Lars,
I know you hate these discussion about what we exactly mean by which class, but I am the one that hates sloppy use of terms
so please bear with me.
PWM is just ONE specific class of digital amplifiers. You can also have Puls Repetition Modulation, where the puls width remains constant but the repetition freq varies. I haven't seen it in commercial audio, but have seen it in SMPS circuits. Presumably there is also a technology with alternating polarity pulses, varying in length to modulate the audio. Don't know the exact name, something with Biphase Modulation.
Jan Didden
I know you hate these discussion about what we exactly mean by which class, but I am the one that hates sloppy use of terms
so please bear with me.PWM is just ONE specific class of digital amplifiers. You can also have Puls Repetition Modulation, where the puls width remains constant but the repetition freq varies. I haven't seen it in commercial audio, but have seen it in SMPS circuits. Presumably there is also a technology with alternating polarity pulses, varying in length to modulate the audio. Don't know the exact name, something with Biphase Modulation.
Jan Didden
I agree with jannemann on the definition of PWM: A fixed frequency rectangular whose duty-cycle is varied.
So are you using a self-oscillating topology with either hysteresis or delay control (or both) ?
In this case what leads you to the assumption that there is only noise generated and no "real" harmonics ?
I know that delta-sigma modulation (i.e. one bit modulation with a fixed sampling rate) generates a "noise-floor" that is consisting of odd order harmonics of the input signal if this is consisting of a single sinusoidal signal. For anything else it looks much more complicated. It is interesting however that IMD seems to be quite low for delta-sigma modulation.
Regards
Charles
It all adds up to PWM, no matter if it is the number of pulses, the length, the time, delay, etc. etc. As long as you have only 1's and 0's to put out, the only way you can modulate the signal is by altering the (or modulating) the Pulse Width. Hence Pulse Width Modulation. Does'nt need a fixed frequency to fit the definition. ...
I agree with you two guys completely, the class term should not be used in a sloppy manner. Therefore i will refrain from even using any 'class' terms. Since there seems to be such widely spread confusion about the definition of each class.
Maybe one of you 'definition of terms' experts could make a list (in this thread) with the precise definition of each amplifier class.
However i admit there is one exception to the PWM / 1-0 rule, namely Frequency Modulation. In theory it would be possible (but not nessescarily beneficial) to make a frequency resonant mode controlled power amplifier. A parallel LC circuit is exited with a squarewave of various frequency. At it's resonance the p-p voltage goes up, at other freq's it goes down. So you can control the output voltage by controlling the frequency of the square wave. Anyone gonna build one of these ??
My topology in your terms would be:
Delay controlled.
(with some feedback filtering as used in DS converters)
But since the modulator is integrating (producing a triangular wave) you can also add an external clock signal to syncronize the amplifier to other amplifiers or the DAC.
In this case it is no longer free running, but locked to the clock generator. (Typically 300 kHz - 1 MHz).
Delay controlled.
(with some feedback filtering as used in DS converters)
But since the modulator is integrating (producing a triangular wave) you can also add an external clock signal to syncronize the amplifier to other amplifiers or the DAC.
In this case it is no longer free running, but locked to the clock generator. (Typically 300 kHz - 1 MHz).
Lars Clausen said:
What definition? your own? I think you might be confusing
terminology with definition. Usually the name of something
does not tell the whole truth, since it would be unbearably
long and complex. As I remember the definition, I agree with
Jan and Charles that PWM means fixed frequency with
varying duty cycle, and of course just two output values.
I don't think I have any textbook to refer to, or at least none
I can find for the moment. However, searching the web for
"pulse width modulation" seems to give a very coherent
picture. All documents I looked at that gave some form of
definition at all agreed with what I said above. The following
has the crispest definition.
http://controls.ame.nd.edu/~bill/microcontroller/main/node39.html
If you don't like maths, then just stick with what I said above.
Some documents also handled special cases like saturated
PWM, which seems not interesting for audio.
I think what you are after Lars might be some term that
encompasses all different technologies that has just two
output values, both pulse width modulation and pulse density
modulation and also other imaginable things like varying both
frequency and pulse length (guess there are no more
possibilities). I don't know if there is such a
term, but I agree such a term would be useful. If there is
already a term for this, someone will probably tell us soon,
and it would not be wise to introduce a new term if there is
already an established one.
Otherwise, what about just calling it a switching amplifier or
a binary amplifier?
Jocko: If i understand right Class G is switching between different power supply levels, but using ananalog power stage ... right?
You are right about the distortions in the "********" amplifiers
( don't know what to call them anymore ) Just some don't have distortions harmonic to the audio signal, but random noise instead.
You are right about the distortions in the "********" amplifiers
( don't know what to call them anymore
) Just some don't have distortions harmonic to the audio signal, but random noise instead.
