Amplifier Output Impedance (Damping Factor) and Speakers

[pogo]

Audio reproduction has many factors and variations for all the graphs and measured data. Our ears are the ultimate measuring system.

Correct, testing is not incorrect in what it measures, but in what it does not measure.

 

[KSTR]

Can you deposit 'vary a bit' with numbers/facts? This goes exactly in the direction where I never received any feedback from Benchmark, see also here: Link

For example, in the most basic power amplifier -- the single ended voltage follower (MOSFET source follower or Bipolar emitter follower -- the output impedance varies with the momentary current through that MOSFET or Bipolar transistor. The higher the momentary current, the lower the output impedance.

 

[Suffolkhifinut]

Indeed output impedance of an amp (and its rescaling to "damping factor") can vary a bit with dynamic operating conditions (which directly translates to additional distortion).
Still output impedance does never dominate the damping the drivers sees as long as it remains low in comparision to the driver impedance. Below 1/10th (DF > 10) it has zero effect on damping...

Recently read a review on the EAT E-GLO I tube integrated amp with an output impedance that never falls below 1 Ohm. The reviewer Ken Kessler describes the sound as ‘warm and cuddly, could it be underdamped?

 

[Suffolkhifinut]

Correct, testing is not incorrect in what it measures, but in what it does not measure.

Sometimes when listening it isn’t what we hear, it is what’s missing.

 

[SIY]

Correct, testing is not incorrect in what it measures, but in what it does not measure.

Sometimes when listening it isn’t what we hear, it is what’s missing.

Audio as fortune cookie.

 

[KSTR]

Recently read a review on the EAT E-GLO I tube integrated amp with an output impedance that never falls below 1 Ohm. The reviewer Ken Kessler describes the sound as ‘warm and cuddly, could it be underdamped?

No real correlation to damping here. If anything, the reviever just noticed the frequency response change of that amp with his speakers vs an amp with high DF... distortion profile might also have played a role.
As this was most certainly a sighted review, the dominant cause might well have been just alone the ususal bias that comes with sighted testing.

 

[Geert]

For example, in the most basic power amplifier -- the single ended voltage follower (MOSFET source follower or Bipolar emitter follower -- the output impedance varies with the momentary current through that MOSFET or Bipolar transistor. The higher the momentary current, the lower the output impedance.

But in practise (closed loop) not with a factor 5 to 10 as Pogo thinks, no?

 

[Suffolkhifinut]

Audio as fortune cookie.

Some of us spend more time listening rather than looking at graphs.

 

[SIY]

But in practise (closed loop) not with a factor 5 to 10 as Pogo thinks, no?

Even that has minimal effect, especially when you add in the inevitable cable resistance. There's likely 50-100 milliohms right there.

 

[SIY]

Some of us spend more time listening rather than looking at graphs.

I've seen that cookie before. I think it was at Golden Dragon, just outside of Baltimore.

 

[mhardy6647]

Hegel amps have a very high damping factor and generally get very good reviews. Is it a coincidence?

No idea -- but I will add that retail price per watt is often a confounding variable when "review outcome" is a parameter of interest.


Are they expensive*?

_______________
* I don't know the brand, so I have no idea.

 

[mhardy6647]

I seem to remember the original AR amp had transformer coupled speaker outputs? I have to say I loved it back then...

(EDIT) The AR Amplifier, like many of the early ss amplifiers (of which it wasn't - particularly - one), used transformers for interstage coupling.
No OPTs in the AR Amplifier.

 

[DonH56]

Some of us spend more time listening rather than looking at graphs.

I see this implied insult all the time, but realized with a 60-80 hour workweek I probably do indeed spend more time per week looking at graphs, schematics, lab equipment, test results, and all of that rather than listening. Does not make the graphs wrong, nor my listening sessions invalid, but guilty as charged. I remember when engineering knowledge was valued rather than being phrased as an insult, sigh.

Calculating dynamic output impedance is challenging (yes, done it, years ago) but large-signal circuit simulators do a good job of it. As mentioned earlier it is such a small effect that it is usually neglected. Feedback handles it nicely, natch. Along with changes due to supply voltage changes, temperature changes, process variations, and a host of other factors that influence output impedance (and thus damping factor). Where it typically matters is at clipping especially if the feedback loop opens due to slew limiting or whatever. Then things can get squirrely.

