Boris_The_Blade said:
Just you anticipate me before make the same questions as you.
I have to add only one remark. There is an antagonism between THD and IMD; The question it is that: which of the two distortions is more valuable? According to my books the measurement of IMD it is more valuable because its product there is in the audible spectrum from human ears; the IMD calculation according to SMPTE standards -which give the worst results from other standards such as CCIF- it is executed by applying in the input of DUT a mix of two sinus, 60HZ/1Vpp and 7KHZ/0,25Vpp. The difference between these two fundamental frequencies presented in the output of DUT which expressed as, and only, eiter the 2nd or 3rd order factor harmonics - produced by the fundamental frequencies- amplitude sum divided by the amplitude of the high frequency it is the IMD percentage while the rest harmonics can be omitted because they are outside from the region of measurement (and the region of audible spectrum). In conclusion so the THD as the IMD caused from the nonlinearity of the amplifier. Yet the cheap or free PC analysers offers the measurement of these two distortions. But the TIM? This distortion related with the slew rate (or rise time) mainly and its percentage is dependent from the errors of whole the feedback loop. And it is of most importance as i think me at least, because it represents the audible dynamic characterisics of the amplifier. It has two main poles, the low frequency and the high frequency. Also there are many nested and well hiden poles inside the main endmost frequency poles which are impossible to summarized in a total sum by any expensive instrument exists. There is also an antagonism between the TIM and the THD; by improving the one it worsen the other and vice-versa.
In practice now, i think that these all free PC visual instruments (such RMAA) are useless because they add in the plots the significant amount of the sound card's distortions. Moreover, the parts offered today (semiconductors, resistors, caps etc) are of so big quality in relation with they offered before 5 to 10 years and back, thus only by making a gross mistake in the design or in the implementation of the circuit there is the possibility of unacceptable THD or IMD presentation. So it remains the job of managing the TIM in a such manner to be in equilibrium with the THD and IMD in conjuction with the better as it is possible transient response of the amplifier.
The ways of estimating the TIM, the rise time (or the slew rate if you preffer) was analyzed up to death in many threads in this forum, and i think there is not need to proceed in their description furthermore.
Fotios
Everyone has their opinion, but I feel I can tell a difference with amps that distort less than others.
IMO and experience........
Amps with a stronger output drive sound the best. The bass is stronger, and tighter with less boominess, and the high trebles come out clearer and louder.
On my 350W subwoofer amp, the output drive is overbuilt. Even when driven with a flat signal with no EQ, the amp sounds fairly bassy when playing full range. Even better, the trebles come out strong and clear. Under normal use, I only use it for lows for a sub, but the amp is plenty capable of full range.
Other amps I've owned and built with weaker drive, get muddy bass, and duller trebles. The bass difference is more noticable than the treble, but I can definately tell a difference.
IMO THD DOES Matter, mainly for treble clarity.
in my Sub amp, I changed the preamp quad-opamp from a upc324 to an On-Semi MC33079 and the difference was VERY noticeable! Solid bass, and perfect treble. The MC33079 has 0.002% THD. The difference was like comparing low bitrate to high bitrate Mp3's.
IMO and experience........
Amps with a stronger output drive sound the best. The bass is stronger, and tighter with less boominess, and the high trebles come out clearer and louder.
On my 350W subwoofer amp, the output drive is overbuilt. Even when driven with a flat signal with no EQ, the amp sounds fairly bassy when playing full range. Even better, the trebles come out strong and clear. Under normal use, I only use it for lows for a sub, but the amp is plenty capable of full range.
Other amps I've owned and built with weaker drive, get muddy bass, and duller trebles. The bass difference is more noticable than the treble, but I can definately tell a difference.
IMO THD DOES Matter, mainly for treble clarity.
in my Sub amp, I changed the preamp quad-opamp from a upc324 to an On-Semi MC33079 and the difference was VERY noticeable! Solid bass, and perfect treble. The MC33079 has 0.002% THD. The difference was like comparing low bitrate to high bitrate Mp3's.
nitrate said:
First of all: who is in position to guaranteed that a class A amp. it is more remarkable than a class AB to drive a pair of speakers?
Do you want my oppinion? A class A implementation mainly addressed to remove absolutelly the cross distortion and thus to improve the THD. For this purpose it uses large ammounts of energy produced by powerfull (but with low voltage level) supplies. IMHO the most of energy translated in heat and the less for driving the load. And these all damned Hi-End loudspeakers are very heavy loads. Although its most active parts such the speakers are relativelly easy loads, the rest such the box, the x-overs, the damping materials etc make finally this load very heavy to drive. Although i am not busy with tube amplifiers, i believe that they are prefferable in such situations instead a class A solid state unit. And this owed mainly in their tubes brutal force which is not spended in heat.
