Disney_SK said:
Why don't you try UcD? They have the feedback taken after the output LC filter and as a result of that a high damping factor that is constant over a wider ferqeuncy range. Don't worry about the sound quality for full range or sat use, in my experience they sound much better than my Tripath based Marantz amp. Just try it, if it is no good, then at least you can say it from first hand experience. I think you will be happily surprised (as most others). I tried them in both full range and limted range (active speakers), they work great for both.
Best regards
Gertjan
Hi All!
I think some clarification regarding the relevance of a so-called high damping factor (DF) needs to be examined before one thinks of exotic designs. As some have suggested there is a lot of sales-fodder here, too often taken as gospel.
Basically (I presume we are all fimiliar with reversable electro-mechanical basics), the movement of a voice coil (in this instance) in a magnetic field after cessation of signal is dampened by the degree of back e.m.f. that it is able to generate. The current available for this action is a function of the generated voltage and the TOTAL resistance (impedance) in the loudspeaker circuit. If we disregard impedance and just take the minimum total d.c. resistance in the circuit, this will be the sum of amplifier output resistance, cable resistance AND voice coil resistance. Most practical 8 ohm loudspeakers have a voice coil resistance of some 5 ohm minimum. Thus, I am afraid that the maximum meaningful (real) damping factor possible is 8/5 = 1.6! The popular definition of damping factor as (loudspeaker impedance)/(amplifier output impedance) is meaningless as an indication of loudspeaker "braking effect" since it ignores the largest series impedance.
One will often find comments by reviewers that they could not really detect the difference between amplifiers of DF 30 and those higher, for the simple reason above. Measurements have borne this out.
Further, improving the damping factor by feedback alone or mainly, without an intrinsic low impedance output stage, could be hazardous without thorough measurements as to the full effect. Audible effects are often the result of improper feedback behaviour and have little to do with a higher DF per se. (The dangers of using high feedback to reduce output impedance have been illustrated in a landmark article by Otala and Lammasniemi in Wireless World way back in November 1980).
There is one exception, and that is where a negative output impedance is generated by use of combined negative voltage and positive current feedback, also called motional feedback. In such a design the loudspeaker and cable resistance (the latter is usually negligible) can be cancelled thus giving a "real" DF of infinity. But the design procedures are tricky and it is arguable whether the audible improvement is worthwhile (excepting perhaps in the region of the main loudspeaker bass resonance, which also involves the enclosure design).
I do not wish to antagonise anyone who claimed to have found the effect of a high DF audible. Such claims have been made and I respect people's testimony. It is nonetheless subjective and has to my knowledge never been borne out by proper blind tests; on the contrary. As said above, there could be other reasons for an audible difference; it is highly unlikely to have resulted simply from an improved "braking" of the driver motion.
Regards.
I think some clarification regarding the relevance of a so-called high damping factor (DF) needs to be examined before one thinks of exotic designs. As some have suggested there is a lot of sales-fodder here, too often taken as gospel.
Basically (I presume we are all fimiliar with reversable electro-mechanical basics), the movement of a voice coil (in this instance) in a magnetic field after cessation of signal is dampened by the degree of back e.m.f. that it is able to generate. The current available for this action is a function of the generated voltage and the TOTAL resistance (impedance) in the loudspeaker circuit. If we disregard impedance and just take the minimum total d.c. resistance in the circuit, this will be the sum of amplifier output resistance, cable resistance AND voice coil resistance. Most practical 8 ohm loudspeakers have a voice coil resistance of some 5 ohm minimum. Thus, I am afraid that the maximum meaningful (real) damping factor possible is 8/5 = 1.6! The popular definition of damping factor as (loudspeaker impedance)/(amplifier output impedance) is meaningless as an indication of loudspeaker "braking effect" since it ignores the largest series impedance.
One will often find comments by reviewers that they could not really detect the difference between amplifiers of DF 30 and those higher, for the simple reason above. Measurements have borne this out.
