Some kind questions on linearity and distortion - Page 3 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 22nd March 2013, 07:59 PM   #21
expert in tautology
diyAudio Member
 
bear's Avatar
 
Join Date: Apr 2002
Location: New York State USA
Art of Electronics is a book that is not written in the usual dry text style. It is easy to read.

Get it.

Used is good.

First edition, used, is fine. Hardcover, good binding!
__________________
_-_-bear
http://www.bearlabs.com -- Btw, I don't actually know anything, FYI -- [...2SJ74 Toshiba bogus asian parts - beware! ]
  Reply With Quote
Old 24th March 2013, 02:12 PM   #22
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Hi ! in the meantime a kind but straight question
Does a transistor have an optimum point of work ?
Thanks a lot
Regards,
gino

Last edited by ginetto61; 24th March 2013 at 02:15 PM.
  Reply With Quote
Old 24th March 2013, 03:45 PM   #23
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
A transistor in isolation? No.

A transistor in a particular circuit, with particular DC voltages and signal voltages and sources and loads? Yes, maybe. Finding it is called 'design'.

People may try to tell you that a 2N12344 always works best at 8.1342mA. This is a signal that you can ignore what they tell you.

Once again, don't keep looking for simple answers to complex questions. How long is a piece of string?
  Reply With Quote
Old 24th March 2013, 04:48 PM   #24
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Ok but just one last question about linearity
What is the widely accepted test to assess the linearity of a circuit ?
is it the distortion test ?
Can we say that if a circuit has a high distortion is not linear ?
You can very well understand that one thing is to read a graph of distortion (like those in the Stereophile reviews for instance me too i understand that and found bad units ) another one is to be able to design an equipment (and this is not for everyone)
If this blessed distortion test is THE TEST everything is much simpler
Because then i would buy one sim SW and changing the parameter for just one bjt until i find a circuit with low distortion that then i will try to build
It could be the case that just a one bjt line stage optimized could have low enough distortion, at least for me
And i would leave with humble pleasure more complex topologies to talented designers

And moreover i think that to start a circuit with just one bjt is the best
Even at school they start with simple circuits I guess
If one is not able to understand this kind of circuit is better not to start completely
But still my main question is about the common test used to assess the linearity
Just think that I thought it were the frequency response test
Thanks again
Kind regards,
gino

Last edited by ginetto61; 24th March 2013 at 05:00 PM.
  Reply With Quote
Old 24th March 2013, 05:44 PM   #25
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Linearity (or lack of it) is a circuit property. Distortion is the result of what non-linearity does to a signal. It can be measured by looking for harmonics or intermodulation. Distortion (almost) always varies with signal level, so a useful distortion measurement includes more than one signal level. Distortion often, but not always, increases with larger signals.

Frequency response is something different. It is caused by filters, intended or unavoidable. Both filtering and non-linearity can change the shape of a waveform, but on the latter causes distortion. People sometimes find this confusing, and claim that a filter has 'distorted' a signal. At some point in the ensuing discussion it usually emerges that they either don't understand, or don't believe, Fourier theory.

All of this is standard textbook stuff.
  Reply With Quote
Old 24th March 2013, 06:07 PM   #26
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Quote:
Originally Posted by DF96 View Post
Linearity (or lack of it) is a circuit property. Distortion is the result of what non-linearity does to a signal.
It can be measured by looking for harmonics or intermodulation.
Distortion (almost) always varies with signal level, so a useful distortion measurement includes more than one signal level.
Distortion often, but not always, increases with larger signals.
It could seem weird to you but this for me is extremely important, and i thank you sincerely
From day one i heard of this damned linearity and all the circuital tricks to improved the low intrinsic linearity of transistors
But i was confused about which test shows the non linearity
For me this becomes now a cornerstone, a reference to judge those circuital tricks
What i have found always at least strange of this distortion spectra is the fact that i send a single tone in the input and at the output i get the armonics
Other people by the way are very much less worried by this fact
So maybe it is a my problem
Anyway thank you. This was my dilemma

Quote:
Frequency response is something different. It is caused by filters, intended or unavoidable. Both filtering and non-linearity can change the shape of a waveform, but on the latter causes distortion.
People sometimes find this confusing, and claim that a filter has 'distorted' a signal.
At some point in the ensuing discussion it usually emerges that they either don't understand, or don't believe, Fourier theory.
All of this is standard textbook stuff.
Perfect ! no more annoying questions
I will try to understand something from the book recommended me here
By the way i like very much the idea of sim SWs
My feeling is that they are an exceptional tool and a great point to start
If i am not wrong i can change parts in the circuit and the SW automatically can give a distortion spectrum,
This is for me now that i understand the importance of this test is amazing. I am honest.

I have just an observation
When i look at this distortion graphs on Stereophile I see that they are measured with pure resistive load (600, 10k, 100kohm etc.)
But from the spec of some amps i see also some capacitance at the input, in some case even more than 500pF
I wonder if this capacitance would impat on the distortion performances of a preamp
They should carry out the tests with a though load, like 10k+1000pF
Just to be on the safe side
Strangely i see this tests often shown in the datasheets of op-amps
Thank you sincerely
Kindest regards,
gino

Last edited by ginetto61; 24th March 2013 at 06:10 PM.
  Reply With Quote
Old 24th March 2013, 06:51 PM   #27
diyAudio Member
 
Join Date: Feb 2009
Location: UK
Quote:
Originally Posted by ginetto61 View Post
What i have found always at least strange of this distortion spectra is the fact that i send a single tone in the input and at the output i get the armonics
This is simply the result of representing a waveform in terms of Fourier analysis.

