"I repeat, every loudspeaker has an optimum amplifier (output impedance)."
Sorry but absolutely disagree with you on this.... and on most of the other points you raise....
Still, it would be a boring world if we all agreed about everything
Easy enuff to prove, next time you are in the neighborhood drop in.
dave
If we want build an amp for as many passive loudspeakers as possible, we better take a voltage amp. So it is clear why commercial manufactors do that. I would make it too.
But if we want to build an aktive speaker with aktiv crossover it is woth to think about current driving. Because now you developmental for the whole system with the speaker.
So i say, current driving is only for aktiv speakers and only for suitable speakers. You can also drive a horn with every speaker but it makes no sense.
And when you decide to build a current amp, you do not have any problems with damping.
But if we want to build an aktive speaker with aktiv crossover it is woth to think about current driving. Because now you developmental for the whole system with the speaker.
So i say, current driving is only for aktiv speakers and only for suitable speakers. You can also drive a horn with every speaker but it makes no sense.
And when you decide to build a current amp, you do not have any problems with damping.
Any competently designed commercial speaker should be designed to work at it's best when driven from a voltage source in order to be compatible with any amplifier.
IMHO if a speaker only sounds at it's best when driven from say 1R impedance then it is not competently designed.
The Catch-22 of the current market realities. I refered to it earlier as momentum,
One would be silly not to consider current driving all but the bass driver with a current amp to get the lowest distortion possible.
dave
Why? Only if you are not interested in maximum quality. Besides this is a diy forum, i give a rat's *** about commercial loudspeakers (except mining them for things to do, and not to do).
And besides it is a generalization that is not possible. Take a Bryston, a good PP pentode tube amp (one with lotsa feedback, and 1 with very little), a good SET, and a Firstwatt F1 or F2. Not many speakers at all that are compatible with all those.
That is an insult to all those very competent speaker designers catering to the (small, but growing) market for speakers compatible with SETs and amps like the Firstwatts mentioned. (including myself)
And now with the outlook of the potential for a step up in quality that Nelson's work with SITs promise, even more important.
dave
PS: what amps and speakers do you use?
I stand by everything I said.
I use a variety of amplifiers, mostly of my own design, all solid state apart from my Leak Stereo 20. The latter has been modified for best performance
Speakers? again a variety but I mostly use my KEF Reference 105.3/4 which funnily enough is one of the few speakers that would work with all the amps you mention as it is 93dB/W and has KEF's conjugate load matching and so is a pure resistive load!
As you will have gathered, I am not a fan of horn speakers, single speaker full range drivers or most valve amps and therefore we probably are at opposite ends of hi-fi's scale of beliefs!
Live and let live eh Dave?
You could but why would you want to? This was mainly done in the days of no decent complementary pairs. IIRC some McIntosh solid state amps still use transformer coupling! Not entirely sure of that though....
There is a fascinating Fostex FE166E IB loudspeaker design in the diyAudio articles section. Morgan Jones' Arpeggio loudspeaker:
http://www.diyaudio.com/forums/diyaudio-com-articles/158899-arpeggio-loudspeaker.html
What he does is exploit the 2-3 ohm output impedance of a small SET valve amplifier to get that unpromising 6" High efficiency Fostex unit working in a decent sized 20L box. He actually increases the output impedance to 6.5 ohms, which increases the Q of the driver to get some bass wallop without going all boomy! He then tames the rising top-end with a RC network to counter the RL nature of the top-end and keep the impedance flat.
Point is that you really couldn't normally use that Fostex unit in IB configuration with any bass extension with a highly damping voltage amp. Clever stuff from an ex-BBC sound engineer!
http://www.diyaudio.com/forums/diyaudio-com-articles/158899-arpeggio-loudspeaker.html
What he does is exploit the 2-3 ohm output impedance of a small SET valve amplifier to get that unpromising 6" High efficiency Fostex unit working in a decent sized 20L box. He actually increases the output impedance to 6.5 ohms, which increases the Q of the driver to get some bass wallop without going all boomy! He then tames the rising top-end with a RC network to counter the RL nature of the top-end and keep the impedance flat.
