An amplifier & speaker are a system, they have to be symbiotic.
Althou not always the case, it is often that a speaker with a crossover will have a roller-coaster impedance response (as seen in Eric's plots). An amplifier with low output impedance (voltage amp) will ignore that, an amplifier with high output impedance (current amp) will vary in output power and have FR variations that mirror the impedance plot.
If you have a speaker with flat impedance response, typical of FR speakers, Joe Rasmussen’s latest multiways, a few others and active speakers as OllBoll describes, the amplifier impedance makes little difference. You can take advantage of amps that tend towards current amp status and these will often sound better (as Oli points out a speaker is a current device and using a current amp avoids using the impefect speaker impedance as an I/V converter).
One other thing to consider is that almost every speaker has an impedance rise at the rersonant frequency of the box/speaker. If the output impedance is too low one can get one can get an overall overdamped response and thin bass. It is also possible to use a highish output impedance amp and the resonance peak to get an extension in the bass response.
So, the amp needs to be matched to the speaker, there is no universal choice. In an ideal world, you will design a speaker with perfectly flat impedance and use a current amp. In the real world something like OliBoli describes is more realistic.
dave
Althou not always the case, it is often that a speaker with a crossover will have a roller-coaster impedance response (as seen in Eric's plots). An amplifier with low output impedance (voltage amp) will ignore that, an amplifier with high output impedance (current amp) will vary in output power and have FR variations that mirror the impedance plot.
If you have a speaker with flat impedance response, typical of FR speakers, Joe Rasmussen’s latest multiways, a few others and active speakers as OllBoll describes, the amplifier impedance makes little difference. You can take advantage of amps that tend towards current amp status and these will often sound better (as Oli points out a speaker is a current device and using a current amp avoids using the impefect speaker impedance as an I/V converter).
One other thing to consider is that almost every speaker has an impedance rise at the rersonant frequency of the box/speaker. If the output impedance is too low one can get one can get an overall overdamped response and thin bass. It is also possible to use a highish output impedance amp and the resonance peak to get an extension in the bass response.
So, the amp needs to be matched to the speaker, there is no universal choice. In an ideal world, you will design a speaker with perfectly flat impedance and use a current amp. In the real world something like OliBoli describes is more realistic.
dave
Thanks, matching amps to loudspeakers is very easy when your amp is not utter rubbish. Those with significant output impedance can benefit from flattening loudspeaker impedance as seen by the amp. These speakers have simple RLC and RC networks in parallel with amp out terminals tailored for exact impedance peaks. You can turn any loudspeaker with such networks into one that can be good partner to any amp, as long as nominal impedance is within acceptable limits that amp is comfortable working with. A dead short of a loudspeaker is obviously not our goal.
Happy New Year and God bless, my friends!
These networks can be physically placed anywhere you prefer, directly on amp OUT, or at speaker terminal IN from the outside of loudspeaker enclosure, or directly on crossover board where the low and high pass filters are, as long as it's electrically as depicted.
Attachments
Last edited:
Yes. Once upon a time, KEF implemented complex crossovers with networks like that, to make the impedance peak-free (I don't recall ever seeing measurements of the results). Their advertising quoted Richard Small as saying "We've made your amplifier twice as powerful!" and he got a lot of flak for it. Later on when I talked to him about that he laughed and said that yeah, perhaps that had not been a good phrasing.
What he meant was that many amplifiers have trouble delivering current into impedances with severe phase angles which many speakers have. The amps will exceed the transistors' SOA (Safe Operating Area),* distorting and/or going into protection mode. All kinds of evil things may happen at the outputs which are not good for sound quality or your speakers. The KEF setup would let many amplifiers deliver more actual power to the speakers. This is probably more relevant to receivers than to massive outboard amps. I don't think KEF does this any more, at least they don't market it, probably because the huge capacitors and inductors needed to try and notch out bass impedance peaks get really expensive.
*I don't know what the tube equivalent is. A KEF benefit with tubes would be flatter frequency response due to less impedance interaction.
and now to
- Low noise,*
- Low output impedance**
- Low distortions (there are many kinds, but we don't really know how relevant each is or what is a meaningful number)
- Rated to deliver significantly more power into 4 ohms compared to 8. This indicates the amplifier hopefully has enough "beefiness" to deliver into the varied impedance of an 8 ohm speaker ha ha
- Crosstalk
but really we have no idea how to measure the "sweetness" or various other aspects of an amp, and many folks don't believe in any such things. Personally, I suspect there are dynamic behaviors that we haven't conceived and thus don't measure, and distortions that don't get measured either. Another aspect that rarely gets measured is HOW the amp clips-gracefully? Huge positive and negative DC spikes? Ultrasonic oscillation? Usually we don't know.
What is definitely true is you can find a lot of amps whose adherents LOVE their sound, but the amps "measure bad"
*but watch out: many amps specify S/N (Signal-to-noise) by comparing their raw noise to maximum power. This is not too relevant-what IS relevant would be the noise level compared to 1 watt, since THAT is what you would hear out of a tweeter.
**Also not measured in standardized ways, especially when called "damping factor" PLUS this varies with frequency...
Tough question. I firmly believe there is still a lot of mystery about what actually makes good sound. I tend to want an amplifier that is an accurate amplifier, i.e. with:
- Low noise,*
- Low output impedance**
- Low distortions (there are many kinds, but we don't really know how relevant each is or what is a meaningful number)
- Rated to deliver significantly more power into 4 ohms compared to 8. This indicates the amplifier hopefully has enough "beefiness" to deliver into the varied impedance of an 8 ohm speaker ha ha
- Crosstalk
but really we have no idea how to measure the "sweetness" or various other aspects of an amp, and many folks don't believe in any such things. Personally, I suspect there are dynamic behaviors that we haven't conceived and thus don't measure, and distortions that don't get measured either. Another aspect that rarely gets measured is HOW the amp clips-gracefully? Huge positive and negative DC spikes? Ultrasonic oscillation? Usually we don't know.
