Interesting read on AES/Engineering observations of tubes and transistors

[LL21]

i make no comment and definitely dont want to start the 174,000th war on tubes vs solid state...just for people's reading enjoyment...

http://www.theaudioarchive.com/TAA_Resources_Tubes_versus_Solid_State.htm

 

[c1ferrari]

Thanks for the link

 

[LL21]

Thanks for the link

i went back and the sources/reports are quite dated (not that tubes are brand new technology or anything)...nevertheless interesting reading. Certainly seems broadly consistent (to me, a novice/non-techie) with what i have read from people like Atmasphere who is obviously knowledgeable being a manufacturer and a top one at that.

 

[DonH50]

Nice link, thanks! A lot of good points brought together.

I have posted many times over the years on the intrinsic distortion characteristics of tubes (factorial) vs. transistors (exponential). However, circuit implementation also plays a very big role. Also, tubes are often more linear because they have more negative feedback built in (intrinsic) to the device, so "more linear without external feedback" might be more appropriate. In any event tubes also benefit from negative feedback to stabilize their gain/bias/distortion characteristics and that narrows the gap a little.

Back in college I designed and built a differential tube pre-amp. It was fun but most commented that it sounded "too solid-state". I unbalanced it to add more even-order distortion and people liked it better, go figure.

 

[LL21]

Wow...great to see some serious people on this thread with technical knowledge. I have read your posts elsewhere and enjoyed [trying] to understand them......certainly trying to learn from them.

Do either of you have tubes in your own systems?

 

[LL21]

Well, I use SET amp occasionally when I want to hear what it sounds like. Other times ss. Its true that some tubes, for example, DHT type triodes, are the most linear amplifying devices commercially available (VMOS fets are very near as good but no longer made for general consumption, I think Nelson Pass had some made for him special at one point). The triode does have an internal feedback mechanism and that's why its so linear. The more linear a device is, the less IMD products occur,and that's IMO one of the big advantages of tube gear in general. The clipping mechanism is an advantage if you are not using external feedback around the tube, if you are, you have lost most of that advantage. People clip music a lot more than they know. Many units with power meters have slow reactions so don't show the truth on the peaks anyway. For example if you are playing at about 1 watt, and your music is 20db dynamic, then peaks will reach 100W. So, power is important for true fidelity, but since tastes and preferences come into play, and peaks are fast, many get away with truly way underated power in their systems. Tubes do react differently during those clipping events, and their harmonic spray can be compressive as it says in the article, and thus the higher (and typically lower power frequencies) can get louder during the clipping events (soft clipping anyway in the triode) and so as it says can sound louder to your ear, remember fletcher munson.

Thanks for that...all good to know. As ever an educational read. Thanks.

 

[bonzo75]

Well, I use SET amp occasionally when I want to hear what it sounds like. Other times ss. Its true that some tubes, for example, DHT type triodes, are the most linear amplifying devices commercially available (VMOS fets are very near as good but no longer made for general consumption, I think Nelson Pass had some made for him special at one point). The triode does have an internal feedback mechanism and that's why its so linear. The more linear a device is, the less IMD products occur,and that's IMO one of the big advantages of tube gear in general. The clipping mechanism is an advantage if you are not using external feedback around the tube, if you are, you have lost most of that advantage. People clip music a lot more than they know. Many units with power meters have slow reactions so don't show the truth on the peaks anyway. For example if you are playing at about 1 watt, and your music is 20db dynamic, then peaks will reach 100W. So, power is important for true fidelity, but since tastes and preferences come into play, and peaks are fast, many get away with truly way underated power in their systems. Tubes do react differently during those clipping events, and their harmonic spray can be compressive as it says in the article, and thus the higher (and typically lower power frequencies) can get louder during the clipping events (soft clipping anyway in the triode) and so as it says can sound louder to your ear, remember fletcher munson. nice link.

Yes my Lampi 7 dac is DHT.
Also understand that guitarists like Jimmy Page, Clapton, and most rock and blues (not heavy metal) guitarists preferred tube amps on their guitar - Page's Marshal was upgraded with KT88s

 

[LL21]

Yes my Lampi 7 dac is DHT.
Also understand that guitarists like Jimmy Page, Clapton, and most rock and blues (not heavy metal) guitarists preferred tube amps on their guitar - Page's Marshal was upgraded with KT88s

I think Allnic 5000 DAC is also DHT...very interesting to read about its intrinsic low distortion...

