"What's your reasoning?" and not "What's your belief?".

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
john curl said:
Folks, Walt Jung is working to putting Barrie Gilbert's article: 'Are Op Amps Really Linear?' on Walt's Website. By working together, Walt and I are reconstructing the entire article for people to read for themselves.

Excellent! Thanks John and Walt. It was a real shame that the article was removed from the EDN site, so I'm glad it will be back on the web again. Given the differing interpretations of Gilbert's results, I'm sure it will lead to lots more lively discussion.
 
our conclusion is that negative feedback is problematic
Feedback does "hall of mirror" effect of non-zero errors produced by any stage (especially in final stage), back to differential, re-amplifies it to final stage, then back to differential, then goes to the final stage again, and so on. Producing high order distortions, Mr. JC said there is somekind of audio amplifier distortions up to FM frequencies.

What about local feedback (adjusted gain) in every audio power amp stage. Does it better than global feedback, since no global feedback does not generate "hall of mirror" distortion repetition?
 
I'm interested in that article, John.

Some explanations might help for visitors to this thread. Amplitude Modulation (AM) and Frequency modulation (FM) and Phase Modulation (PM) have their origins in radio transmission. They are methods for encoding an RF carrier signal with an information signal, usually audio, but could be anything. Normally, the information signal is of much, much lower frequency range than the carrier.

In AM, the instantaneous amplitude of the carrier is proportional to the amplitude of the signal. The signal amplitude is less than the carrier amplitude - otherwise you get "over-modulation" which corrupts the signal. The spectral content includes the carrier and two "sidebands" for each frequency in the signal, at the carrier plus and minus the signal frequency. The amplitude of the sideband spectra depends upon the amplitude of the signal.

In FM, the instantaneous frequency of the RF carrier is proportional to the amplitude of the signal. The carrier amplitude is constant. FM is good for radio because environmental noise almost entirely causes amplitude errors and the FM detector can be made insensitive to amplitude changes. This is why FM sounds quieter than AM. FM spectra includes sidebands, like AM, except instead of their being one pair for each signal frequency there are multiple sidebands at equal intervals. The spread of sidebands is a complex relationship between the signal frequency and the deviation of the carrier cause by the signal. Simply, an FM signal produces a spread of multiple sidebands rather than just a pair as in AM.

In PM, the instantaneous pahse of the RF carrier is proportional to eht amplitude of the signal. PM and FM are inter-related as any change of phase requires an temporary change of frequency and vice-versa. PM is used rather than FM because it is easier to have a highly accurate and stable carrier oscillator and then implement the PM in a gain stage rather than designing an oscillator that is both accurate and stable and variable in frequency. Like FM, a spread of sidebands is created.

In broadcast FM the stations need to be very careful about the spread of the sidebands so as not to cause intereference with nearby stations. Since the sideband spread is related to signal frequency and amplitude range, the stations must limit these and monitor the sidebands continually to ensure no breach is made of their allocated bandwidth window. This also leads to use of compression and bandwidth limiting techniques which can make music sound particularly sanitized.

In an audio amp there isn't FM going on as such. But non-linearities will cause a spread of distortion spectra to appear, some of which resemble AM, FM or PM.
 
traderbam said:
Vam = sin(2.pi.Fc.t).(1 + Vs)

Vfm = sin(2.pi.Fc.(1 + Vs).t)

Vpm = sin(2.pi.Fc.t + Vs)

where,
Fc is the carrier frequency
Vs is the signal voltage

Yep, that's my point !

Vpm = sin(2.pi.Fc.t + Vs)
Now i set Vs = carrierfreq 40%, gives following: (except the cos)

cos(2*pi*t*f + sin(2*pi*t*f)*0.4)

Looks like attachement,
A classical 2nd harmonic, this time 180°...

Mike
 

Attachments

  • pm.gif
    pm.gif
    2.7 KB · Views: 178
Mikek swiftly dodges (for the umpteenth time) another direct question when asked about his practical experience.

This is then followed by an ill-aimed salvo from the moderators.

I can't see any personal stuff between traderbam and mikek in this thread - where are you looking?

Andy.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.