diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Cathode Follower, where is the 3d H gone? (http://www.diyaudio.com/forums/tubes-valves/215705-cathode-follower-where-3d-h-gone.html)

Mosquito 6th July 2012 01:26 AM

Cathode Follower, where is the 3d H gone?
 
3 Attachment(s)
Hi gents
I was playin on LTSpice around the idea of a tube line amplifier, capable of driving a solid state amp trough some long cables,
While working at it, I found that this one, not optimally biased, exhibits an FFT behaviour that seems to me, somewhat curious,
It is a simulation error, or such a circuit can perform that way in the real world?

TheGimp 6th July 2012 04:03 AM

I've seen harmonic cancellation in multi-stage amps before. They simulate and measure nearly the same.

Harmonic content varies with bias point. If you get two stages biased at the optimal (but possibly different) points they may do considerable harmonic cancellation. However, it usually is very configuration specific.

revintage 6th July 2012 05:55 AM

1 Attachment(s)
Hey M,

Error due to to short time!

.tran 0 100m 60m 1u

This is what you should expect when doing a proper FFT:

powerflux 6th July 2012 06:49 AM

indeed. do not absolutely trust the LT FFT analysis. it's a great tool, but toying with time-steps, simulation time, compression and other options can give wildly different graphs for the same circuit. Always measure real-world performance and if there should be a 3rd harmonic, there will be.

DF96 6th July 2012 11:48 AM

A 'negative' amount of 3rd should have warned you that the FFT was in trouble. FFT does lots of calculations on samples of your waveform. Small numbers coming out of the FFT are the result of lots of cancellations of big numbers, so more likely to be be meaningless than bigger numbers (e.g. fundamental) if you have insufficient data, wrong windowing etc.

revintage 6th July 2012 01:23 PM

Hey,
Donīt complicate this to much, its only a matter of knowing how to use a program. Mosquito didnīt know how to use the LTSpice FFT, thats all. With the .tran commando I gave, it will work like intended.

DF96 6th July 2012 01:37 PM

Simply following your advice may make it work, but he still won't know why it makes it work. I generally find understanding is a useful thing. I deliberately kept my comments very simple - did I mention orthonormal basis sets for inner-product spaces?

piano3 6th July 2012 02:26 PM

in infinite dimensions?

DF96 6th July 2012 02:35 PM

Finite dimensions for the FFT. Infinite for a real Fourier analysis.

Mosquito 6th July 2012 03:35 PM

Hi again
Many thanks You all for the sharp comments.
In fact looking at the first image , the small ammount of negative 3dH must be warned me about somethig gone wrong. Lesson learned.
LTspice , and particulary the FFT analysis, are powerfull tools, but if the user isn't completely aware about what he is doing, it's somewhat like give a razor blade to a monkey... :-))
Again thanks Lars, DF96 and the others for the useful info and tips. I'm here just to learn.


All times are GMT. The time now is 09:12 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright Đ1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2