Further, improving the damping factor by feedback alone or mainly, without an intrinsic low impedance output stage, could be hazardous without thorough measurements as to the full effect. Audible effects are often the result of improper feedback behaviour and have little to do with a higher DF per se. (The dangers of using high feedback to reduce output impedance have been illustrated in a landmark article by Otala and Lammasniemi in Wireless World way back in November 1980).
There is one exception, and that is where a negative output impedance is generated by use of combined negative voltage and positive current feedback, also called motional feedback. In such a design the loudspeaker and cable resistance (the latter is usually negligible) can be cancelled thus giving a "real" DF of infinity. But the design procedures are tricky and it is arguable whether the audible improvement is worthwhile (excepting perhaps in the region of the main loudspeaker bass resonance, which also involves the enclosure design).
I do not wish to antagonise anyone who claimed to have found the effect of a high DF audible. Such claims have been made and I respect people's testimony. It is nonetheless subjective and has to my knowledge never been borne out by proper blind tests; on the contrary. As said above, there could be other reasons for an audible difference; it is highly unlikely to have resulted simply from an improved "braking" of the driver motion.
Regards.
Hi Guys, the best bass performance that i have heard was from an open loop D class amplifier that used an PFC based SMPS, variable voltage switching regs, running via a its digital input.
What i mean by 'best' - tighest most nimble articulate bottom end i have heard.
Speakers were Tannoy TD10's.
So thats digital input D class with no feedback - Damping factor was about 40.
THD was about 0.1%
Stunning...
Sheriff
What i mean by 'best' - tighest most nimble articulate bottom end i have heard.
Speakers were Tannoy TD10's.
So thats digital input D class with no feedback - Damping factor was about 40.
THD was about 0.1%
Stunning...
Sheriff
Thanks for replies, anyway, I think most of you misunderstood the meaning of this thread.
Somebody may say he likes lower damping factor, because sound with high one sounds to him "dead". Somebody might say there is no need to even try to raise DF (most of you, i see), but somebody might like to have high DF.
There have been done some measurements by some people I know, what should I not mention here, that say that the CLOSEST and UNEXACT, but MOST EXACT from all possible amplifiers' parameters we use, that determines the deepness and tightness of bass, is mentioned Damping Factor.
So please, shall we stop making conclusions about how high damping factor is the best? Shall we come back to the subject of this thread? I would really appredicate that.
Thanks for understanding.
Somebody may say he likes lower damping factor, because sound with high one sounds to him "dead". Somebody might say there is no need to even try to raise DF (most of you, i see), but somebody might like to have high DF.
There have been done some measurements by some people I know, what should I not mention here, that say that the CLOSEST and UNEXACT, but MOST EXACT from all possible amplifiers' parameters we use, that determines the deepness and tightness of bass, is mentioned Damping Factor.
So please, shall we stop making conclusions about how high damping factor is the best? Shall we come back to the subject of this thread? I would really appredicate that.
Thanks for understanding.
In the datasheet of TDA2500, the only thing mentioned about Damping Factor, is:
"To maximize damping factor and reduce distortion and noise, the modulator feedback connections should be routed directly to the pins of the output inductors L0".
I have emailed Tripath for some additional information, and I have got this reply:
"The only other advice that I can give for improving damping factor is to choose an output inductor that has very low DC resistance."
So we have two conditions, which we have to implement to obtain highest possible DF.
I had an idea to lower the output resistance by using more parallel mosfets, e.g. two to four per rail. But Tripath chips have limited the total gate charge, which can not be exceeded. So I was thinking about some additional gate-driving circuitry, that would be driven by the chip and would be able to supply more charge into the gates, say, around 200nC. But this is pretty tricky, because we work with spread-spectrum frequencies up to 1MHz and I am not sure whether it would be better choice to use discrete buffer, or some IC buffer. I think discrete buffer might be a better choice.