Lars Clausen said::However i admit there is one exception to the PWM / 1-0 rule, namely Frequency Modulation. In theory it would be possible (but not nessescarily beneficial) to make a frequency resonant mode controlled power amplifier. A parallel LC circuit is exited with a squarewave of various frequency. At it's resonance the p-p voltage goes up, at other freq's it goes down. So you can control the output voltage by controlling the frequency of the square wave. Anyone gonna build one of these ??
I think this would be an interesting experiment, but the efficiency would not be good. The frequency would have to go very high above the base frequency where full power output would be obtained. Otherwise, the dynamic range would be poor. Maybe I am off on this view, but was interested in taking a stab at the idea
For more information on Class D amplifier technology, see: http://www.aes.org/sections/la/PastMeetings/2003-02-25.html.
BTW, very few Class D amps are actually digital. Most use totally analog means of modulation.
BTW, very few Class D amps are actually digital. Most use totally analog means of modulation.
hi.
i think that there are several good reasons for using class-d/pwm amplifiers in high power applications , but i also think that there is a lot of hype and that the use of pwm amplifies in hifi/home applications isnt all that obvious.
i have been involved in listening tests comparing 3 different types of pwm amplifiers and a "normal" amplifier too (a current feedback amp like the alexander amp)
no doubt the latter presented the best sound quality of these.
the pwm amplifiers involved were ; icepower , tripath and our own discrete pwm amp.
i do realize that different people hear different things , as an example;
> have now owned a pair of monoblocks based on Ice-power (semi-diy) and I think they are superior in certain areas. I would never buy anything else actually.>
>the trancparency and directness is stunning. Everything alse just sound dull or artificial compared.>
in the mentioned comparison the icepower were sounding quite dull and distant like in other reviews i have read i do think they have the best steady state specs but they could very well have the lowest sound quality (like we found).......
a good reason for sticking with traditional amps for a while is the verdict i read about lc audio zappulse some time ago ; "the sound quality is certainly below hifi standards but they are good at receiving radio , you twist the output inductors to tune in to different stations......"
as for low price i suggest that lars clausen publish the schematic of his low price pwm amp , in this way the diy folks here can make a couple and compare them to the amps they allready have, this i think would give a broad basis of comparisons and very valuable feedback for lars too...
i for one have been a bit reluctant to pay some $350 for what seems to be $20 worth of parts
rgds karsten madsen
ps. i havent heard the zappulse amp so i cant comment on it....
i think that there are several good reasons for using class-d/pwm amplifiers in high power applications , but i also think that there is a lot of hype and that the use of pwm amplifies in hifi/home applications isnt all that obvious.
i have been involved in listening tests comparing 3 different types of pwm amplifiers and a "normal" amplifier too (a current feedback amp like the alexander amp)
no doubt the latter presented the best sound quality of these.
the pwm amplifiers involved were ; icepower , tripath and our own discrete pwm amp.
i do realize that different people hear different things , as an example;
> have now owned a pair of monoblocks based on Ice-power (semi-diy) and I think they are superior in certain areas. I would never buy anything else actually.>
>the trancparency and directness is stunning. Everything alse just sound dull or artificial compared.>
in the mentioned comparison the icepower were sounding quite dull and distant like in other reviews i have read
i do think they have the best steady state specs but they could very well have the lowest sound quality (like we found).......a good reason for sticking with traditional amps for a while is the verdict i read about lc audio zappulse some time ago ; "the sound quality is certainly below hifi standards but they are good at receiving radio , you twist the output inductors to tune in to different stations......"
as for low price i suggest that lars clausen publish the schematic of his low price pwm amp , in this way the diy folks here can make a couple and compare them to the amps they allready have, this i think would give a broad basis of comparisons and very valuable feedback for lars too...
i for one have been a bit reluctant to pay some $350 for what seems to be $20 worth of parts
rgds karsten madsen
ps. i havent heard the zappulse amp so i cant comment on it....
Well Karsten, if you can produce our ZAPpulse for 20$: congratulations! You have a job!
Furthermore the retail price is NOT 350$, check your facts Karsten
Lastly i will give you right in your conclusions about pwm (may i use that term?) vs. analog amps for home use.
You were absolutely right about one year to 18 months ago. However evolution works, even in this area. Today i think there are several good pwm amp's available for home use also.
Icepower is (in my opinion) a very good amplifier, i have heard the latest versions, and i can not say anything bad about them.
Maybe you compared with some 2-3 year old modules, and then i might be inclined to agree with you in your conclusions. However the new ones are really good!
Furthermore the retail price is NOT 350$, check your facts Karsten
Lastly i will give you right in your conclusions about pwm (may i use that term?) vs. analog amps for home use.
You were absolutely right about one year to 18 months ago. However evolution works, even in this area. Today i think there are several good pwm amp's available for home use also.
Icepower is (in my opinion) a very good amplifier, i have heard the latest versions, and i can not say anything bad about them.