What is definitely true is you can find a lot of amps whose adherents LOVE their sound, but the amps "measure bad"
*but watch out: many amps specify S/N (Signal-to-noise) by comparing their raw noise to maximum power. This is not too relevant-what IS relevant would be the noise level compared to 1 watt, since THAT is what you would hear out of a tweeter.
**Also not measured in standardized ways, especially when called "damping factor" PLUS this varies with frequency...
To kind of follow up on Dave's statements, head unit, while you are describing a beefy SS amp, you should keep an open mind. Some of the best amps/music I ever heard were from amps which were noisy and had high output impedance by comparison. Seductive liars if you will.
It is absolutely worth the experience, even if you don't buy them.
Best,
E
It is absolutely worth the experience, even if you don't buy them.
Best,
E
Perhaps you meant a sleeper? (looks bad, works well). Maybe the most common starting point for output impedance (not specific to speakers) is Zo=RL before considerations of the desired drive mode. Certainly that is the way it is with passive power transfer.
..and where RE of the speaker is 6.5Ω, anything below two-thirds of an ohm becomes insignificant. I'd also like to point out that while a Voltage source is a convenience in eliminating variations due to impedance, and that these variations should be eliminated (unless there is a reason to keep them, and there are some), ..this premise doesn't qualify as a justification to demand low output impedance.
What would be a reason (or more reasons) to keep impedance variations in regard to sq?
Well, that might be a rule of thumb but that does not say anything at all about the minimum impedance or if the speaker is critical. The crossover often creates an (unnoticed or simply ignored) dip in the impedance, sometimes even critical. And there are tweeters which become very low on impedance too. Both does not show on the Re, not at all.
In general, I agree. But these variations are only 'eliminated' if the amp can keep the voltage stable. The higher the output impedance, the quicker the voltage drops. That might not be significant on most speakers but there are some where it actually does matter. The power supply plays a role in that too if the current the speaker draws becomes very high. But that's not the only issue. On some critical or high Qt speakers you can hear an immense difference between low and medium output impedance amplifiers, the low output impedance amps often sound a lot more controlled, firm on such speakers. That means, there actually are some cases which justify the demand for low output impedance.
For example if the speaker was developed on such an amp. That's then a combination which may sound great but once moved to another amp, things start to get ..not so nice. There are remarkable numbers of commercial wideband and horn speakers without impedance compensation, intended for tube amps. These speakers are often the argument, the sheer 'evidence' why tube amps 'sound better'. Well, at least for some high-enders.
No, don't misunderstand me, I've got nothing generally against tube amps but these are the amps where the speaker development and it's enviroment becomes really important.Seeing we are talking about 'damping factor' of amplifiers, regardless of whether they are tube or solid state, the fact remains that the voice coil resistance is electrically in series with the voice coil inductance, which incidentally, does all the work. So for a nominal impedance of 8 ohms the voice coil resistance may well be 6 ohms more or less. This DC (Re) is essentially added to the output impedance of the amplifier, let alone that which is added to it by the passive crossover and wiring. So the end result is a damping factor of less than 1 for any amplifier output stage when connected to a loudspeaker of any design.
Amplifier damping factor is calculated with a purely resistive load of 8 ohms across the output terminals of the power output stage. It is meaningless when connected to a loudspeaker and quoted as a specification, it is typically used as a selling point, and not much more. Loudspeaker drivers with large magnet structures generate much larger back EMF's than weaker magnet structures and as such imposes a greater feedback signal (due to irregular cone movement) on the driving current that comes from the amplifier. The amplifier then attempts to correct any error in the cone movement by negative feedback to outgoing signal. Thus drivers with large magnet structures tend to be better damped in cone movement, particularly at the lower bass frequencies and thus 'tighter, bass but not so deep and loose.
So amplifiers with a huge 'damping factor' are of little benefit.
C.M
Amplifier damping factor is calculated with a purely resistive load of 8 ohms across the output terminals of the power output stage. It is meaningless when connected to a loudspeaker and quoted as a specification, it is typically used as a selling point, and not much more. Loudspeaker drivers with large magnet structures generate much larger back EMF's than weaker magnet structures and as such imposes a greater feedback signal (due to irregular cone movement) on the driving current that comes from the amplifier. The amplifier then attempts to correct any error in the cone movement by negative feedback to outgoing signal. Thus drivers with large magnet structures tend to be better damped in cone movement, particularly at the lower bass frequencies and thus 'tighter, bass but not so deep and loose.
So amplifiers with a huge 'damping factor' are of little benefit.
C.M
For example if the speaker was developed on such an amp. That's then a combination which may sound great but once moved to another amp, things start to get ..not so nice. There are remarkable numbers of commercial wideband and horn speakers without impedance compensation, intended for tube amps. These speakers are often the argument, the sheer 'evidence' why tube amps 'sound better'. Well, at least for some high-enders.
Thanks.
I think I get it now.
Due to the high impedance of the amp the resulting fr will not be ideal due to the varying impedance of the speaker.
Now if there`s a synergy of high amp impedance and speaker fr the result will be "good" anyway but it will be a different story on a different speaker/amp.
Not in my experience.
Apart from connector resistance, cable resistance, etc... which will influence df there is a big difference between a high df amp and a low df amp, especially on large subwoofer drivers.
Also cable diameter and length will matter more than one might think.
Of course df is not everything.
But it can make a significant difference.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.