Point is that you really couldn't normally use that Fostex unit in IB configuration with any bass extension with a highly damping voltage amp. Clever stuff from an ex-BBC sound engineer!
sorry i had to cut short what i mentioned earlier, but i had to leave for work....
what i found was that you could start out at 20hz with an output impedance of 100microohms, and it would stay constant until you get to 200hz. in this frequency range, the primary contributor to the output impedance is the feedback factor. after 200hz, the feedback factor begins to drop, and the output impedance increases accordingly. as the feedback factor drops with frequency, the output resistance becomes proportionally more important. at the unity gain frequency of the amp, the output impedance IS the output resistance. so, while things that you do to improve the output resistance (paralleling output devices, lower emitter resistor values, bypass caps across the filter caps) don't matter much below the turnover point of the amp's open loop gain curve, they do matter at higher frequencies. what matters most overall, however is the open loop gain of the amplifier, and the amp's gain-bandwidth product. these two factors also matter most for minimizing distortion, so if you design for one (low distortion), the other (low output impedance) will be present as well. before the fire-arrows fly, i'm strictly talking about the effects of open loop gain and GBW here as they relate to output impedance and distortion (i know that a lot of other things must be done right for low distortion, but i'm singling out these two parameters because they have the same effect on distortion and output impedance, for the same reasons)
what i found was that you could start out at 20hz with an output impedance of 100microohms, and it would stay constant until you get to 200hz. in this frequency range, the primary contributor to the output impedance is the feedback factor. after 200hz, the feedback factor begins to drop, and the output impedance increases accordingly. as the feedback factor drops with frequency, the output resistance becomes proportionally more important. at the unity gain frequency of the amp, the output impedance IS the output resistance. so, while things that you do to improve the output resistance (paralleling output devices, lower emitter resistor values, bypass caps across the filter caps) don't matter much below the turnover point of the amp's open loop gain curve, they do matter at higher frequencies. what matters most overall, however is the open loop gain of the amplifier, and the amp's gain-bandwidth product. these two factors also matter most for minimizing distortion, so if you design for one (low distortion), the other (low output impedance) will be present as well. before the fire-arrows fly, i'm strictly talking about the effects of open loop gain and GBW here as they relate to output impedance and distortion (i know that a lot of other things must be done right for low distortion, but i'm singling out these two parameters because they have the same effect on distortion and output impedance, for the same reasons)
While there is any loop gain (i.e. across the audio range for most SS amps) the bare output impedance of the output stage is reduced by the feedback, so it is wrong to say that feedback sets the LF o/p Z and the output stage sets the HF o/p Z. The o/p Z at the feedback point is set by both together. After that, any output network may modify it.
re-read whai i said, carefully. that's exactly what i said. the output resistance is a larger portion and has more effect on the output impedance at higher frequencies, but the output impedance doesn't match the output resistance until the unity gain frequency is reached (and we don't use audio power amplifiers at their unity gain frequency anyway). even in a poorly designed amp with a GBW of 600khz, the output impedance at 20khz will be (assuming an Av of 30 for a 100W amp) Rout/30. or 1/30th of the output resistance since at that point the feedback ratio would be about the same as the closed loop gain because we're at the turnover point for the closed loop gain).
edit: i had written something else there, but caught my math error.....
Last edited:
When it comes to speakers designed by ex-BBC engineers I like these:
PMC Ltd
However they use Brystons rather than SETs and T/L rather than IBs.
Very popular with the worlds leading classical record labels.
Anecdotally all I can say that whenever I replaced one amp with one having a higher DF bass SQ improved. The first one I used had a DF of 50, followed by 266 and eventually 400. But I can't really draw any fast conclusion from this as the overall amp quality also improved. However using two identical amps, one bridged one not, had me and their owner prefer the non-bridged amp due to the slightly tighter bass. The amps were MC2 Audio E25s so I don't think the difference in potential power output played a role.