Someone mentioned the influence of negative feedback. Tripath amplifiers use switching feedback from modulated audio signal. But shall we try to add also the mostly used in class-AB amplifiers feedback from the final output back to the input stage of the amplifier? Any reasons, why should this not be done?
"To maximize damping factor and reduce distortion and noise, the modulator feedback connections should be routed directly to the pins of the output inductors L0".
I have emailed Tripath for some additional information, and I have got this reply:
"The only other advice that I can give for improving damping factor is to choose an output inductor that has very low DC resistance."
So we have two conditions, which we have to implement to obtain highest possible DF.
I had an idea to lower the output resistance by using more parallel mosfets, e.g. two to four per rail. But Tripath chips have limited the total gate charge, which can not be exceeded. So I was thinking about some additional gate-driving circuitry, that would be driven by the chip and would be able to supply more charge into the gates, say, around 200nC. But this is pretty tricky, because we work with spread-spectrum frequencies up to 1MHz and I am not sure whether it would be better choice to use discrete buffer, or some IC buffer. I think discrete buffer might be a better choice.
Someone mentioned the influence of negative feedback. Tripath amplifiers use switching feedback from modulated audio signal. But shall we try to add also the mostly used in class-AB amplifiers feedback from the final output back to the input stage of the amplifier? Any reasons, why should this not be done?
Hi,
I think you greatly misunderstand/missed the point.
You've got tunnel vision on achieving a high damping factor, it has been very well explained why it is just not a useful performance related measurement, in fact it is misunderstood and typically nothing more than a marketing number.
If you really do want to do this, negative impedance is the way to go, not a tripath chip which already has feedback well defined/set in stone, that's why they told you your only option is a good component selection and layout.
By worrying only about a high damping factor you're going down a long hard road which will lead to a dead end, and your friend is wrong to place that kind of importance on it, if anything it shows he lacks understanding, and most likely wrong to think the effective damping factor of the amps he's using is anywhere near 800.
You need to reread the posts in this thread and understand them.
I think you greatly misunderstand/missed the point.
You've got tunnel vision on achieving a high damping factor, it has been very well explained why it is just not a useful performance related measurement, in fact it is misunderstood and typically nothing more than a marketing number.
If you really do want to do this, negative impedance is the way to go, not a tripath chip which already has feedback well defined/set in stone, that's why they told you your only option is a good component selection and layout.
By worrying only about a high damping factor you're going down a long hard road which will lead to a dead end, and your friend is wrong to place that kind of importance on it, if anything it shows he lacks understanding, and most likely wrong to think the effective damping factor of the amps he's using is anywhere near 800.
You need to reread the posts in this thread and understand them.
Disney_SK,
As one of the lenghtier contributors (post #22 - I am not going to box it all), I am sorry if I bored you .... but how it was off-thread?
I read again your first post. If it was all about Tripath I did miss the point and apologise, but you did ask some fairly general questions: "Maximising DF" and how you could employ feedback to improve DF. (...which on its own is not unequivocally a good idea, but I will not go into that then.)
Pardom me for concluding that you saw some advantage in a high DF! And by now you must have noticed that it is NOT important. Should we have misled you by showing you how to achieve something that should not have been done in the first place?
I have always tried (and believe succeeded mostly) to be respectful concerning the ideas of others. But a lot of people have put a lot of things on the internet in very convincing narrative, but not always with as much scientific basis. And I am sorry, but anybody who tries to make a case for a high DF per se......IS WRONG. (Sorry if that appears off-thread.)
Regards
Johan
As one of the lenghtier contributors (post #22 - I am not going to box it all), I am sorry if I bored you .... but how it was off-thread?
I read again your first post. If it was all about Tripath I did miss the point and apologise, but you did ask some fairly general questions: "Maximising DF" and how you could employ feedback to improve DF. (...which on its own is not unequivocally a good idea, but I will not go into that then.)
