoot is a very short transient that has negligible impact on the end result.
Attachments
Experience is the greatest teacher. Sharing your previous experience to someone else can be either helpful or a bad favor. It depends.
Experience is the greatest teacher. Sharing your previous experience to someone else can be either helpful or a bad favor. It depends.
that is why we had to be careful about what we post here, i have seen a lot of food fights here about trivial things.....in the end what matters is that you are
able to build with the resources at hand...
for a newbie, it is important to get the learnings right the first time, and i am saddened to see that oftentimes, it is the snake oils that they learn more easily.....
Yes and no. LT spice is not the end of it, just a rough and approximate beginning. There are multiple non-linearities not only in active devices, but in all passive components as well. Many of these non-linearities have been explored in the past to varying degrees of depth, but this information largely did not get to to the books, and is difficult to come by.
A simple example is capacitor. For computer modeling, all capacitors are the same, and distortion introduced by capacitors is so small that it can be neglected. However, those sophisticated in the craft know that capacitors are not the same. Many, if not all, will add coloration to sound. To understand this, it is necessary to examine several types of distortion in capacitors due to dielectric absorption, voltage-dependent changes in capacitance, electromechanic transduction, flicker, contact effects, etc.
- Status
- Not open for further replies.