Hi Rodolfo,
your technical dialectic surpasses me and I have not comment details. Of found it is always a my tendency to "simplify" and your "professionalism" that specifies. From this point of view we compensate.
In fact even in this comments ( es: they end now endless Zout in current driver, and point 2 ( effects of the "back_EMF-return-current" on the internal eddy current in the magnetic circuit ) am not entirely of accord, but only by a point of view of phenomenon of "quantity" and "consequence" produced. To is not a lot of important...
Ciao
Mauro
your technical dialectic surpasses me and I have not comment details. Of found it is always a my tendency to "simplify" and your "professionalism" that specifies. From this point of view we compensate.
In fact even in this comments ( es: they end now endless Zout in current driver, and point 2 ( effects of the "back_EMF-return-current" on the internal eddy current in the magnetic circuit ) am not entirely of accord, but only by a point of view of phenomenon of "quantity" and "consequence" produced. To is not a lot of important...
Ciao
Mauro
ingrast said:
and
how come... sorry: no, no, no... no! 😱
No matter how you drive the spk, if you have distorsion on either spk voltage or current (or both), the distorsion is there, and you'll find it on the "output" (sound!).
...why on heart should it magically vanish!?
Of course the relations between spk voltage->sound and spk current->sound are different, so it may require some (very possibly non-trivial) calculations to quantitatively compare the two with each other and with the distorsion of the sound.
That's one of the reasons why I insist in saying that to be meaningful, any distorsion measurement can't be done simply on the voltage output of the DUT (amplifier) connected to a dummy load.
They should be done on the actual output of a complete system. That is, measure the distorsion on the sound produced by the DUT connected to some real-life loudspeaker!
(of course this way you can't easily isolate the distorsion generated by the DUT from that produced by the test speaker, but you CAN COMPARE different amplifiers and see which one will produce less distortion in much more lifelike conditions).
BTW: does anybody ever tried to do so? I'd be very curious to see some results...
On a different note, but still related to Back EMF, taking a typical 12db/octave LC (series L ,parallel C) low pass passive crossover, we still would like lots of electrical damping, correct? Well, there isn't ANY near the crossover region no matter how good the amplifier's damping factor is. In fact the better the amp damping factor is, the worse the problem is near the crossover region, same with less lossy Ls and Cs. The amp can be thought of as a short circuit looking back into it's output terminals, so the woofer's back EMF just 'sees' a parallel LC at resonance, i.e. infinity impedance at the crossover frequency, so NO damping.
Put a 15R resistor across the woofer or lower if you dare to give the back emf somewhere to dump some of its current, and hear a cleaner sound that's nearer to nicely damped active amplification. It won't do much for bass 'tightness' but it will reduce colouration.
Put a 15R resistor across the woofer or lower if you dare to give the back emf somewhere to dump some of its current, and hear a cleaner sound that's nearer to nicely damped active amplification. It won't do much for bass 'tightness' but it will reduce colouration.
mauropenasa said:
Mauro:
I apologize if my post resulted confusing.
What I wanted to make clear, is the issues are different dependent on the assumptions made.
I want to **stress** this point, since unless one is coherent with the integrity of the reasoning chain, the unavoidable consequence is confussion.
In the scenario of voltage-makes-sound, current is irrelevant. Think on it, reflect.
The consequence is current distortion is a non-issue, weird as it may look. (If current mattered, then the starting hypothesis should not be true).
In the scenario of current-makes-sound, the reverse is true, that is voltage distortion is irrelevant (for the same reasons, if it were relevant then the hypothesis should be contradicted).
Depending on which theory you embrace, you make the amplifier stiff for voltage or stiff for current.
I want to rise this observation since I have seen posts proposing a middle ground. I see no advantage to this stand unless there is ample, repeatable, confirmed evidence that it is better. If that were the case, then it should be possible to construct a better electroacustical coupling model that supports it. Not that I deny it, but am not aware of work in this direction (what of course can be my fault).