 

[zztop7]

Single note on single string

Yes my Lampi 7 dac is DHT.
Also understand that guitarists like Jimmy Page, Clapton, and most rock and blues (not heavy metal) guitarists preferred tube amps on their guitar - Page's Marshal was upgraded with KT88s

Find a single note on a single string that you like [tuned properly of course]; go to an excellent electric guitar store; pick a guitar & plug into warmed up quality tube amp [some are left with their heaters running]; pick your note [tuned]. Now switch to a warmed up quality solid state amp; pick your note. You decide.
zz.

 

[microstrip]

Unfortunately most people post interesting technical experiences on SET and old tube amplifiers having significant distortion, covering mainly the old tube sins and limitations.

We should however remember that modern tube devices, particularly the preamplifiers, have very good technical specifications and still keep the good unique characteristics of the "tube sound" referred in the papers quoted in the starting post.

 

[DonH50]

I have owned tube gear and ran a tube-based system for many, many years although SS components (mostly power amps) also rotated through. I do not currently have tube gear in my system. I have some tube gear in a closet, but sold my ARC D79 a few years ago and have not gotten my Counterpoint SA-220 out. I also have an ARC SP3a1, modified, and an old Eico EL34-based amplifier. My current system is all SS is will most likely stay that way. I like the sound of tubes but the prices have gotten too high for me and I don't really have room now (my main system, in the media room, is used for music and movies, probably more the latter).

My SP3a1 had extremely low distortion, lower than most SS gear back then and probably now (don't recall exactly but it was something like 0.001 ~ 0.005 % at 1 or 2 Vrms output). It would also put out large voltage swings into a high-impedance load (10's of volts). The D79 was rated at 1 % or somthing like that but again it was very low up to clipping, at least into my Magnepans. Don't recall the SA-220's specs nor what I measured. I worked at a couple of audio stores back then, and consulted for a couple more, so had access to a lot more gear than now. Back then McIntosh were the amps with among the highest distortion into speaker loads by my measurements (and that of some others) but they were highly-regarded nonetheless. ARC was the choice of many well-heeled audiophiles for tube gear, and Dynaco (often heavily modified) for the rest.

 

[microstrip]

(...) People clip music a lot more than they know. Many units with power meters have slow reactions so don't show the truth on the peaks anyway. For example if you are playing at about 1 watt, and your music is 20db dynamic, then peaks will reach 100W. So, power is important for true fidelity, but since tastes and preferences come into play, and peaks are fast, many get away with truly way underated power in their systems. (...)

Tom,

I think that most typical people listening to acoustical music (non amplified) do not have clipping problems. I have carried measurements in my system and a few friends at their typical listening levels - an easy measurement using a calibration level CD track at -40dB and found no one had such problem. Only when they want to impress friends for a few minutes using audiophile tracks they need high power.

We should not forget that the usual oversimplified fast calculus for power needs using the anechoic speaker efficiency data does not consider the room gain due to reflection of sound - unless you have a really absorptive room, you have to either carry measurements or use a good simulator to know your power needs.

 

[microstrip]

(...)

My SP3a1 had extremely low distortion, lower than most SS gear back then and probably now (don't recall exactly but it was something like 0.001 ~ 0.005 % at 1 or 2 Vrms output). It would also put out large voltage swings into a high-impedance load (10's of volts). (...)

Thanks Don. Did it sound like a tube or SS preamplifier?

 

[DonH50]

I don't want to get into that debate. To me both sound very good and in blind testing, within their normal operating ranges, SS and tube sound very much the same to me. Not true for power amps! And of course when pushed into saturation/soft clipping you could immediately tell.

The biggest advantage I found was that it was virtually impossible to overload/clip the phono stage of my SP3a1, unlike the majority of SS preamps I tried. The biggest disadvantage was the noise floor -- it was (almost) as quiet as SS with the right tubes, but tube selection could be a pain. If I ever pull out my TT, I'll rebuild my SP3a1. Right now it is pretty heavily modified; I should probably put it back to stock and send it to ARC if they still have that SP3a1 rebuild program they had a couple of years ago.

