perfeto !!!!!
a simple question becoms a titan fight ......
which i think that bring us to the final conclusion which that many of the amplifiers presented in the forum have oytstanding performance but will not stand any abuse .....
then again the diference between comercial amps and diy amps is the protection methods like zobel, inductor, VI limmiters and so on ....
then the truth is somewhere in the midle ....you canot have everything .....
i wonder how most of these amps will perform with abuse .....
but thats me
a simple question becoms a titan fight ......
which i think that bring us to the final conclusion which that many of the amplifiers presented in the forum have oytstanding performance but will not stand any abuse .....
then again the diference between comercial amps and diy amps is the protection methods like zobel, inductor, VI limmiters and so on ....
then the truth is somewhere in the midle ....you canot have everything .....
i wonder how most of these amps will perform with abuse .....
but thats me
Re: Re: Function of Output Inductor
First, high-end solid state amps without NFB (no matter local, global or both) simply don't exist.
But even an amp with only error correction in the OP stage (like HEC) and no global NFB at all, needs a coil if loaded with a low ESR cap.
One more thing: please stop bashing "high amount of global negative feedback", by calling it a band aid. Such nonsense hurts your credibility.
PMA said:
First, high-end solid state amps without NFB (no matter local, global or both) simply don't exist.
But even an amp with only error correction in the OP stage (like HEC) and no global NFB at all, needs a coil if loaded with a low ESR cap.
One more thing: please stop bashing "high amount of global negative feedback", by calling it a band aid. Such nonsense hurts your credibility.
Re: perfeto !!!!!
-----------------------------
The Lab Workers/Designers at National Semiconductors, with the simple tools they posses
may not know and understand as much of amplifiers as we do.
But anyway, each time some great audio designer or whoever at our forum critizise these employees writing all them datasheets for us, It makes me doubt some of the things posted here.
My own opinion on this is clear:
If left with the choice
1. --- to believe in some Official Published data by those that made them amplifiers
and
2. --- some stuff posted by a 'moniker' or even 'guru' telling me otherwise in this forum
I have often no problems to make up my mind who to trust.
At least in the FIRST Place.
For example,
When Scott Wurcer gives a good piece of info on AD797,
I will listen carefully.
At least I would be reluctant to argue very much with him.
He was around when the chip was designed, tested and the datasheet public information written.
========================
=============================================
LM3875.
What do they recommend for such an output.
No doubt Overture series of 'gainclones chips' are among the most used ever Power Amplifiers of all times.
Why?
Because they please the users.
If not, they would not be such a sucess.
... there are others making such power amps .. with less sold units.
It might be interesting to see what The National Technical Staff
consider a fair typical Output for speakers.
And compare to what 'our' best & most qualified forum posters say.
See my attachment from LM3875 datasheet
Regards Lineup
The more they know, the less they know about them selves & their own egos. At least at times looks so, in my 'humble' opinion.sakis said:
-----------------------------
The Lab Workers/Designers at National Semiconductors, with the simple tools they posses
may not know and understand as much of amplifiers as we do.
But anyway, each time some great audio designer or whoever at our forum critizise these employees writing all them datasheets for us, It makes me doubt some of the things posted here.
My own opinion on this is clear:
If left with the choice
1. --- to believe in some Official Published data by those that made them amplifiers
and
2. --- some stuff posted by a 'moniker' or even 'guru' telling me otherwise in this forum
I have often no problems to make up my mind who to trust.At least in the FIRST Place.
For example,
When Scott Wurcer gives a good piece of info on AD797,
I will listen carefully.
At least I would be reluctant to argue very much with him.
He was around when the chip was designed, tested and the datasheet public information written.
========================
=============================================
LM3875.
What do they recommend for such an output.
No doubt Overture series of 'gainclones chips' are among the most used ever Power Amplifiers of all times.
Why?
Because they please the users.
If not, they would not be such a sucess.
... there are others making such power amps .. with less sold units.
It might be interesting to see what The National Technical Staff
consider a fair typical Output for speakers.
And compare to what 'our' best & most qualified forum posters say.
See my attachment from LM3875 datasheet
Regards Lineup
Attachments
Hi Lineup,
As usual, most published Zoebel networks are wrong, including this one from NS.
If the load is 8 Ohm, then the R across the coil should also be 8 Ohm.
Furthermore, a parallel cap across the load is missing.
With 0.7uH and 8 Ohm the correct value is ~11nF.
Cheers,
Edmond.
PS: Zoebel rules!
As usual, most published Zoebel networks are wrong, including this one from NS.
If the load is 8 Ohm, then the R across the coil should also be 8 Ohm.
