Home Propeller Head Plaza

Technical and scientific discussion of amps, cables and other topics.

[ Follow Ups ] Thread: [ Display  All  Email ] [ Propeller Head Plaza ]
[ Alert Moderator ]

Funny thing about tubes.

Funny thing about tubes.

One of the many things we can thank marketing for is the old association between “goodness” and low THD figures. “Look, buy this its good see how low those numbers are”.

In reality, distortion is more interesting and like all measurements, one must interpret them.
Your ears hear in bandwidths called a “Bark” (who knows why) and exhibit a property called masking.
If one has two tones, the higher one being quieter, one finds that until its level exceeds a certain amount it is inaudible. The greater the difference from the lower frequency, the more easily it is heard.
When the level and frequency are below the threshold, it is said to be masked by the lower tone and is inaudible.
Harmonics as in harmonic distortion are “free sound” generated by non-linear processes.
These are 2,3,4,5 etc times the frequency of the fundamental and while some are musically related, non are part of the actual signal and are not desirable (I think).
Thus, the higher the harmonic number is, the lower the level it must be in order to be inaudible (masked by the fundamental).

It is perfectly possible to have one amplifier have a greater THD than another but less audible distortion at the same time.
Imagine one amplifier with 5% 2nd harmonic and little else and another with .05% THD but all high orders

Triode and signal tubes have (compared to a transistor) a very gradual non-linearity which tends towards low order distortion. A 6sn7 triode run properly is still the lowest distortion widest bandwidth single device available and response into the MHz is normal for tubes, even the biggest slowest (electronically) tubes are wide band devices so far as audio. Many popular “old style” tubes are transmitting or RF rated and operate to 30MHz routinely.
Lest this sound like an unbounded endorsement of all tube amplifiers, many of today’s tube amplifiers are audibly colored, parts selected “by sound” and using circuit topologies rejected ages ago because they had problems. The use of “many” signal handling transformers is particularly puzzling.

Transistors have a more abrupt non-linearity but more importantly, are much less of a wide bandwidth device. A much larger current gain is possible with transistors BUT open loop bandwidth is hard ($) to achieve. On the other hand, the larger gain allows more feedback to be used to extend the bandwidth and lower the distortion. 20dB of feedback reduces distortion by nearly 20 dB also BUT in reducing the THD, it leaves behind higher order residuals. In other words, if the amp had a 5th harmonic as the highest distortion component open loop, with the loop closed, the highest component might now be the 15th harmonic.

Large amounts of negative feedback, especially if used to hide a greater non linearity is a “rubber glove for a leaky pen” fix from an engineering standpoint.
I suppose cost is the issue here, one can solve a circuit requirement many ways, often in audio the task is partly to make it as cheaply as you can. In that case, a circuit with huge nfb will measure well at 1KHz and be flat to 20KHz.
Set out to design a feed back free transistor amplifier that has low distortion and low output impedance and you see why negative feedback is a popular cost saving tool.

Lastly, in case distortion seen in light of masking seemed simple enough, one also has to fold in the fact that ones ears do not have “flat response”. At the level it takes to be just perceptible, it only takes about 0dB SPL to be audible from 2 to 5KHz. On the other hand, at 100Hz it takes 30dB (1000 times) more to be detectable and at 20Hz about 80dB, at 2Hz, about 120dB. The high frequency end has a very rapid roll off, at 10KHz it takes about 20dB and at 20KHz (for a young person) at 20KHz, about 75 dB SPL.
Now, add to this the fact that the shape of the hearing curve changes a little also with increasing level.

For those considering a “super tweeter” keep in mind, while some peoples ears can go up high, it is just like with low frequencies, it takes MUCH more SPL to be detectable. For example the 80dB louder the 20Hz tone has to be compared to the 2 – 5 KHz range to be detectable, is one hundred million times more energy.

Anyway hope this made sense.
Tom Danley

Danley Sound Labs


This post is made possible by the generous support of people like you and our sponsors:
  Schiit Audio  

Follow Ups Full Thread
Follow Ups
You can not post to an archived thread.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]