The under appreciated impedance curve

The quick answer: The weaker the amp, the more its output in the frequency domain will look like the impedance curve.

And for those who want to know, here is the proof of that, with a well known push pull tube into the Stereophile simulated Kantor speaker load (which is a moderate easy load)

Simulated Kantor load (solid black wavy line)
https://www.stereophile.com/images/archivesart/scan58.jpg

Push pull tube 4ohm tap frequency response (+ - 5dB!!)  into that same Kantor simulated load (solid black wavy line)
https://www.audiofast.pl/assets/Uploads/pict2-prima-rec.jpg

See the amps frequency response take a dive at 30hz, 200hz and 4.5khz (more of a tone control than a ruler flat response)
Now look at the Kantor loading at 30hz, 200hz and 4.5khz, see how the load dictates what the response of the amp is doing at the same frequencies??? No wonder they sound coloured!!!

Cheers George

Prof,

It really only takes a little math to understand how it works.


An ideal amp:


[ input voltage x gain ] = Output Voltage


Gain is often around 20x (26 dB)


The reality :


[Input voltage x gain ] = Voltage lost in output stage + Voltage at Speaker terminals


Tube amps will loose more in the output stage. You can analyze it more or less like series resistors.

Vr1 = Vamp out x (   R1 / ( R1 + R2))


Where R1 = Amp output impedance  and R2 = Speaker impedance


Of course, R2 varies based on frequency.

For Solid State, R1 may be < 0.001 Ohms. For tube it is often in the range of 1-2

The quick answer: The weaker the amp, the more its output in the frequency domain will look like the impedance curve.

Thanks Eric.  I guess that's a good rule of thumb.  Though I always hear "it's totally different between SS and tube amps how they react to speaker loads" so I think I'm still a bit in the dark.

I have heard various things about matching a tube amps output tap to the speaker impedance.

On one hand some say if you have an 8 ohm speaker, you should be using an 8ohm tap output for your tube amps or you aren't getting all the power from your amp.

On the other, some have said that outputting from the 4 ohm tap will generally entail better sound/less distortion at the expense of less full output power, so *if* you will still have enough power for your speakers from the 4 ohm tap, better to use it even with 8 ohm speakers.

Confusing things more to me are that stereophile measured the CJ premier 12 amps I own as actually dropping power output into higher impedances (139W into 4 ohm speaker load, only 98W into 8 ohm speaker load).  Though the review amps came set for 4 ohm tap output...I don't know if that explains it and if switching the amps to an 8 ohm tap would retain their full power output to an 8 ohm speaker load?

I've always over powered my speakers thinking it would take care of any dips.. if any…  I now own a 3-way speaker with a flat curve.. Salk Song3a.. Now I don't have to worry about it anymore..  Yeah, its pretty important to know this measurement.. 

that he could make it with any nominal impedance he chose.

This is true for all cones as well. Off the shelf components are often available at different impedance for exactly the same reasons. Lower impedance = more current = more force against the magnetic field.


However, I have in fact seen Focal create a crossover deliberately to burn power. It had 80 W worth of resistors that didn't need to be there.

Bruce Thigpen stated in his discussion on the design of his magnetic-planar driver, the Linear Field Transducer (LFT), that he could make it with any nominal impedance he chose. Bruce says he chose 4 ohms for the LFT-3, LFT-4, and LFT-6 so as to enable the power amp driving them to provide them with as much power as they are capable of producing (assuming the amp is solid state, of course. SS amps produce more power into lower impedances than into higher. Most tube amps are the opposite.). He doesn’t say why he chose 8 ohms for the LFT-8 (and improved incarnation the -8a and current -8b), but by virtue of it being a nominal 8 ohm speaker, the LFT-8/a/b is more suitable for use with a tube amp than are the magnetic-planar driver 4 ohm Magnepans. By the way, if bi-amped, the LFT-8b’s magnetic-planar midrange driver and ribbon tweeter themselves presents an almost purely resistive 11 ohm load to the power amp, and are therefore an excellent candidate for use with a tube design.

I do think that the community at large does not appreciate how much we can learn.

I second that.

That doesn't mean to be an accusation of anyone in particular.

More more it's talked about, the more they'll make a effort to at least try understand them.

For those how wish to understand impedance/-phase graphs using Ken Kantor's easy simulated speaker loading. 
  
https://www.stereophile.com/reference/60/index.html

https://www.stereophile.com/content/measuring-loudspeakers-part-one-page-6

Cheers George

It isn't all a treatise on you, goerge.

I do think that the community at large does not appreciate how much we can learn.
That doesn't mean to be an accusation of anyone in particular.

The under appreciated impedance curve

Under appreciated?? I don’t think so, to those that find/look and know how to use it.

I believe "it" together with the combined "-phase angle" curve (EPDR) if hard to drive, "becomes the no1 criteria" to selecting the right amp to get the very best from that particular speaker. Probably that’s why it’s the 1st thing John Atkinson measures in Stereophile.

Cheers George

The quick answer: The weaker the amp, the more its output in the frequency domain will look like the impedance curve.

The whole speaker/amp impedance stuff drives me mad. I just seem to have a brain that doesn’t easily grasp electrical theory (not to mention I’ve never studied it).

I know the general principle for speakers seems to be: wild impedance swings - especially low - hard to drive, not great for tube amps.


Low sensitivity...hard to drive.

But I remain unclear exactly "why" this makes speakers ’hard to drive’ i terms of the actual sonic consequences.

I can more easily grasp low speaker sensitivity requiring more amplifier power/wattage. I can get "hard to drive" to some degree there, where an amplifier is going to run out of steam much sooner in terms of turning up volume.

But what are the sonic consequences of impedance curves in a speaker making it "hard to drive?"

