[Argoncat]

I have been ignorant of amplifier / speaker pairing for most of my audiophile journey but am starting to question that logic, specifically the importance & impact of amplifier output impedance or damping factor.  So true to this hobby, there does not appear to be a single theory around optimal damping factor as similar priced units from Vinnie Rossi (L2iSE) and Audionet (Watt) have radically different damping factors, ~80 (into 8 ohms) for the Vinnie Rossi L2iSE and ~1000 for the Audionet Watt.  Both highly respected machine builders but very different amplifier output impedance ratings.   My current (and much more affordable) Prima Luna Dialogue has a damping factor of <5, which seems extraordinary low in comparison to other amplifiers..  With respect to our GR Research speakers, NX-Studios in my case or NX-Otica's for many, how important is amplifier output impedance (damping factor) to maximizing sound potential from our chosen speakers?  Should I use damping factor as a guide in my search for an integrated amplifier or is it an inconsequential attribute of amplifier design?

[mkrawcz]

Higher damping factor, woofers go “thump”. Lower damping factor woofers go “boom”. Tube amps with output transformers tend to have a very low damping factor and the result is less tightness in the bass. But you get the lovely tube characteristics. I went with the hybrid PS Audio BHK 300s which have tubes on the input stage and MOSFETs on the output. These amps give you the best of both worlds.

[opnly bafld]

Higher damping factor, woofers go “thump”. Lower damping factor woofers go “boom”. Tube amps with output transformers tend to have a very low damping factor and the result is less tightness in the bass. But you get the lovely tube characteristics. I went with the hybrid PS Audio BHK 300s which have tubes on the input stage and MOSFETs on the output. These amps give you the best of both worlds.

What is higher and what is lower?

[HAL]

The damping factor of an amplifier is its output impedance.  It is a spec for any amp.  It varies with frequency, but is only reported as a single number.  Usually a measure of how well the amp will deal with the back EMF from the speaker generated by its motion.

The speaker impedance also varies with frequency.  These impedances are in the case of solid state amps usually orders of magnitude different.  For a tube amp output transformer about one order of magnitude different.  There is no way to match them unless you add a series resistor that just burns power, and you loose the ability to have the amp output stage damp the speakers back EMF.

If you are thinking about the speaker Q which is a Theil-Small parameter, that governs the response in a box or open baffle woofer loading. The lower the amplifier impedance, the more control over the back EMF of the driver and its damping you have.  Put a crossover between the amp and speaker and you do loose control over the drivers damping.

[mkrawcz]

What is higher and what is lower?

I don’t know. 

[HAL]

The damping factor of an amplifier is its output impedance.  It is a spec for any amp.  It varies with frequency, but is only reported as a single number.  Usually a measure of how well the amp will deal with the back EMF from the speaker generated by its motion.

The speaker impedance also varies with frequency.  These impedances are in the case of solid state amps usually orders of magnitude different.  For a tube amp output transformer about one order of magnitude different.  There is no way to match them unless you add a series resistor that just burns power, and you loose the ability to have the amp output stage damp the speakers back EMF.  There is no such thing as a high damping factor speaker, only the amp.  High would be 50+ and low would be below 50.  Some solid state amps have DF's of 1100 or more.  Your speaker wire has more impedance than the amps output stage.

If you are thinking about the speaker Q which is a Theil-Small parameter, that governs the response in a box or open baffle woofer loading. The lower the amplifier impedance, the more control over the back EMF of the driver and its damping you have.  Put a crossover between the amp and speaker and you do loose control over the drivers damping.

Sorry meant to edit the original post.

[Argoncat]

Did I understand this correctly, that use of a crossover network in the speaker design renders the amplifier unable to control driver damping and shielded from back EMF?  If this is the case, why worry about amplifier output impedance / damping?  HAL, perhaps that was your point and I missed it.. 

[HAL]

If you analyze the effect of adding a passive crossover network between an amp and speaker you will find that very quickly the amp output damping has dropped to the point it does not make any difference.  Even low gauge speaker cables will do that, so DF becomes a bit of a moot point.

The only way you really take advantage of the amps damping factor is if the drivers are directly connected with short wires to the amp with an active crossover driving the amp. I believe this is why single driver speakers and servo driver speakers are the optimal way to interface to speaker drivers.  The other issue with single driver speakers usually is the frequency response of the driver without any compensation network.  That is easy to correct with an active crossover.