Rodolfo
UnixMan said:
...how come... sorry: no, no, no... no! 😱
No matter how you drive the spk, if you have distorsion on either spk voltage or current (or both), the distorsion is there, and you'll find it on the "output" (sound!).
...why on heart should it magically vanish!?
Yes UnixMan, it magicaly vanishes since it does not contribute to sound (depending on the starting hypothesis). Please read the previous post.
Rodolfo
sbrads said:
Well, though in fact it is not strictly that way, the overall rationale is correct.
Now ask me why I vow for active crossovers 😎 .
Rodolfo
Now lets assume a current drive amp, say with 10k output Z, which is pretty high as current sources (the audio power variety) go. What we see is that the damping in the bass region, which was still there with the voltage drive, is now gone. So, it appears that as far as bass control (or lf damping) is concerned, voltage drive is still better in this configuration. QED.
Jan Didden
Jan Didden
Hi Jan,
Was taking a break. Your Post#49.
Pity many Electronics World letter writers don't see it like that. Cheers.
Hi darkfenriz,
Does Rod Elliot not cover something like the test you suggest in Post#33 ?
Hi tlf9999,
Post#36. In view of the fact that you appear to be accusing me of using an invented term 'back-EMF' I took the trouble to look it up in one of my American reference books in case there are differences between our terminologies. All I can say is that I trust you have since brought your AC-theory up to date.
Hi Ashok,
In your Post#71 you said the tube amp had better sound but the solid state better damping.
Did the loudspeaker have a crossover, such that the Prima Luna was damping cable/circuit 'Q' where the Creek was not ?
Hi sbrads,
And that damping resistor needs to have a decent power rating.
I fried one last week testing my new amplifier circuit and was wondering why the 'amplifier' (unknown at that stage) was not damping properly when it should have been; I experienced exactly what you write about.
Low amplifier output impedance maintains cable/crossover 'Q', so unless there are correct Zobels, damping resistors and impedance compensation circuits, then composite loudspeaker systems cannot fail to sound different when driven by different amplifiers.
Cheers ........... Graham.
Was taking a break. Your Post#49.
Pity many Electronics World letter writers don't see it like that. Cheers.
Hi darkfenriz,
Does Rod Elliot not cover something like the test you suggest in Post#33 ?
Hi tlf9999,
Post#36. In view of the fact that you appear to be accusing me of using an invented term 'back-EMF' I took the trouble to look it up in one of my American reference books in case there are differences between our terminologies. All I can say is that I trust you have since brought your AC-theory up to date.
Hi Ashok,
In your Post#71 you said the tube amp had better sound but the solid state better damping.
Did the loudspeaker have a crossover, such that the Prima Luna was damping cable/circuit 'Q' where the Creek was not ?
Hi sbrads,
And that damping resistor needs to have a decent power rating.
I fried one last week testing my new amplifier circuit and was wondering why the 'amplifier' (unknown at that stage) was not damping properly when it should have been; I experienced exactly what you write about.
Low amplifier output impedance maintains cable/crossover 'Q', so unless there are correct Zobels, damping resistors and impedance compensation circuits, then composite loudspeaker systems cannot fail to sound different when driven by different amplifiers.
Cheers ........... Graham.
Hi Graham,
What I meant to say was the Prima Luna had very tuneful tight bass while the Creek did not. About the damping , I meant that the Creek probably had a better damping factor than the Prima Luna. So in theory the Creek should be 'tighter'.
It's possible that the Creek did not damp the cable/crossover Q .
Is there any test jig that can show up these differences ( on a scope)?
To my ears it sounded as if the transient spectrum via the Prima Luna was maintained intact in time and amplitude while the Creek lost out in the HF content and has HF delay ( duller sound!).
We should try to replicate this with some test signals . If the smeared transient looses much of its bite then possibly we will have to figure out where the problem occured. Audio bandwidth is certainly not the problem.