Pardom me for concluding that you saw some advantage in a high DF! And by now you must have noticed that it is NOT important. Should we have misled you by showing you how to achieve something that should not have been done in the first place?
I have always tried (and believe succeeded mostly) to be respectful concerning the ideas of others. But a lot of people have put a lot of things on the internet in very convincing narrative, but not always with as much scientific basis. And I am sorry, but anybody who tries to make a case for a high DF per se......IS WRONG. (Sorry if that appears off-thread.)
Regards
Johan
Johan Potgieter said:
I do not want to read pardon from you.
Maybe I asked for help the bad way.
But to me, this all seems to me ... like I am asking you "how to do this?", and you tell me "do not do that, because it is not important, and if somebody thinks it is important, he IS WRONG".
I am really sorry, but should we not be more tolerable to each other and our experiences, which are, as we see, different?
I understand that you may not, and either do not want to tell me how to do something, if you believe that it is wrong to do. But please, let's be tolerable. I know some people who put hundreds of thousands dollars into test equipment, mailny to see, why one amp reproduces bass better than other etc. And they were very surprised by the conclusions they had to make! And so did I! And pardon me, but I will follow the valuable results of these experiments.
I think that is the way to go. We have to be open-minded.
I would like to thank everybody who had a wish to help me.
I will follow the main given rules, which clearly lead to maximize damping factor, that is low resistance inductor and good feedback routing.
Anybody who agrees, that high damping factor might be important, and has other ideas to tell, is very welcome to help.
Damping factor
Dear Disney,
Consider the following. If an amplifier has a DF of say 500 with an 8 ohm load at say 500Hz and below, this means that it's output impedance is 0.016 ohms.
Now Jan-Peter correctly says consider the following added resistance which are ALL IN SERIES WITH THE speaker circuit. You have contact resistance at both the speaker and amplifier connectors and this 4 times. Let us assume at each has a resistance of 0.002 ohm (This indicates a very good connector believe me!) So now we have 0.008 ohms added to the Zout of the amplifier. The effective DF is now 333.33 - dropped from 500 which could be measured at the speaker terminals of the amplifier.
All good so far. To be generous let's assume the speaker cable has a length of say 3 metres (10'). Let's assume that you are using #8 wire as well. This has a resistance of 0.6281 ohms per 1000 feet (304.8m) The two legs of the speaker cable is now 20 feet or about 6 metres. Great and now we add in the resistance of this 6m of wire. Do the calculation - 20 divided by 1000 and multiply by 0.6281 and you get 0.012 ohms. Now add this to the 0.008 and the 0.016 and we have 0.036 ohms of total series impedance. New DF is now 8/0.036 = 222.22. So now the DF is less than 50% of what the amplifier measured and this new DF is the effective damping factor.
All nice and dandy - well hold your horses, not so fast. One important paramter we have not added in is the resistance of the voice coil of the speaker. This resistance is 100% IN SERIES with all those other resistance (The speaker circuit is a simple series circuit of an amplifier in series with some cable, in series with some connectors and in series with - you gussed it THE RESISTANCE OF THE VOICE COIL)
So now what we have is an 8 ohm speaker which typically has a DCR of about 6 ohms. Add this 6 ohms to the 0.036 and we have 6.036 ohms. Your DF is now effectively 8/6.036 = 1.325.
So the small resistance can be neglected and so can the Zout of the amplifier. The DCR of the speaker determines the effective DF
The late Julian Hirsch wrote a white paper on this issue.
Dear Disney,
Consider the following. If an amplifier has a DF of say 500 with an 8 ohm load at say 500Hz and below, this means that it's output impedance is 0.016 ohms.