Maybe you compared with some 2-3 year old modules, and then i might be inclined to agree with you in your conclusions. However the new ones are really good!
hi .
thanx lars but i allready have a job
i think you forgot to comment on a few points , dont you think its a good idea to make a pwm amplifier available to the diy community , could be a simple version of your zappulse if you like.
this way your points about comparisons between pwm amps and traditional amps in hifi applications could have some more weight (imho) and on op of that you would get a lot of unbiased feedback helping to improve on the pwm amp topology and giving lots os hands-on experience to those who have opinions about pwm amps based on little (on none) practical experience.
feel free to contact me directly if you want to discuss this further (offline) and/or if you think we could realize a project like this this in some kind of cooperation.
if i may ask , why is the schematic of the zappulse amp secret when you have published schematics for your other amps?
i can only comment on the icepower and tripath boards i have heard and im not sure exactly what version of the icepower modules i have heard , i think its the ones sold through high-fidelity allthough i didnt get them there.
i guess you could say the same about tripath boards, to my knowledge there have been several revisions to them as well.
rgds karsten madsen
ps. my price of $350 seem to be for 2 boards , sorry but i just emembered you had a offer for 2 boards + 12v supply for around $350.....
thanx lars but i allready have a job
i think you forgot to comment on a few points , dont you think its a good idea to make a pwm amplifier available to the diy community , could be a simple version of your zappulse if you like.
this way your points about comparisons between pwm amps and traditional amps in hifi applications could have some more weight (imho) and on op of that you would get a lot of unbiased feedback helping to improve on the pwm amp topology and giving lots os hands-on experience to those who have opinions about pwm amps based on little (on none) practical experience.
feel free to contact me directly if you want to discuss this further (offline) and/or if you think we could realize a project like this this in some kind of cooperation.
if i may ask , why is the schematic of the zappulse amp secret when you have published schematics for your other amps?
i can only comment on the icepower and tripath boards i have heard and im not sure exactly what version of the icepower modules i have heard , i think its the ones sold through high-fidelity allthough i didnt get them there.
i guess you could say the same about tripath boards, to my knowledge there have been several revisions to them as well.
rgds karsten madsen
ps. my price of $350 seem to be for 2 boards , sorry but i just emembered you had a offer for 2 boards + 12v supply for around $350.....
km and dylanmrjones: Thanks for the suggestion, it seems like a good idea, however there is one problem. A ZAPpulse can not be implemented with full sound quality on less than a 4 layer PCB. This is because the ZAPpulse does not produce harmonic distortions, but rather random noise. And the only way to keep the random noise down is by the use of a 4 layer board with shielding ground planes, and a lot of HF capacitors.
I think most DIY'ers would try to build their implementation on 1-2 layer boards, and thus get a degraded performance.
But i might be able to find another good pwm amplifier, that is not ZAPpulse, but will give good results as a DIY project. Let me look into it. How about your pwm, km, maybe you can publish that in here??
Have a nice day from rainy and cold Denmark
I think most DIY'ers would try to build their implementation on 1-2 layer boards, and thus get a degraded performance.
But i might be able to find another good pwm amplifier, that is not ZAPpulse, but will give good results as a DIY project. Let me look into it. How about your pwm, km, maybe you can publish that in here??
Have a nice day from rainy and cold Denmark
Lars Clausen said:Jocko: If i understand right Class G is switching between different power supply levels, but using ananalog power stage ... right?
You are right about the distortions in the "********" amplifiers
( don't know what to call them anymore
""********" amplifiers", which ones do you mean ?
ClassG or analog in general ?
Bernhard
Bernhard: I would not use a term starting with the word: 'Class'.
So what i am referring to is some types (but not every) of 'amplifiers with a coil in the output for making square waves with various duty cycles into an analog audio signal that can feed a loudspeaker' do not produce harmonic distortions but rather produce random noise products instead...
Some people call them Class D. (Others call them other classes).
So what i am referring to is some types (but not every) of 'amplifiers with a coil in the output for making square waves with various duty cycles into an analog audio signal that can feed a loudspeaker'
do not produce harmonic distortions but rather produce random noise products instead...Some people call them Class D. (Others call them other classes).
If class AB amps haven't yet killed off class A, I don't think class D will ever replace class AB completely. I could see it getting into the mid-fi section of audio in the next few years, probably suited more towards battery powered boomboxes. I read an article about class D amps finding their way into cell phones and other portable equipment where battery life matters. As far as full range quality goes, I don't think class D will ever get to where AB is. For PA and huge subwoofer amps, class D will always win but I don't think it will completely replace AB.
Lars Clausen said:So what i am referring to is some types (but not every) of 'amplifiers with a coil in the output for making square waves with various duty cycles into an analog audio signal that can feed a loudspeaker'
And yet you can measure harmonic distortion at their output.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.