Quote:
In mathematics, Fourier analysis is the study of the way general functions may be represented or approximated by sums of simpler trigonometric functions.
This says that a repeating waveform may be synthesised by adding together other waveforms at the correct frequencies, amplitude and phase. It can be shown that any waveform can be represented by a sum of sinusoids at multiples of the fundamental frequency i.e. sinusoids are elemental components of any repeating waveform.

If your circuit has a tendency to clip, then its output can no longer be represented by a single sinusoidal waveform. However, by Fourier synthesis you could sum sinusoidal waveforms at multiples of the fundamental frequency (i.e. harmonics), at the correct amplitudes and phase, and arrive at the clipped waveform exactly. Fourier analysis is the opposite of Fourier synthesis, and automatically finds the amplitudes and phases of these harmonics from the clipped waveform. The energy of these harmonics vs. that of the (desired) fundamental at the amplifier's output is the THD measure, basically.
  Reply With Quote
Old 25th March 2013, 07:59 AM   #28
diyAudio Member
 
Join Date: Sep 2009
Location: Stavanger (NORWAY)
Quote:
Originally Posted by CopperTop View Post
This is simply the result of representing a waveform in terms of Fourier analysis.
This says that a repeating waveform may be synthesised by adding together other waveforms at the correct frequencies, amplitude and phase. It can be shown that any waveform can be represented by a sum of sinusoids at multiples of the fundamental frequency i.e. sinusoids are elemental components of any repeating waveform.
If your circuit has a tendency to clip, then its output can no longer be represented by a single sinusoidal waveform.
However, by Fourier synthesis you could sum sinusoidal waveforms at multiples of the fundamental frequency (i.e. harmonics), at the correct amplitudes and phase, and arrive at the clipped waveform exactly.
Fourier analysis is the opposite of Fourier synthesis, and automatically finds the amplitudes and phases of these harmonics from the clipped waveform.
The energy of these harmonics vs. that of the (desired) fundamental at the amplifier's output is the THD measure, basically.
Hello ! I am sorry but this is both extremely interesting and completely beyond my ability to understand
You mean that the distortion of a device can be predicted with math without measuring it ?
This is really fascinating.
I do not think that i would be able to calculate even a basic circuit
But i think i could be able to use one of these sim softwares
From what i understand a lot of audio designers simulate their designs before actually build them.
This at least for me could be a good point to start.
Simulate a very basic circuit, try to optimize it with the sim SW and then build it and ... listen
It would be better to measure it before listen anyway
I think I am done for now.
Minimalism and linearity ... now i have good answers to both this questions.
Thank you very much again.
Have a nice day !
Kind regards,
gino
  Reply With Quote
Old 25th March 2013, 11:48 AM   #29
diyAudio Member
 
Join Date: Feb 2009
Location: UK
Quote:
Originally Posted by ginetto61 View Post
Hello ! I am sorry but this is both extremely interesting and completely beyond my ability to understand
You mean that the distortion of a device can be predicted with math without measuring it ?
Hi ginetto

The main point was in answer to your question about how a sinusoidal input to a distorting amplifier circuit can give rise to multiple 'frequencies' at the output. They are only separate frequencies in the sense that they are harmonics i.e. integer multiples of the incoming frequency and locked in phase, so they do not sound like multiple signals, merely a change in the timbre of the waveform. Fourier analysis of the measured output would show that a sine wave clipped by an amplifier contains a number of harmonics, but on an oscilloscope it would show as a fixed repeating waveform - which it is. A computer simulation of the action of the amplifier based on the components' known characteristics would, indeed, predict the clipping, and thus the THD, without needing to measure it.
  Reply With Quote
Old 25th March 2013, 12:28 PM   #30
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Quote:
Originally Posted by ginetto61
You mean that the distortion of a device can be predicted with math without measuring it ?
Audio electronics is a branch of electronics; in fact one of the simpler branches. Electronics is a branch of electrical engineering. Electrical engineering is a branch of applied physics. In physics understanding involves being able to make calculations using mathematics. Therefore circuit understanding involves being able to estimate its properties, including distortion. I say 'estimate' because we sometimes do not know the required parameters for a particular device in enough detail to make an accurate prediction. Fortunately audio design is usually sufficiently noncritical that an estimate is good enough.

So a circuit designer should be able to predict, roughly, things like gain, frequency response and distortion before he measures or simulates. The alternative is blindly copying others, or random fiddling (in reality or on a computer) until the circuit comes out right. Some people call this 'design', but it is not design in any serious engineering sense of the word.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Some questions about amplifier classes, distortion and efficiency Plecto Solid State 27 15th January 2013 09:00 AM
Yamaha's kind of ERCO: Zero Distortion Rule "ZDR" (M50, M70, A520) tiefbassuebertr Solid State 3 28th October 2010 08:44 AM
Marshall MS-2 as distortion pedal questions. Black Mark UK Chip Amps 7 30th June 2009 06:10 PM
distortion due to Miller non linearity David Lewis Solid State 0 8th April 2005 06:55 PM
relation between ampere, voltage and distortion or linearity oliverniekamp Pass Labs 4 17th June 2002 09:38 PM


New To Site? Need Help?

All times are GMT. The time now is 08:06 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2