 

[Robh3606]

We should not forget that the usual oversimplified fast calculus for power needs using the anechoic speaker efficiency data does not consider the room gain due to reflection of sound - unless you have a really absorptive room, you have to either carry measurements or use a good simulator to know your power needs.

Hello Micro

It works out to be about -3db for each doubling of the listening difference vs. -6db free field for your average room. You can also measure peak SPL and work backwards. You can get a very close using anechoic speaker sensitivity and -3db per doubled distance.

Rob

 

[Atmasphere]

Yep, push pull and differential allow the odd harmonics to flourish.

This statement is not quite accurate.

What is real is that push-pull and differential cancel even ordered harmonics. But if the design is done properly, the odd orders will be as low or lower than a single-ended design. In the case of a preamp, it might be quite difficult to measure this distortion, which will be almost entirely the 3rd harmonic if the circuit is fully differential. The 5th and higher don't really show up unless single-ended and balanced are shared in the same circuit, for example a single-ended input in an amp but push-pull output.

The article at the link of the OP was pretty accurate. My main concern was there was not a lot of discussion about loop negative feedback, which as time goes by, I am finding is poorly understood. But that was not the focus of the article in the first place.

 

[LL21]

...What is real is that push-pull and differential cancel even ordered harmonics. But if the design is done properly, the odd orders will be as low or lower than a single-ended design...
The article at the link of the OP was pretty accurate. My main concern was there was not a lot of discussion about loop negative feedback, which as time goes by, I am finding is poorly understood. But that was not the focus of the article in the first place.

Hi Atmasphere...thanks for weighing in here. Always enjoy reading your very informative posts...great to learn little by little...

 

[microstrip]

Micro, want to share with us this test you use with the cd calibration track. Most folks don't have a scope, which is what I use at the speaker terminals to see clipping, and is accurate as your eyes are I suppose.

We can assume that no digital music will have higher output than 0 dB. I then play a 1000 Hz test track at -40 dB and measure the voltage output at the amplifier output terminals using a decent AC voltmeter - the cheaper ones only are accurate at 60Hz - with the preamplifier at the maximum level I use for normal music listening. From there PowerMax = (100 x measurement x 100 x meausrement) / 8 W, as power amplifiers are specified at 8 ohms. If the computed power in these conditions is less than the specified power of the amplifier you can be sure you are not clipping your amplifier.

The 100 factor comes from -40 dB. Remember P = V x V / R. Do not carry this measurement with a 0 dB track - it will probably damage your system and hearing!

 

[Geardaddy]

Great article Lloyd. Two quotes stood out for me:

“Engineering studios show that any amplifier adds distortion as soon as the overload point is reached. The tests show that all amplifiers could be overloaded to a certain degree without this distortion becoming noticeable. It may be concluded that these inaudible harmonics in the early overload condition might very well be causing the difference in sound coloration between tubes and transistors.”

and

“There is a close parallel here between electronic distortion and musical tone coloration that is the real key to why tubes and transistors sound different.”

The references are somewhat dated, and things have obviously moved on. I would be interested for these engineering societies to re-examine this issue in light of advances such as transistor design (Thermal-track) classD typology (Hypex in particular) and recent hybrids like those implemented by Devaliet.

One question I have for engineer types like Ralph relates to "timing errors" at the amplifier stage. I know that Goldmund, a Swiss engineering based company, makes the point that "high bandwidth" amplifiers are somehow more resistant to time and phase errors intrinsic to the amplifier itself. True or false? I would be interesting to test this with the vertex software to see if such an amplifier translated into lower "error" rates....

 

[LL21]

Great article Lloyd. Two quotes stood out for me:




and



The references are somewhat dated, and things have obviously moved on. I would be interested for these engineering societies to re-examine this issue in light of advances such as transistor design (Thermal-track) classD typology (Hypex in particular) and recent hybrids like those implemented by Devaliet.

One question I have for engineer types like Ralph relates to "timing errors" at the amplifier stage. I know that Goldmund, a Swiss engineering based company, makes the point that "high bandwidth" amplifiers are somehow more resistant to time and phase errors intrinsic to the amplifier itself. True or false? I would be interesting to test this with the vertex software to see if such an amplifier translated into lower "error" rates....

Totally agree, Geardaddy...i also would love to read an update of this article. BTW, i did not know about the Goldmund high bandwidth thing...i think Gryphon does that as well...flat to 350khz or something.