Second: The D. Self book it is not more than a collection from well known topologies already and some curious ideas of the author based mainly on his (and a dream of all of us) Audio Precision analyser. Maybe from his enthusiasm, caused from this instrument, he proceeded in various approaches from which the only important and practical was the EF in the VAS stage. John L. Hood he had made many of such propositions enough years before. An example it is the CFP output examination. Yes, it is by far most responsible in bandwidth in relation with EF output but in practice (i made also me my experiments) it is dangerous in supply levels from +/-40V and above to
.. So, with a such topology we can obtain as much a power about 80W/8Ù. This power it is enough for my bookshelf Mission 780 speakers. Third: The secret (according to my experiments) it is that. The supply level it is of bigger importance. A level of +/-60V at least it is appropriate. Of course we must forget a class A amplifier in such voltage level. Also a CFP it is continuously under the danger of thermal runaway in a such voltage level. Thus, by applying the subtractive method (the method of Socrates, because except philosopher was a big psychologist in ancient Greece) we resulted in the EF class AB amplifier. It is so simple.
As for the rest, the thermal compensated CCS and the current mirrors are usefull for the stability. I heat personaly the double diode polarized CCS. And the large electrolytic bootstrap capacitors. These big power (due to their high supply voltage and the big power transformers accordingly used) amplifiers are amazing in reproduction due to their almost zeroed output impedance owed to large number of output transistors that comprise, are in possition to send the cone of any Hi-Fi speaker out of the magnet assembly without any heared distortion. I remembered a fany event, when i had some Crest and Peavey P.A. amplifiers in my stock, with one friend. He had a pair of heavy Spendor loudspeakers and he tried to listen music by driving these with a NAD 2 X 100W and he was disapointed because its error selection. Take this Crest i said to him to check your speakers. If i was not with him during the "audition tests "
sure from his enthusiasm the cones of speakers will catapulted on the opposite wall without they burned. The speakers burned from the distortion in mid-high frequencies, even in low audition levels, and not from clear power. From this simply the cones catapulted 
Fotios
That is exactly what i'm getting at!! <<<@ Andy L
So it seems i should try and measure the IMD befor i condemm this amp as sounding bad. It may show up a measurable discrepancy that i can begin to hunt down. If i cant see whats wrong then i cant really fix it. Just like Andy L said befor, i'm having the same phenominon were a throw together amp built to the lowest standards ( also similar to a JLH amp ) sounds great in comparison even though on paper its under powered and of high distortion.
So, who's gonna volunteer to give me a crash course in IMD distortion measurement then?? LOL
Leigh
So it seems i should try and measure the IMD befor i condemm this amp as sounding bad. It may show up a measurable discrepancy that i can begin to hunt down. If i cant see whats wrong then i cant really fix it. Just like Andy L said befor, i'm having the same phenominon were a throw together amp built to the lowest standards ( also similar to a JLH amp ) sounds great in comparison even though on paper its under powered and of high distortion.
So, who's gonna volunteer to give me a crash course in IMD distortion measurement then?? LOL
Leigh
For what it's worth, IMO the way the amp reacts to having what is in reality a "low impedance signal generator", ie the speaker, hung on it's output plays a large part. Also high damping factor isn't nearly as important as claimed for good sound (My opinion ). Any chance of you posting the circuit ?
Regards Karl
). Any chance of you posting the circuit ?Regards Karl
If you have the old version of RMAA you can set whatever tones you like for the IM measurement. The latest version I think have a fixed setting with 60Hz + 7kHz or something like that. Possibbly the latest pro/licensed version allows custom settings.
As an alternative you can download ARTA and use that for IM measurement and multitone measurements.
/Peter
As an alternative you can download ARTA and use that for IM measurement and multitone measurements.
/Peter
I'm at work at the mo but as soon as i get home i'll post the circuit. I must say though its not just this particular design i'm probing, if its a fact that low THD does not automatically give good performance then i want to know why. It's not about improving my amp design at all. I'll post the circuit tho then you can all see why i'm at this dilema.
Leigh
Leigh
If we knew the answer to that I think a lot comes down to how we perceive sound. Turning the argument round slightly, why does a musician prefer an old instrument over a modern equivalent ? The modern equivalents, computer designed and all, are technically superior but as for that elusive sound, the sound that makes the hairs on your neck stand up, we've a long way to go. I have found it the same with audio.
Regards Karl
I think a lot comes down to how we perceive sound. Turning the argument round slightly, why does a musician prefer an old instrument over a modern equivalent ? The modern equivalents, computer designed and all, are technically superior but as for that elusive sound, the sound that makes the hairs on your neck stand up, we've a long way to go. I have found it the same with audio. Regards Karl
Pros and Cons of THD measurement
I have posted a summary of what I believe to be the value and shortcomings of THD measurements over in the permanent negative feedback thread.
I've tried to explain why it is desirable to have very low levels of THD. I have also tried to cite the numerous sources of sonic degradation that may escape a low THD measurement. THD is neither necessary nor sufficient to achieving excellent sonics; so why is THD a valuable tool? That is the sort of question I've tried to answer.
Cheers,
Bob
I have posted a summary of what I believe to be the value and shortcomings of THD measurements over in the permanent negative feedback thread.
I've tried to explain why it is desirable to have very low levels of THD. I have also tried to cite the numerous sources of sonic degradation that may escape a low THD measurement. THD is neither necessary nor sufficient to achieving excellent sonics; so why is THD a valuable tool? That is the sort of question I've tried to answer.
