[Eric Strasen]

In an Email discussion with an electronics engineer friend, I mentioned that a mutual friend was building a tube amplifier for me, designed to be most efficient with an 8-ohm load for my ADS L-990 towers.
I got the following interesting reply from him, and I quote:
"You mentioned 8 ohms Z. You know that lower impedances in audio speaker systems are sort of a joke. If you ever run an impedance curve on any speaker system it varies all over the place with frequency, enclosure volume/shape, reflected waves in the room, and the political orientation of the technician. It's sort of a 'middle of the road' thing."

It never occurred to me to measure speaker impedance during operation. Guess I figured an 8-ohm speaker was always 8 ohms no matter what.

[neekomax]

In an Email discussion with an electronics engineer friend, I mentioned that a mutual friend was building a tube amplifier for me, designed to be most efficient with an 8-ohm load for my ADC L-990 towers.
I got the following interesting reply from him, and I quote:
"You mentioned 8 ohms Z. You know that lower impedances in audio speaker systems are sort of a joke. If you ever run an impedance curve on any speaker system it varies all over the place with frequency, enclosure volume/shape, reflected waves in the room, and the political orientation of the technician. It's sort of a 'middle of the road' thing."

It never occurred to me to measure speaker impedance during operation. Guess I figured an 8-ohm speaker was always 8 ohms no matter what.

Very interesting topic. 

I've been thinking about this a lot, as I've been researching different types of amps, and one of the things I've noticed is the crazy frequency response charts when a simulated speaker load is shown to certain tube amps in particular. I'm pretty crap when it comes to physics and electrical stuff, but AJ, the designer of my speakers, pointed this out to me when I first ordered them and we were discussing suitable amplification. This article talks about this in an understandable manner, at least to me.

For example, this is a modern 70 wpc pp monoblock KT88 tube amp that sells for about $2,800. From Stereophile:

"...output impedance was extremely high from the 8 ohm tap, at 8.5 ohms at treble and midrange frequencies, and rising to 9.1 ohms at 20Hz. It was around 4.5 ohms from the 4 ohm tap, and still 2.6 ohms from the 2 ohm tap. These impedances will maximize power transfer into loads that equal the transformer-tap rating, but will introduce large variations in frequency response with real-word loudspeakers. Into our standard simulated loudspeaker, for example, there were ±2.2dB variations in response from the 4 ohm tap (fig.7), rising to ±3dB variations from the 8 ohm tap (fig.9), and even ±1.6dB changes from the 2 ohm tap."



 

[Eric Strasen]

Neekomax,
Thanks for that most enlightening reference about SET amplifiers and how they react with speaker systems. A Canadian friend has built several high-power SET designs, using old triodes (50's) or super power transmitter tubes, such as 311CH's.
The two-channel stereo amp he is building for me will have PP milsurp 1625's in audio output, driven by 6SL7's (or 6SN7's -- my choice).
I am most curious how this will sound, as opposed to my current amplifier (Adcom GFA-555). Am also replacing my GFP-555 preamp with a Van Alstine Transcendence 8+, as soon as Frank builds it.
This will be my first venture back into tubes since the late 1960's, when I stupidly sold my Dyna Stereo 70 and PAS-3 for peanuts.
And maybe it's time to dive into my ADS L-990's to check out the crossovers. They are rapidly approaching middle age... 

[musinteg]

The difference between any valve amplifier and a solid state is the output impedance of the amplifier. Every amplifier has some sort of an output impedance. By this I mean that the amplifier , at the output stage has an internal resistance that ends up in series with the speaker.

Firstly, phase angle is going to be completely left out of the discussion.

Basic electronics says, if you have 2 resistors in series, the current through them will be the same, and the voltage across each resistor, will be a product of the current of the entire circuit ( I ) multiplied by the resistance that you are looking at ( R ).

now....you have the amplifier resistance, lets call it Ra...........it is in series with the speaker resistance Rs.

AMP -------Ra---------------0--------------!
                            amp + terminal       !
                                                           !
                                                           !
                                                          Rs
                                                           !
                                                           !
                                                           !
 AMP--------------------------0--------------

                             amp - terminal

I put the Ra before the amp + terminal because it is an internal characteristic of each amplifier.

lets do the math.............if Ra is 1 ohm and Rs is 9 ohms......the total circuit resistance is 10 ohms.

If the amp is generating 1 volt, we will drop 0.1 volts internally across Ra and the other 0.9 volts across the speaker Rs.

This holds true only when Rs is constant......but it isn't........all speakers ( some much better than others ) have changing Rs with frequency.

in a bass reflex box you will get Rs spikes 1 octave below bass port frequency, a dip at port frequency ( box tuning freq ), another spike approx 1 octave above port freq, then another dip through the upper bass and more changes at the various crossover frequencies.