My layman brain translates impedance variations this way: when impedance dips downward, at those particular frequencies it’s like the speaker becomes a lower sensitivity speaker, putting higher demand on the amplifier. So if you just went with a calculation for the demand on an amplifier’s power by looking mainly at the higher, flat points of an impedance curve - say it’s around 6 ohms, you could miss that it actually drops way down at one frequency to 2.5 ohms, in which case what you *really* need is an amp that will drive 2.5 ohms well, otherwise you live with possible frequency response deviations from the speaker. In other words, in figuring out the amp power requirements for a speaker, you should pay attention to the lowest dips in impedance, especially figure out where they happen in terms of their possible sonic consequences and if you need an amp that will drive *that* part of the impedance with necessary volume.

How off am I in that thinking?

I tried to grasp the consequences of impedance in that other great thread about amps, but I’m not sure I really got it.

That’s why I’ve used the comparison of two speakers, one I owned (Thiel 3.7) and one I’m interested in (Joseph Audio Perspectives).

https://www.stereophile.com/content/joseph-audio-perspective-loudspeaker-measurements

https://www.stereophile.com/content/thiel-cs37-loudspeaker-measurements


Looking at stereophile’s measurements and JA’s comments, he measures a pretty low sensitivity for the Joseph Perspectives, 84db, so my mind immediately thinks "hard to drive, needs a lot of power."

But then after the impedance measurements, JA says

"the Joseph is a very easy load for the partnering amplifier to drive. Not only is the phase angle relatively low, the impedance remains above 8 ohms at almost frequencies, with a minimum magnitude of 6.27 ohms at 135Hz. "

Oh...ok..now they are "easy" to drive. What does that mean for my Conrad Johnson Premier 12 tube amps? Not too powerful at 140W per side in terms of that speaker’s sensitivity, but the impedance is "easy" so....good to go?


Then there are the measurements for the Thiel 3.7s. JA measures a sensitivity of 90.7. So my brain says "Oh, easy to drive, doesn’t require much amp power to go pretty loud."
But then I read JA say:

Thiel specifies the impedance being nominally 4 ohms, with a minimum of 2.8 ohms. I actually found the minimum impedance to be 2.4 ohms at 125Hz. The difference between 2.8 and 2.4 ohms is academic, either mandating use of an amplifier that has no problem delivering high currents.

Whoops, now they sound "hard to drive." What does that mean for my tube amps?

Which is "harder to drive?" The speakers with the lower sensitivity and "easy" impedance curve? Or the speaker with the higher sensitivity and "harder" impedance curve? And what are the possible sonic consequences for each speaker in terms of being driven by a tube amp, like my CJ Premier 12s?

(My CJs seemed to drive the Thiels quite well btw...as well as any number of other speakers I’ve had, including even my small MBL speakers which are something like 82db sensitivity and I think fairly low....but somewhat even...impedance).

Thanks for any unraveling of my ongoing confusion...

What do you look for in the impedance curve?

The minimum value and the frequency at that value. Also, whether the curve is flattish over multiple octaves and, if so, the average impedance over that range.
That said, the map is not the territory. This is even more true for amp measurements. No way to know how amp+speaker will sound until you actually listen to the combo.

Anytime I get asked about the suitability of a speaker with our gear I always look for the impedance curve. All amplifiers make more distortion into lower impedances so I look for that sort of thing- it can be pretty detrimental in the bass where a lot of energy can be needed. The reason its so detrimental is distortion components can extend to some very high frequencies- and we might be talking about some very high orders- not just 7th but well above the 20th and the like. The ear finds this irritating!

You can reduce the amplitude of the harmonics by simply presenting the amp with an easier load. This has the effect of smoother sound with greater detail, regardless of the amp.

Tubes in general like to drive higher impedances solely due to a higher output impedance. Our smaller amps are the most popular, so I impedance curves are very helpful. I'm not a fan of Stereophile but their measurements of speakers have been very helpful in the past.

ESL's are notorious, and I've seen down to 1/3 of an Ohm at the top of the frequency range. It's something hopefully purchasers learn about before they buy.

Of course, from the dealer's' perspective, they should jump on this as it gives them a reason to sell a bigger amp.

I can't for the life of me understand why manufacturers aren't required to state minimum impedance, and would be nice to know what frequency it occurs at as well.  That's not a hard ask, is it?  Nominal impedance is a near USELESS spec to me.  

One example, my buddy's Martin Logans dip to 2 Ohms, but it occurs at 2Khz so not as tough a load as it would be at 200Hz.  This is such critical information to know when purchasing a speaker, I almost can't believe they don't make it more transparent. 

I cant think of a single reason why any speaker should be of a lower impedance other than perceived higher power with most S.S. amps. Fact remains that higher impedance speakers are almost universally easier to drive regardless of amplifier type. 

Stereophile has had a long love affair with Wilson. Gives me pause after numerous auditions of Wilson products. 

I agree that the slopes within the frequency plot should not be too steep. I believe an 8ohm nominal speaker plot should average ~8ohms and not dip below 4 ohms or reach greater than 16 ohms. Often this is not the case. I wish this  and phase angle information, were available for all speakers.  

  

 

Erik, I agree this measurement is overlooked by many audiophiles (except by the technically oriented).
I look to see if the curve is relatively flat. As you state, a clearly defined hump is common and usually acceptable.
If the specs of the speaker state 8 ohms nominal, I want to see the curve between 8 ohms and 4 ohms. And if it is close to 8 ohms throughout the frequency spectrum then the rated sensitivity can be given less importance.

I check to see if there are any drastic peaks in the low  frequencies (about 60Hz or less) and then ignore them since they are typically
resonances inside a wood cabinet speaker. At such low frequencies they should not affect bass reproduction.