Hasn't anyone come up with a bench test that can show up such audible differences without connecting up a speaker ?
Cheers.
The crossover was 6db/ octave and the cable was 10 feet of stranded OFC cable ( Taiwanese origin). Cable resistance must have been less than 0.1 ohms.
Hi darkfenriz,
It is on Rod's site somewhere, and in my CD roms somewhwere too due to Windows crashes/rebuilds. The quickest way to find it would be searching Rod's 'esp' (Elliot Sound Products) website.
Hi Ashok,
And yet I think it likely that if you A-B checked those amplifiers at a low lounge listening levels you would find it hard, if at all possible, to tell them apart.
6dB/oct is not a difficult load.
The Prima Luna would still have good native open loop performance at hf because the output valve grids are well decoupled from the anodes, this is further improved by a limited amount of NFB, probably not more than 20dB, and the final damping figure is unlikely to be anything like the Creek's, so without the virtual amplifier end short circuit any single loudspeaker circuit back-EMF will be less actively cable impedance modified and re-developed across all loudspeaker circuits.
This is where a parallel connected 10 ohm resistor at the output terminals and a 0.22 ohm resistor in series with the loudspeaker might have reduced Creek induced characterisations. If you ever try this, I should be pleased to hear of your findings.
During the early seventies I found the sound from tetrodes was much better when driving reactive loudspeakers. This was because lower internal anode electron reflection and re-radiation to G2 when hot allowed better grid controlled anode current flow. Did that version of the Prima Luna have EL34 pentodes or KT88 tetrodes.?
I wonder if the Creek also lost its timing because of the internal stabilisation components so frequently used to maintain global stability, but which simultaneously introduce inductive phase change to NFB control.
A solid state amplifier can have superior sine thd performance but its 'first cycle distortion' becomes impaired as a direct result, and thus, with music being an endless stream of first (non-sine) cycles, it loses clarity; the louder the worse.
One way of checking whether an amplifier has such an 'inductive' characteristic is to open the global loop but leave all stabilisation components in place. If it sounds muffled and does not have a good frequency response then it is most unlikely to sound good as a finished product.
Such an amplifier can reproduce bass or voices or treble powerfully, but not coherently when they are all sounding together, and even less so when the loudspeaker has a high inductance bass driver and additional back-EMF inducing crossover circuitry.
This was the reason for me developing my last 100W A//AB amplifier circuit. I wanted something that could better KT88s, which already offered better dynamic reproduction than all other solid state designs I had heard except for maintaining very low frequency 'electronic' bass instrument power output, or extreme fineness of hf detail, both being due to output transformer imperfections rather than the tubes, though not overall being worse than than the offerings of most 'hi-fi' solid-staters.
If that amplifier had KT88's, then I can't help but think you'll need to DIY or have a very deep pocket to purchase better sound !
I have always maintained that 50W of KT88 performs as well as 100W and sometimes more of solid-state, except at the very lowest of bass frequencies, so what was the power of the Creek, and by any chance did their sound output levels appear roughly similar ?
Cheers ........... Graham.
It is on Rod's site somewhere, and in my CD roms somewhwere too due to Windows crashes/rebuilds. The quickest way to find it would be searching Rod's 'esp' (Elliot Sound Products) website.
Hi Ashok,
And yet I think it likely that if you A-B checked those amplifiers at a low lounge listening levels you would find it hard, if at all possible, to tell them apart.
6dB/oct is not a difficult load.
The Prima Luna would still have good native open loop performance at hf because the output valve grids are well decoupled from the anodes, this is further improved by a limited amount of NFB, probably not more than 20dB, and the final damping figure is unlikely to be anything like the Creek's, so without the virtual amplifier end short circuit any single loudspeaker circuit back-EMF will be less actively cable impedance modified and re-developed across all loudspeaker circuits.