Furthermore, a parallel cap across the load is missing.
With 0.7uH and 8 Ohm the correct value is ~11nF.
Cheers,
Edmond.
PS: Zoebel rules!
For the illiterate, 'oe' stands for 'ő' (o-umlaut). As this character is not always correct displayed on other monitors, I prefer to be on the safe side, and use 'oe', which is the same as 'ő' in German.
BTW, traderbam, you really don't know much, do ya? Even a *.rar file was a mystery for you.
Cheers.
BTW, traderbam, you really don't know much, do ya? Even a *.rar file was a mystery for you.
Cheers.
Edmond Stuart said:
If you want to be on the safe side you should perhaps spell his name correctly. His name was Zobel with a plain ordinary "o", without any omlaut or other ornaments. Since he was born in the USA, we can rule out that he originally spelt his name differently and changed it after immigrating, since he neved immigrated.
traderbam said:You cast an aspersion oen National's circuit, Edmoend. The least yoeu can doe is explain why. Peoeple are trying toe learn soemething here.
Regrettable, I've to repeat my myself: you really don't know much, do ya? Even a *.rar file or a Zoebel network is a mystery for you.
BTW, do you really think you're funny with your stupid "oe", or are you just stoned.
Zőbel
Hi Christer,
Hmm...In the past I had a long discussion with Ovidiu (syn08) on this topic and he assured me (with good arguments) that it is Zőbel. Apparently, he is a purist, nothing wrong with that.
Cheers,
Edmond.
BTW, what about installing a spelling checker on your browser.
Christer said:
Hi Christer,
Hmm...In the past I had a long discussion with Ovidiu (syn08) on this topic and he assured me (with good arguments) that it is Zőbel. Apparently, he is a purist, nothing wrong with that.
Cheers,
Edmond.
BTW, what about installing a spelling checker on your browser.
Re: Zőbel
His name is obviously of european origin, so if your post showed the right character on my computer, it seems like hungarian origin then, not german (language) origin as I thought. Of course it is possible that his family kept the original spelling of the name even after immigrating, and then Otto Julius Zobel himself used to simplify it. He obviously did spell it with a plain "o" himself, though:
http://en.wikipedia.org/wiki/Otto_Julius_Zobel
Edmond Stuart said:
His name is obviously of european origin, so if your post showed the right character on my computer, it seems like hungarian origin then, not german (language) origin as I thought. Of course it is possible that his family kept the original spelling of the name even after immigrating, and then Otto Julius Zobel himself used to simplify it. He obviously did spell it with a plain "o" himself, though:
http://en.wikipedia.org/wiki/Otto_Julius_Zobel
The following simulations may help some people understand why the RL network, in addition to the RC, guarantees maximum reliability and stability for any amplifier and with any load.
It's quite easy to find a range of RLC loads capable of bringing *any* amplifier without a LR output network into oscillation and self destruction, but an amplifier with the proper RC and RL output networks is unconditionally stable and can't be disturbed by any load.
I have modelled the load impedance and its phase, as seen from the output stage, for various load capacitances. I have taken the output network of the LM3886 example since it's reasonably well designed. Note how the impedance is nicely resistive and phase is always near 0 above 2Mhz, which is the minimum unity gain bandwidth product of the LM3886 (typical is 8Mhz) according to the datasheet.
This is the purpose of the RC and RL networks, keeping phase and gain controlled and load independent in the critical unity gain crossing region.
Note how very small capacitances like 3.3nF or 10nF reveal that the L and R value choices are a bit on the low side, while bigger capacitance values are the most harmless ones.
Have fun! With this information some people could even end up learning how to design stable amplifiers (after many years making unstable ones)
No ears were harmed during the writting of this post
It's quite easy to find a range of RLC loads capable of bringing *any* amplifier without a LR output network into oscillation and self destruction, but an amplifier with the proper RC and RL output networks is unconditionally stable and can't be disturbed by any load.
I have modelled the load impedance and its phase, as seen from the output stage, for various load capacitances. I have taken the output network of the LM3886 example since it's reasonably well designed. Note how the impedance is nicely resistive and phase is always near 0 above 2Mhz, which is the minimum unity gain bandwidth product of the LM3886 (typical is 8Mhz) according to the datasheet.
This is the purpose of the RC and RL networks, keeping phase and gain controlled and load independent in the critical unity gain crossing region.
Note how very small capacitances like 3.3nF or 10nF reveal that the L and R value choices are a bit on the low side, while bigger capacitance values are the most harmless ones.
Have fun! With this information some people could even end up learning how to design stable amplifiers (after many years making unstable ones)
No ears were harmed during the writting of this post