Now Jan-Peter correctly says consider the following added resistance which are ALL IN SERIES WITH THE speaker circuit. You have contact resistance at both the speaker and amplifier connectors and this 4 times. Let us assume at each has a resistance of 0.002 ohm (This indicates a very good connector believe me!) So now we have 0.008 ohms added to the Zout of the amplifier. The effective DF is now 333.33 - dropped from 500 which could be measured at the speaker terminals of the amplifier.
All good so far. To be generous let's assume the speaker cable has a length of say 3 metres (10'). Let's assume that you are using #8 wire as well. This has a resistance of 0.6281 ohms per 1000 feet (304.8m) The two legs of the speaker cable is now 20 feet or about 6 metres. Great and now we add in the resistance of this 6m of wire. Do the calculation - 20 divided by 1000 and multiply by 0.6281 and you get 0.012 ohms. Now add this to the 0.008 and the 0.016 and we have 0.036 ohms of total series impedance. New DF is now 8/0.036 = 222.22. So now the DF is less than 50% of what the amplifier measured and this new DF is the effective damping factor.
All nice and dandy - well hold your horses, not so fast. One important paramter we have not added in is the resistance of the voice coil of the speaker. This resistance is 100% IN SERIES with all those other resistance (The speaker circuit is a simple series circuit of an amplifier in series with some cable, in series with some connectors and in series with - you gussed it THE RESISTANCE OF THE VOICE COIL)
So now what we have is an 8 ohm speaker which typically has a DCR of about 6 ohms. Add this 6 ohms to the 0.036 and we have 6.036 ohms. Your DF is now effectively 8/6.036 = 1.325.
So the small resistance can be neglected and so can the Zout of the amplifier. The DCR of the speaker determines the effective DF
The late Julian Hirsch wrote a white paper on this issue.
My two cents....;
There is an difference in sound quality between amplifiers and sometimes you can get a relation with the DF. However to reach a good low DF the amplifier need a lot of loopgain. The higher the loopgain the lower the DF.
Now...what MOER explains in his posting is that the theoretical DF of an amp can be high, but in practice the rest of the resistance included the DC coil restiance has much more influence on the DF in reall life.
So...the difference what you think you hear is not the difference in DF but the difference in loopgain......
Jan-Peter
There is an difference in sound quality between amplifiers and sometimes you can get a relation with the DF. However to reach a good low DF the amplifier need a lot of loopgain. The higher the loopgain the lower the DF.
Now...what MOER explains in his posting is that the theoretical DF of an amp can be high, but in practice the rest of the resistance included the DC coil restiance has much more influence on the DF in reall life.
So...the difference what you think you hear is not the difference in DF but the difference in loopgain......
Jan-Peter
Disney_SK said:
So why stick to tripath? With amps that take feedback after the filter, damping factor can be higher as the coils DC resistance will be effectively reduced by the feedback. As some suggested, you could even make a more than infinite damping factor (read negative damping factor) feeding back the current that flows through the speaker and use that as positive feedback. this way you can create even a negative output resistance, basically you can compensate part of the DC resistance of the voice coil of the speaker, thus lowering the Q-factor of the speaker and getting tighter bass.
Best regards
Gertjan
Re: Re: Re: Damping factor
Yeah, though......
I however have the feeling we forget something in this reverse effect.
Many, many moons ago I just fumbled around with a laser interferometer and a speaker just to see what happens with resonance. And there was quite a difference whether you lowered resistance (used resistor pad) from 1 Ohm to .1 Ohm and a short circuit.
Speaker was a 9710 (8 Ohms) from Philips then. But I have no data anymore.
I fear I have to leave this to a real practizing freak ......
Jan-Peter said:
Yeah, though......
I however have the feeling we forget something in this reverse effect.
Many, many moons ago I just fumbled around with a laser interferometer and a speaker just to see what happens with resonance. And there was quite a difference whether you lowered resistance (used resistor pad) from 1 Ohm to .1 Ohm and a short circuit.
Speaker was a 9710 (8 Ohms) from Philips then. But I have no data anymore.