[RPM123]

It has been my experience that the bass of some speakers sounds more "natural" with a low damping factor (DF) - say 50 or less. I have had owned DeVore Fidelity speakers for a number of years and they tend to sound more lifelike with amps with a low(er) DF. Which is why so many owners tend to prefer tube amps to drive them, although there are some SS amps with low damping that are fine martch.

I use a Sugden amp that has DF of around 22 and it sounds great with my DeVores. I once tried another SS amp with a DF of around 250 or so with DeVores and I thought the bass was way over damped sounding. The bass was so tight that it did not sound natural to me. There was no "bloom". It sounded rather...constipated. Some may prefer that type of sound.    I recall reading about some amps that have a switch to alter the DF so that it is suitable for a wider range of speakers. Too bad more manufacturers do not offer this feature.

[HAL]

I will explain my info in a simple example.

An amp with a DF of 50 is the same as saying the output impedance of the amp is 0.16Ohms.

If you take an 8ft run of 14awg speaker cable its DC resistance is 0.063Ohms.  That is ~1/4 the impedance of the amp output, and that impedance will go up with frequency. As pointed out the damping is most needed for the woofer control.  Now the DF is 36.  Smaller gauge cables will drop the DF farther.

[Rusty Jefferson]

Did I understand this correctly, that use of a crossover network in the speaker design renders the amplifier unable to control driver damping and shielded from back EMF?  If this is the case, why worry about amplifier output impedance / damping? ....

I don't think HAL is saying there will be no difference between how well an amplifier with low output impedance controls woofers compared to one with high output impedance. Many solid state amplifiers have damping factors between 1000-20,000 in the bass region. The high DF amplifiers can maintain better control of the woofers compared to a transformer coupled tube amplifier with a damping factor of 5-10, driving a variety of passive loudspeakers. The subjective difference will be tighter (possibly leaner depending on the speaker design) more articulate bass.

As RPM123 points out, some speaker builders design their speakers to perform a particular way to sound subjectively better with a high output impedance amplifier (TC tube amp), and also the opposite. I have a pair of studio monitors whose owners manual states "...these speakers were designed to be used with modern low output impedance amplifiers..." i.e. don't use a TC tube amplifier.

Damping factor is something to consider in amplifier selection for a given speaker, but for many speakers, enjoyment can be had using either type (high/low damping factor) because of our subjective preferences.  At least one equipment builder, David Berning, made a few models of amplifiers with variable output impedance to help his amplifiers better match speaker loads. I've done double blind listening tests with Dave and several of his amplifiers and easily heard the difference between the different output impedance settings.

[jjss49]

this element of system matching is wayyyy underappreciated and not well understood, even for 'experienced' hifi afficianados

[HAL]

Rusty,
Ok, let me take your example of an SS amp with a DF=1000.  To find amp output impedance, it is AmpZ=8/DF.

This means the output impedance is 0.008Ohms.  Using the same 14awg speaker cable example of 0.063Ohms, the cable is 12.5x the impedance of the amp.  The DF at the speakers is 112.  It is higher than the earlier example where the DF=50 to start, but it is not 1000. 

It will have better bass driver control than the first amp.  If that is what the speaker needs then short speaker cables and long interconnects is preferable.

What it says is that short speaker cables are much better for a speaker that needs bass control than long cable.  It also says that bass driver control will change with different speaker cables of different impedance.  That is system synergy.

[planet10]

Yes, this subject is often very misunderstood.

HAL has made some very good points.

Damping factor is a technical term invented by the markering department to make a typically small number, R_out, jnto a large number for th espec sheet.

To look at things in another way.

The R_out of an amplifier interacts this the impedance of a loudspeaker, which can vary wildly. Speakers with a flat impedance curve are rare if they have a crossover in them and almost no speaker will have flat impedance at the speaker resonance.

Note thatgiven cables and voice coils any damping specification above 20-40 is moot.

Today, most amplifiers have low R_out, say less than an ohm or so, and there is little affect of the impedance of speaker. Except perhaps at resonance. We talk about control, it is control of the resonant peak. There can be too much or too little.

But we see a growing interest in highish R_out amplifiers, most notably SETd, and other single ended amplifiers with little or no feedback such as the PASS ACA and the new VFET, the Firstwatt F1/F2 and a few much less known japanese and European SS amplifiers.