Cheers,
Bob
hi nitrate
hi nitrate..i didnt read all of the posts but i can tell something ..i build output capacitorless class a headphone amplifier a week ago..i used opamps that have very very low thd..i used lm4562.it has nearly no thd.i listened to it..the sound is very clear and detailed but i thought that there is something wrong..i replaced opa2134 which has more thd than lm4562..but the sound is better now!and now i think that detailed and pure sound is not always good for my ears..and low slew rate amp maybe can sound better?
here is the schematic of my amp
hi nitrate..i didnt read all of the posts but i can tell something ..i build output capacitorless class a headphone amplifier a week ago..i used opamps that have very very low thd..i used lm4562.it has nearly no thd.i listened to it..the sound is very clear and detailed but i thought that there is something wrong..i replaced opa2134 which has more thd than lm4562..but the sound is better now!and now i think that detailed and pure sound is not always good for my ears..and low slew rate amp maybe can sound better?
here is the schematic of my amp
Attachments
As requested here is the offending circuit.
The Amps origins are to be found here:-
http://www.diyaudio.com/forums/showthread.php?postid=1407918#post1407918
It seems that the more NFB is applied the greater the drop in THD but at the same time the amplifier seems to sound less transparent. I mean the amp does not sound 'bad' but it isnt up to my expectations either!!
Leigh
The Amps origins are to be found here:-
http://www.diyaudio.com/forums/showthread.php?postid=1407918#post1407918
It seems that the more NFB is applied the greater the drop in THD but at the same time the amplifier seems to sound less transparent. I mean the amp does not sound 'bad' but it isnt up to my expectations either!!
Leigh
Attachments
AAMOI why did you opt for a quasi complementary output?. Possibly getting bipolar gain droop, the VAS stge is set to 13mA and the TIP41 has a gain of as low 30 and the MJL21193 25 so it could possibly supply 10 amps max also the ft the TIP41 is only 4MHz which wont help. I have played, about thirteen years ago!, with a Self type topology except it used FETS and had Hawksfords error correction circuitry, the results were very very good. If think the design is in here somewhere, if not I can dig it out. Any questions on error correction should be referred to Mr Cordell .
.
Re: Pros and Cons of THD measurement
Mr. Cordell
With each respect to you, i maintained nearly the same in my post #21:
"""I have to add only one remark. There is an antagonism between THD and IMD; The question it is that: which of the two distortions is more valuable? According to my books the measurement of IMD it is more valuable because its product there is in the audible spectrum from human ears; the IMD calculation according to SMPTE standards -which give the worst results from other standards such as CCIF- it is executed by applying in the input of DUT a mix of two sinus, 60HZ/1Vpp and 7KHZ/0,25Vpp. The difference between these two fundamental frequencies presented in the output of DUT which expressed as, and only, eiter the 2nd or 3rd order factor harmonics - produced by the fundamental frequencies- amplitude sum divided by the amplitude of the high frequency it is the IMD percentage while the rest harmonics can be omitted because they are outside from the region of measurement (and the region of audible spectrum). In conclusion so the THD as the IMD caused from the nonlinearity of the amplifier."""
Our opinions (from experiments) they converge almost and thank you very much for your confirmation.
Fotios
Bob Cordell said:
Mr. Cordell
With each respect to you, i maintained nearly the same in my post #21:
"""I have to add only one remark. There is an antagonism between THD and IMD; The question it is that: which of the two distortions is more valuable? According to my books the measurement of IMD it is more valuable because its product there is in the audible spectrum from human ears; the IMD calculation according to SMPTE standards -which give the worst results from other standards such as CCIF- it is executed by applying in the input of DUT a mix of two sinus, 60HZ/1Vpp and 7KHZ/0,25Vpp. The difference between these two fundamental frequencies presented in the output of DUT which expressed as, and only, eiter the 2nd or 3rd order factor harmonics - produced by the fundamental frequencies- amplitude sum divided by the amplitude of the high frequency it is the IMD percentage while the rest harmonics can be omitted because they are outside from the region of measurement (and the region of audible spectrum). In conclusion so the THD as the IMD caused from the nonlinearity of the amplifier."""
Our opinions (from experiments) they converge almost and thank you very much for your confirmation.
Fotios
nitrate said:
Leigh, realy the circuit it is an uncommon implementation. Because in electronics everything it is under revision forever, i can't express my opinion. Before 5 years i heared a Hi-End class amplifier of Italian origin with quasi compl. output with wonderful reproduction.
Thus the only remark that i have it is the strange mix of misfit between them transistors. Why you use these hi Vce transistors although the voltage supply level of your project it is so low? Everyone, yet the most newcomer, he knows that the higher the Vce the lower the beta (bandwidth) of a transistor. In so low supply you can use transistors with a beta of some hundreds which accordingly helps in the improvement of the transient response. That is exactly what you don't like in hearing.
Also the circuit it does not say nothing by alone and by the simulation. Can you please tell us how much it is the Rise Time by applying a square wave of 10KHz from a real generator in input and measured in output with a real and not a virtual oscilloscope?
Fotios
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.