I fear I have to leave this to a real practizing freak ......
Disney_SK and all,
You choose to single me out for rebuke although others have made similar comments to mine - so be it. And apologising for possible misunderstanding is my right; others read this as well.
Now if we can get back to science:
1. Final paragraphs of both my contributions indicated that I AM respectful of the experience of others - and believe me, in my 50+ years as a professional EE, I had to contend with a lot of that over the whole spectrum.
2. But experiences can be very subjective, and there can be many explanations. I never pushed for my (personal) ideas when mentioning that the "advantages" of high DF (at least per the popular definition) are a fallacy. I simply stated established basic electronic principles, as contained in countless relevant handbooks, and repeated by several honourable members on this thread. There is no snake oil or black magic here. If I disappointed you in not giving information, it was simply because others have done so already.
3. Despite this I wish you success, Disney_SK - I am off this thread as from now.
Regards
Johan
You choose to single me out for rebuke although others have made similar comments to mine - so be it. And apologising for possible misunderstanding is my right; others read this as well.
Now if we can get back to science:
1. Final paragraphs of both my contributions indicated that I AM respectful of the experience of others - and believe me, in my 50+ years as a professional EE, I had to contend with a lot of that over the whole spectrum.
2. But experiences can be very subjective, and there can be many explanations. I never pushed for my (personal) ideas when mentioning that the "advantages" of high DF (at least per the popular definition) are a fallacy. I simply stated established basic electronic principles, as contained in countless relevant handbooks, and repeated by several honourable members on this thread. There is no snake oil or black magic here. If I disappointed you in not giving information, it was simply because others have done so already.
3. Despite this I wish you success, Disney_SK - I am off this thread as from now.
Regards
Johan
Snake Oil?
Yes this is snake oil unless the laws of physics have been repealed. No preprocessing will eliminate the inertia of the cone assembly or the back EMF generated. It can be compensated for to a small degree but this is introducing false artifacts into the signal.
The other equally troubling statement that the voice coil is only driven in one direction is equally false as it is driven to follow the signal in both directions.
When the back EMF goes into the feedback loop it generates a counter signal that overcomes the inertial effects of the cone to some degree beyond mere damping factor or output impedance. This is a counter or negative feedback signal developed across the amps output impedance. This becomes a good case for actually having higher output impedance or lower damping factor as there would be more of this effect. These things are only relevant at the lower frequencies where there are significant cone motions.
Other ways have been used to improve LF fidelity like an actual transducer attached to the cone for feedback generation. This can give excellent and solid bass without having a ton of negative feedback in the amp.
Roger
raintalk said:
Yes this is snake oil unless the laws of physics have been repealed. No preprocessing will eliminate the inertia of the cone assembly or the back EMF generated. It can be compensated for to a small degree but this is introducing false artifacts into the signal.
The other equally troubling statement that the voice coil is only driven in one direction is equally false as it is driven to follow the signal in both directions.
When the back EMF goes into the feedback loop it generates a counter signal that overcomes the inertial effects of the cone to some degree beyond mere damping factor or output impedance. This is a counter or negative feedback signal developed across the amps output impedance. This becomes a good case for actually having higher output impedance or lower damping factor as there would be more of this effect. These things are only relevant at the lower frequencies where there are significant cone motions.
Other ways have been used to improve LF fidelity like an actual transducer attached to the cone for feedback generation. This can give excellent and solid bass without having a ton of negative feedback in the amp.
Roger
I know a lot has been said about DF and the law of diminishing return (that is, extremely high DFs won't really get you any more bang for your buck).
==> Setting all things aside, what are some techniques for lowering DF in a solid state amp? More output transistor pairs? A higher bias current in the output pairs? Higher beta transistors?
-RT
==> Setting all things aside, what are some techniques for lowering DF in a solid state amp? More output transistor pairs? A higher bias current in the output pairs? Higher beta transistors?
-RT
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.