Please explain amplifier output impedance


I have recently read a few loudspeaker reviews which mention that the speakers would likely work best with low output impedance (or high output impedance) amplifiers.

So, what measurement defines low output impedance (or high output impedance) on an amplifier? What's the numerical value of low and high output impedance, and what is "average"?

Also, what specification of a loudspeaker provides info that would indicate using an amplifier with particular output impedance?

Thanks in advance for explaining this in laymen’s terms. :)
tvad
Thank you, Duke.
The 27 ohm maximum was at about 3 kHz, which is not a good place to have a response anomaly
Indeed, although rather benign compared to a (expensive) branded speaker I onced measured with a friend...! And you had few components in the signal path messing things up, which is a bonus. After all, a push-pull (or an OTL?) amp may work wonderfully with that... who knows...
Mr Berning sez
By the way, (Sean,) what (..you..) would think of an amplifier with negative output impedance
Not to hi-jack the conversation -- but hasn't Mr Pass tried/experimented with something like that?
I repeat, a low output amplifier impedance does not guarantee proper speaker damping. It often actually overdamps the speaker. The subject of overdamping, critical damping and underdamping is a well-studied topic in engineering

I agree that damping is well studied in engineering, however, in engineering "damping" is NOT the same thing as the audio "damping factor" term used to describe an amplifier. Actually "damping factor" is simply a reflection on the amp output impedance relative to to the speaker load. This is an "indication" of how well an amplifier can be expected to handle a reactive load. It should not taken as gospel ...there is a lot more to an amplifier than damping factor.

Once the output impedance is low relative to the load then the "damping" of the driver (suspension/magnet/coil)and box will dominate the response...The amplifier simply puts the correct voltage over the speaker terminals which always pushes the driver towards the proper position (as determined by the audio signal) by supplying the appropriate current.

In theory more current can be supplied for a given amplifier voltage if the output impedance is lower...i.e more "push" in the right direction.

Apologies if I have oversimplified things but I don't see how amplifier damping factor can be regarded in quite the same way as the critical damped response that is often saught in most speaker designs.
Hi Sean, I can't think of an amplifier that that qualifies as 'universal use'. For any example you can think of I can think of a speaker it does not work with. On that account your argument seems to fall apart. It is also not true that I 'prefer vented speakers', in fact what I prefer is speakers that sound right regardless of their technology. I use headphones (Stax, Grado) and master tapes for reference to avoid the pitfalls that you (?might?) have assumed that I have fallen prey to.

We've used a variety of speakers over the years including acoustic suspension ('sealed'), horns, bass reflex, magnetic and ES planars. On all of these speakers it was easy to demonstrate that our amps were making correct bass against the typical power house transistor amps which do not.

I should also point out that I do use a set of design rules based on reality and easily proven. Nor do we lack in the measurement department for them (see our website for details). IOW we respect the human ear and the rules that *it* uses, rather than made up rules that it does not use.
David@davidberning.com: I understand that electrical damping, mechanical damping, amplitude linearity, transient response, loading characteristics, etc... are all different yet inter-related subjects. Achieving excellent results in one area may end up compromising performance to some extent in another area. This is where design variables come into play and the engineer / designer must do some juggling in order to achieve what they think will result in the best balance.

As far as your comments go about an amp with a negative output impedance, you are right, this is not that difficult to do. The end result would typically be a design that was highly stable into various loads, but not very linear in terms of transient response, dynamics and desirable sonics. Once again, there's a design trade-off involved due to the ( limited ) technology available to mankind as we know it today.

Having said that, i think that far too many concentrate on certain areas while neglecting others, resulting in equipment that might measure well under certain test conditions, but perform FAR less accurately / musically in the real world. That's because the product is not well rounded, but instead, concentrates on measuring / performing certain measurable tasks better than others.

If you would be so kind, can you refer me to what issue Stereophile reviewed the ZH-270 in? I know i have this issue, but haven't been able to locate exactly which one it is.

Atmasphere: There are no "perfect" amplifiers, so of course, NOTHING is "universal" by design. Having said that, some amps are FAR less sensitive to loading characteristics while providing more consistent performance from speaker to speaker and system to system.

Unless one is interested in working with "chameleon" components that change both electrical and audible measurements at the drop of a hat, selecting gear that is of a more "universal" nature makes building a reasonably decent sounding and consistent performing system a little easier. Otherwise, one runs into problems with crediting certain changes to one specific component ( or cable ) change, when in reality, that one change triggered other changes in the chain. This results in a cumulative effect, making it harder to both recognise and disect exactly what is going on and why.

When something like this happens, stability is compromised, resulting in a wider tolerance of potential results. There is also less potential to transfer said results into another system by following similar approaches with slightly different componentry. This is why so many "recommendations" here are NOT universal by nature, but extremely system and personal preference dependent.

An ideal amplifier would have an output stage that would mimic the load impedance that it sees on a real time basis, adjusting as dynamic conditions and frequency are altered. Such a design would require unlimited voltage and current potential with phenomenal transient response. We don't have anything anywhere near this, so all one can do is to choose the lesser of evils and shoot for some form of electrical consistency with pleasant sonic results.

As far as my belief / stated comments regarding your preference of speakers, i based my comments on the speakers that i've seen you set your gear up with at various shows. I had assumed that you would have selected support componentry that you not only find sonically desirable, but also brings out the best qualities of your gear. From what i can recall, what i have seen you working with under those conditions were all vented designs. Granted, show installations may not always be what a manufacturer would recommend under optimum installation conditions, but one would think that they would at least strive to deliver something close. Forgive me for making any assumptions in that area if i am mistaken. Sean
>
No worries Sean! I agree with much of your comments regarding the ideal amplifier also. The issue for me personally when I set forth on the designs I did was how to get around the ill effets of negative feedback. Some of our early designs used it, and we pressed the limits of the tubes to produce super wide open loop bandwidth, in an attempt to put the propogation delay effects out beyond 20KHz. The result was full power bandwidth to nearly 30 MHz!

It didn't solve the feedback issues though (which I have previously discussed). We had set the feedback up so that we could switch it. In the end we chose zero feedback, and in the process discovered that a different set of design rules emerges.

I've had some discussion of these rules and the competing (more accepted, Voltage Paradigm) rules in the past. Both have challenges to overcome- the Power Paradigm has to find ways to make amplifiers more universal as you say and the Voltage Paradigm has to find a way to get loop feedback to arrive at the input of the amplifier in time to make the correct adjustment, not an approximate one (in order to create a musical presentation). Both are serious challenges!