Output coupling: there are a number SS amplifiers today that are capacitor coupled, but the only SS transformer-coupled brand I can think of off-hand is McIntosh and their output autoformer (balun).

 

[mhardy6647]

Output coupling: there are a number SS amplifiers today that are capacitor coupled

Really? Is this a class D thing?
I would've thought that single-rail class AB soiled solid state amplifiers with DC blocking coupling caps went out with the Swingin' Sixties.
Capacitors are not good things when it comes to audio output.

HH Scott LK60 DSC_0789 by Mark Hardy, on Flickr

The Scott LK60 shown above (and its ready-built sibling, the 260) actually used a quasi-complementary circuit architecture, but since both "halves" rode on one side of zero volts, DC blocking output capacitors were required -- and it's what I had a handy photo of!

Audion Circuit Review

 

[DonH56]

Really? Is this a class D thing?
I would've thought that single-rail class AB soiled solid state amplifiers with DC blocking coupling caps went out with the Swingin' Sixties.
Capacitors are not good things when it comes to audio output.

HH Scott LK60 DSC_0789 by Mark Hardy, on Flickr

The Scott LK60 shown above (and its ready-built sibling, the 260) actually used a quasi-complementary circuit architecture, but since both "halves" rode on one side of zero volts, DC blocking output capacitors were required -- and it's what I had a handy photo of!

Audion Circuit Review

No, most class-D amps have a series-L/shunt-C filter network, though some car audio amps probably use capacitor coupling with the common-mode voltage at half the 12 V'ish battery supply. There have been some articles on AVS about pro amps and I noticed several were AC-coupled. Don't recall the brands off-hand, Behringer might have been one, can't remember the other. It also enables bridged outputs to be quasi-ground referenced with some topologies.

 

[pogo]

Even that has minimal effect, especially when you add in the inevitable cable resistance. There's likely 50-100 milliohms right there.

Are there any valid measurements on this and, if so, on which equipment was this carried out?

 

[Geert]

 

[SIY]

Are there any valid measurements on this and, if so, on which equipment was this carried out?

Your parody skills need work.

 

[Suffolkhifinut]

I see this implied insult all the time, but realized with a 60-80 hour workweek I probably do indeed spend more time per week looking at graphs, schematics, lab equipment, test results, and all of that rather than listening. Does not make the graphs wrong, nor my listening sessions invalid, but guilty as charged. I remember when engineering knowledge was valued rather than being phrased as an insult, sigh.

Calculating dynamic output impedance is challenging (yes, done it, years ago) but large-signal circuit simulators do a good job of it. As mentioned earlier it is such a small effect that it is usually neglected. Feedback handles it nicely, natch. Along with changes due to supply voltage changes, temperature changes, process variations, and a host of other factors that influence output impedance (and thus damping factor). Where it typically matters is at clipping especially if the feedback loop opens due to slew limiting or whatever. Then things can get squirrely.

Output coupling: there are a number SS amplifiers today that are capacitor coupled, but the only SS transformer-coupled brand I can think of off-hand is McIntosh and their output autoformer (balun).

Like you spent most of my working life in the electronics industry. When it comes to HiFi the only instruments that really matter are the ones on either side of your head. Human being make subjective judgements throughout life and live by them whether it’s relationships, work, hobbies etc. If we get back to HiFi I can’t stand the Naim take on sound reproduction yet love the sound of TAD gear, wish I could afford some. Many other people love the sound Naim equipment and wouldn’t wish for anything else. Yet your measurements of Naim gear could never take into account my judgement or theirs. Not intended as an insult just pointing out the way it is for us all.

 

[sergeauckland]

Like you spent most of my working life in the electronics industry. When it comes to HiFi the only instruments that really matter are the ones on either side of your head. Human being make subjective judgements throughout life and live by them whether it’s relationships, work, hobbies etc. If we get back to HiFi I can’t stand the Naim take on sound reproduction yet love the sound of TAD gear, wish I could afford some. Many other people love the sound Naim equipment and wouldn’t wish for anything else. Yet your measurements of Naim gear could never take into account my judgement or theirs. Not intended as an insult just pointing out the way it is for us all.

Not for all of us! I've never cared what anything sounds like to me, only how it measures. If it measures well, that's all I care about. If then I don't like it, that's my fault, not that of the equipment. Then, choosing equipment is easy, just buy on facilities, build quality (which I accept is subjective) and ease of service. None of this agonising over what I think something sounds like.
S.