We used passive Tannoys for this 'experiment'.
PMC Ltd
However they use Brystons rather than SETs and T/L rather than IBs.
Very popular with the worlds leading classical record labels.
Anecdotally all I can say that whenever I replaced one amp with one having a higher DF bass SQ improved. The first one I used had a DF of 50, followed by 266 and eventually 400. But I can't really draw any fast conclusion from this as the overall amp quality also improved. However using two identical amps, one bridged one not, had me and their owner prefer the non-bridged amp due to the slightly tighter bass. The amps were MC2 Audio E25s so I don't think the difference in potential power output played a role.
We used passive Tannoys for this 'experiment'.
I have no idea what you mean by this. Damping factor is precisely defined. It fairly directly relates to electrical Q, which in turn combines with mechanical Q to get the desired total Q. No speculation needed. Choosing total Q is a matter of taste and judgement, then the rest is engineering.Nico Ras said:
O/p Z = (bare o/p Z)/(loop gain) so both contribute equally at all frequencies. Halving 'bare o/p Z' will halve o/p Z, so exactly the same proportion. This may be what you meant, but it didn't appear to me to be what you said.unclejed613 said:
What do I mean, answer the thread question with some facts instead of stupid comment like "I have no idea what you mean". Damping factor can equate to less bass?
"O/p Z = (bare o/p Z)/(loop gain)". Define bare output Z as opposed to some other form of output Z. Does bare output Z include the physical series emitter resistor, or does Re become near zero in bare Z.
Previously you rejected when I said that paralleling output devices results in decreasing output impedance, here you suddenly agree, so what is it? Is there any other way that you would increase the output current of the amp(and thereby lowering the output impedance) to drive and better control lower impedance speakers.
So I say that damping factor does have a definite affect on the capability of controlling the bass performance and sound of a speaker or speaker system.
When the damping factor is low (i.e. high output impedance of the amp) bass would appear to increase due to the speaker flapping around and the note slowly decaying until it comes to rest naturally. Having a high DF (low amplifier output impedance relative to the speaker impedance) would increase control and therefore tightness of the bass so it would appear to have less bass. As you say you decide on the system Q, Inother words what you choose or what bass pleases you and design around this.
Obviously, a critically damped circuit are optimal which is achieved by matching drive impedance with device impedance, but this would be almost impossible to do in audio unless you design both amp and speaker drivers to work together perfectly. This can be achieved by means of speaker driver being included in the negative feedback path, which was pioneered by ServoSound in the late 60's if my memory serves me right.
Last edited:
several times I have tried to 'cure' a poorly designed speaker with sloppy bass, by using differnt amp designs
it never worked
small differences, yes
but by far not enough
no magic happened
bass remained sloppy
but I have heard barely unlistenable thin sounding speakers on SS amps, but really shined with tube amps
so I guess it works ok the opposite way
it never worked
small differences, yes
but by far not enough
no magic happened
bass remained sloppy
but I have heard barely unlistenable thin sounding speakers on SS amps, but really shined with tube amps
so I guess it works ok the opposite way
when I built up my world audio KLS3g, I chose the box size using Bullock on Boxes as my design guide.
I knew that if the Q turned out too high, that I would have a boomy bass sound. I had enough of those and arranged that a small value of extra resistance would be needed to get the box Q to a value I predicted would be about right.
In the end I used 0ryy to equalise the measured box Q in one speaker cable and then a further 0rxx in both speaker cables to bring the box Q to what I thought sounded right. The speakers were tri-wired without speaker terminals from go and thus the added resistor did not affect the Mid nor Treble crossovers.
I did have slide adjustable ports and added mass on the bass cones and still I had to reduce box volume by packing out the bottom of the towers. But I got there. The 0rxx and 0ryy resistor values were quite small.
Basically, I ensured that I didn't need negative damping factor.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.