This is where a parallel connected 10 ohm resistor at the output terminals and a 0.22 ohm resistor in series with the loudspeaker might have reduced Creek induced characterisations. If you ever try this, I should be pleased to hear of your findings.
During the early seventies I found the sound from tetrodes was much better when driving reactive loudspeakers. This was because lower internal anode electron reflection and re-radiation to G2 when hot allowed better grid controlled anode current flow. Did that version of the Prima Luna have EL34 pentodes or KT88 tetrodes.?
I wonder if the Creek also lost its timing because of the internal stabilisation components so frequently used to maintain global stability, but which simultaneously introduce inductive phase change to NFB control.
A solid state amplifier can have superior sine thd performance but its 'first cycle distortion' becomes impaired as a direct result, and thus, with music being an endless stream of first (non-sine) cycles, it loses clarity; the louder the worse.
One way of checking whether an amplifier has such an 'inductive' characteristic is to open the global loop but leave all stabilisation components in place. If it sounds muffled and does not have a good frequency response then it is most unlikely to sound good as a finished product.
Such an amplifier can reproduce bass or voices or treble powerfully, but not coherently when they are all sounding together, and even less so when the loudspeaker has a high inductance bass driver and additional back-EMF inducing crossover circuitry.
This was the reason for me developing my last 100W A//AB amplifier circuit. I wanted something that could better KT88s, which already offered better dynamic reproduction than all other solid state designs I had heard except for maintaining very low frequency 'electronic' bass instrument power output, or extreme fineness of hf detail, both being due to output transformer imperfections rather than the tubes, though not overall being worse than than the offerings of most 'hi-fi' solid-staters.
If that amplifier had KT88's, then I can't help but think you'll need to DIY or have a very deep pocket to purchase better sound !
I have always maintained that 50W of KT88 performs as well as 100W and sometimes more of solid-state, except at the very lowest of bass frequencies, so what was the power of the Creek, and by any chance did their sound output levels appear roughly similar ?
Cheers ........... Graham.
ingrast said:
com'on Rodolfo, both of those "starting hypothesis" are obviously wrong! 😱
One thing is arguing that "the relation between spkr current and sound pressure is more linear than that relating spkr voltage to sound, thus spkrs should be driven with a VCCS" (or vice versa, for those favoring voltage drive), and a whole different thing is pretending that the sound output of a spkr is depending _only_ on either the voltage or the current, which is complete nonsense!!
The job of any transducer is that of converting one form of energy to another. In our case, the loudspeaker does convert some electrical energy from the driving signal to mechanical energy of vibrating air. Right?
By definition, the amount of energy E in a time unit is the power P == dE/dt.
Or the other way around, we can state E = ∫Pdt
Now, as we all know for electricity: p = v * i
If we have some distorsion components on either voltage or current spectra (or both), that distortion components will contribute to the total power supplied to (drawn by) the spkr.
Obviously, whatever we have connected to our spkr, no matter what the source impedances is, no matter wether we have NFB or not, tube or transistor or mechanical generator or anything else, the electrical power supplied to (drawn by) the loudspeaker will be "translated" into mechanical power (of course modulo conversion efficiency, the rest being converted into heat by various means).
Thus if we have distorsion on either current or voltage (or both) we will find distorsion also on the sound, no matter what.
To me, this is obvious: it looks like very basic physics... 🙄
AFAIK, in the known universe Energy cannot disappear... and I see no reason why the speaker should suddenly decide to completely convert into heat all and only the distorsion components on its current if it's driven by a VCVS or those on its voltage if its driven by a VCCS!
Look at the attahed schematic, consisting of a voltage source driving diodes in series with a resistor. The distortion appearing across the resistor is large due to 0.6V of the waveform being lopped off. What happens if the voltage source is switched for a current source? Magically, the distortion across the resistor disappears! What happens to the extra power? It is converted to heat. If the diodes were in parallel with the resistor instead of in series, then it would be the voltage source that yielded the least distortion.