With these amplifiers we can look at the the affects in the bass, and higher up in frequency. These amplifiers offer less control” of the impedance and you often hear the word “boomy” associated with the use of such an amplifier with many speakers. Those speakers have been designed assumming a low R_out amplifier. The speaker has to be matched to the amp. Higher up a fraction of the speaker impedance is convolved with the FR (driven be a low R_out amp).



For instance, if you use a high R_out amp, sat a 2A3 SET, the huge peak in the impedance will be heard as a peak in the FR. Unfortunately that speaker is often recommended for those kinds of amplifiers.

Note that there are loudspeakers that expect to be driven by an amplifier with highish R_out, like the Fostex FExx6 series in any of the horns they ar ehappy in. And they do not have insignificant impedanc ebumps at the bottom but are well behaved above.

You will note my use of the term highish R_out. I am ignoring high highish R_out amplifiers which are current amplifiers. They are much rarer, and a whole subject on their own.

dave

[SoCalWJS]

this element of system matching is wayyyy underappreciated and not well understood, even for 'experienced' hifi afficianados

This x 100.
Synergy.

I know there are some people who are very good at identifying what combinations of equipment sound good together, but I think it is strictly through trial and error.

It would be fantastic if there were a better/easier/"Scientific" way to know in advance before you go through 200 Pre -> Cables -> Amp -> Cables -> Speaker combinations.

But on the other hand, that can make for a great Hobby. 

[planet10]

Was also going to mention thus planset.

It specifically looks at a BR and shows what the box  looks like if you consider the amplifier R_out.

http://wodendesign.com/downloads/King-of-Swingers.pdf

(for a low R_out amplfier teh box is 9L.



dave

[jjss49]

This x 100.
Synergy.

I know there are some people who are very good at identifying what combinations of equipment sound good together, but I think it is strictly through trial and error.

It would be fantastic if there were a better/easier/"Scientific" way to know in advance before you go through 200 Pre -> Cables -> Amp -> Cables -> Speaker combinations.

But on the other hand, that can make for a great Hobby. 

This factor is at the heart of the very divisive and idiotic notion espoused by some speaker makers (notably Harbeth/Alan Shaw) that all amps sound the same, any decent amp will do to enjoy xx speaker - well, into a resistive test circuit, maybe amps will sound the same - but NOT into real world speakers with multi element crossovers and drivers with voice coil/magnet structures that store energy, and present complex capacitive and inductive loads that 'push back' hard on the amp trying to exert influence on them

[HAL]

This factor is at the heart of the very divisive and idiotic notion espoused by some speaker makers (notably Harbeth/Alan Shaw) that all amps sound the same, any decent amp will do to enjoy xx speaker - well, into a resistive test circuit, maybe amps will sound the same - but NOT into real world speakers with multi element crossovers and drivers with voice coil/magnet structures that store energy, and present complex capacitive and inductive loads that 'push back' hard on the amp trying to exert influence on them

If you think of the output impedance of an amp, even with feedback, it varies with frequency and not a single number.  If the amp has a feedback network, the back EMF from the speaker cone when it returns to position after the signal stops has to be absorbed by the amps output stage and also into the amps feedback network.  The lower the amp output impedance (including the cables) the smaller the error voltage getting in the feedback network, so less error is corrected.  Since every amp's design for output stage and feedback network vary, they all deal with back EMF differently.  Some have Zobel and/or Theil networks after the output stage to damp the speaker back EMF for stability purposes.  Some are direct coupled from output stage to speaker cable.  It really is a system of the amp and speaker load and they do not all react the same. 

As a preference, I prefer a direct coupled output stage where the speaker crossove is in the feedback loop.  Those are designs like Parasound and some Pass Labs/First Watt amps.  There are tube amps that are similar in design.  Even better when the speaker is directly connected to the amp output and in the feedback loop.  In my experience planar speakers like that alot. 

It is a fun hobby as the whole reason behind it is better music.  That is the goal and there are lots of ways to do it. 

[planet10]

...the back EMF from the speaker cone...

It should be note din all this, that a speaker driver is current driven, and if you have a voltage amp, the impedance of the loudspeaker converts from voltage to current.

In a current amplifier there is no back EMF. And lots in a voltage amplifier. There is more a less a line between these 2.

dave

[HAL]

It should be note din all this, that a speaker driver is current driven, and if you have a voltage amp, the impedance of the loudspeaker converts from voltage to current.

In a current amplifier there is no back EMF. And lots in a voltage amplifier. There is more a less a line between these 2.

dave

Yes, understand.  Voltage amps are by far more prevalent than current amps for most users.