Speakers don't have diodes in series with them of course, but many of the mechanisms causing distortion do similarly disappear when using a current source, with the extra power turning into heat.
Speakers don't have diodes in series with them of course, but many of the mechanisms causing distortion do similarly disappear when using a current source, with the extra power turning into heat.
Attachments
Hi Graham,
The Prima Luna had KT88's !
It was a 45 watt per channel amp but sounded like it was a 100 watts or more. Very loud levels were possible but after a point it just gave up .I'm sure it clipped somewhere along the way but was very hard to detect by ear.
The Creek played pretty loud too but sounded awful when clipped, not the way the Prima Luna reached the max output. Typical ss vs tube clipping .
I'm inclined to believe that the KT88 does have some special characteristics , especially the bass.
I have a set of brand new KT88 tubes. I bought them to build a PP tube amp. Haven't yet decided what transformer to use and if I should use UL or not and fixed or cathode bias. I also did think of trying out an autobias circuit.
Cheers.
The Prima Luna had KT88's !
It was a 45 watt per channel amp but sounded like it was a 100 watts or more. Very loud levels were possible but after a point it just gave up .I'm sure it clipped somewhere along the way but was very hard to detect by ear.
The Creek played pretty loud too but sounded awful when clipped, not the way the Prima Luna reached the max output. Typical ss vs tube clipping .
I'm inclined to believe that the KT88 does have some special characteristics , especially the bass.
I have a set of brand new KT88 tubes. I bought them to build a PP tube amp. Haven't yet decided what transformer to use and if I should use UL or not and fixed or cathode bias. I also did think of trying out an autobias circuit.
Cheers.
Thank you Graham
found it.
although it was not funny at all to get to know someone have done it before I came to this 'innovative' idea 😀
Hi UnixMan
You miss one vital thing- no one cares about linearity of power as efficiency is often less than 0.5%. Than one should analyse the relationship between quantities being transduced e.g. voltage to velocity (v=BlV), current to force (F=BlI) where B is magnetic induction and l is legth of a coil. Both current and voltege driving will suffer from Bl non-linearity then. I still trust voltage driving, which results in velocity; Force (achieved by current) is highly non-linear because of cone suspension.
regards
found it.
although it was not funny at all to get to know someone have done it before I came to this 'innovative' idea 😀
Hi UnixMan
You miss one vital thing- no one cares about linearity of power as efficiency is often less than 0.5%. Than one should analyse the relationship between quantities being transduced e.g. voltage to velocity (v=BlV), current to force (F=BlI) where B is magnetic induction and l is legth of a coil. Both current and voltege driving will suffer from Bl non-linearity then. I still trust voltage driving, which results in velocity; Force (achieved by current) is highly non-linear because of cone suspension.
regards
Hi Paulo,
Your 'energy conservation' reasoning triggered an explanatory thought.
With a tube amplifier loudspeaker generated back-EMF is partially dissipated within anode resistance and the amplifier's output impedance is too high to reflect out of phase components.
With ultra low impedance solid state output back-EMF that is no longer directly related to on-going drive remains trapped to ring within the cable/loudspeaker system, until re-transduced or dissipated by components or drivers.
Cheers ....... Graham.
Your 'energy conservation' reasoning triggered an explanatory thought.
With a tube amplifier loudspeaker generated back-EMF is partially dissipated within anode resistance and the amplifier's output impedance is too high to reflect out of phase components.
With ultra low impedance solid state output back-EMF that is no longer directly related to on-going drive remains trapped to ring within the cable/loudspeaker system, until re-transduced or dissipated by components or drivers.
Cheers ....... Graham.
Graham Maynard said:
Hi Graham,
Yes granted. There's probably close to 6 billion people who don't even have an idea what I - or you - are talking about. Your point?
Jan Didden
Graham Maynard said:
Hi again, Graham,
As much as I hate to admit it 😉 I do agree with you on this one.
Jan Didden
- Status
- Not open for further replies.