HOLY COW.....what does this mean ?.....

The speakers impedance changes with frequency.......in a solid state amplifier, the Damping Factor refers to how low the amplifiers internal resistance Ra is . The higher the damping factor, the lower Ra.

So if Ra is lets say 0.1 ohm and the speaker has impedance spikes up to 99.9 ohms and dips down to 10 ohms ( i'm exaggerating for effect ).............

Given that same 1 volt output as in the first example.........

at the 99.9 ohm Rs peaks, 0.999 volts are across the speaker, only 0.001 volts are dropped internally inside the amp. So the amp is delivering 99.9% of the theoretical current to the speaker AT THE "spike " FREQUENCIES.

IF the amp had an Ra of say 2 ohms........we are going to drop a bit of voltage internally at the " spike " frequencies.....not so bad.....BUT at the frequencies where the speaker impedance dips to 10 ohms......holy cow......we are dropping 0.16 volts across the Ra and delivering 0.8333 volts to the speaker. So instead of transferring all the current ( and voltage ) to the speaker, you are losing roughly 16% of your current internally in the amplifier.

This is one of the biggest electrical differences between tube and solid state.

NOW......I'm not saying that tubes aren't pleasing to the ear......not at all !!

It's just that the output impedance of the transformer is very critical in working faithfully with a wide variety of speakers.
You may need to do a bit of trial and error test listening to see which speaker works in most harmony with your amp.

There are very few speakers with a nice flat impedance curve.

clear as mud.





 

[G Georgopoulos]

The difference between any valve amplifier and a solid state is the output impedance of the amplifier. Every amplifier has some sort of an output impedance. By this I mean that the amplifier , at the output stage has an internal resistance that ends up in series with the speaker.

Firstly, phase angle is going to be completely left out of the discussion.

Basic electronics says, if you have 2 resistors in series, the current through them will be the same, and the voltage across each resistor, will be a product of the current of the entire circuit ( I ) multiplied by the resistance that you are looking at ( R ).

now....you have the amplifier resistance, lets call it Ra...........it is in series with the speaker resistance Rs.

AMP -------Ra---------------0--------------!
                            amp + terminal       !
                                                           !
                                                           !
                                                          Rs
                                                           !
                                                           !
                                                           !
 AMP--------------------------0--------------

                             amp - terminal

I put the Ra before the amp + terminal because it is an internal characteristic of each amplifier.

lets do the math.............if Ra is 1 ohm and Rs is 9 ohms......the total circuit resistance is 10 ohms.

If the amp is generating 1 volt, we will drop 0.1 volts internally across Ra and the other 0.9 volts across the speaker Rs.

This holds true only when Rs is constant......but it isn't........all speakers ( some much better than others ) have changing Rs with frequency.

in a bass reflex box you will get Rs spikes 1 octave below bass port frequency, a dip at port frequency ( box tuning freq ), another spike approx 1 octave above port freq, then another dip through the upper bass and more changes at the various crossover frequencies.

HOLY COW.....what does this mean ?.....

The speakers impedance changes with frequency.......in a solid state amplifier, the Damping Factor refers to how low the amplifiers internal resistance Ra is . The higher the damping factor, the lower Ra.

So if Ra is lets say 0.1 ohm and the speaker has impedance spikes up to 99.9 ohms and dips down to 10 ohms ( i'm exaggerating for effect ).............

Given that same 1 volt output as in the first example.........

at the 99.9 ohm Rs peaks, 0.999 volts are across the speaker, only 0.001 volts are dropped internally inside the amp. So the amp is delivering 99.9% of the theoretical current to the speaker AT THE "spike " FREQUENCIES.

IF the amp had an Ra of say 2 ohms........we are going to drop a bit of voltage internally at the " spike " frequencies.....not so bad.....BUT at the frequencies where the speaker impedance dips to 10 ohms......holy cow......we are dropping 0.16 volts across the Ra and delivering 0.8333 volts to the speaker. So instead of transferring all the current ( and voltage ) to the speaker, you are losing roughly 16% of your current internally in the amplifier.

This is one of the biggest electrical differences between tube and solid state.

NOW......I'm not saying that tubes aren't pleasing to the ear......not at all !!

It's just that the output impedance of the transformer is very critical in working faithfully with a wide variety of speakers.
You may need to do a bit of trial and error test listening to see which speaker works in most harmony with your amp.

There are very few speakers with a nice flat impedance curve.

clear as mud.

That's a speaker problem!!,how can it be corrected?,only with another speaker!!
thanks for your nice post,cool...

[JLM]

Another vote for active (one amp per driver) or at least powered (amp between crossover and drivers) where the designer matches amp with driver or speaker (and maybe adds circuitry to compensate?) rather than us dummies trying to make best guesses on match ups.

Of course many other reasons to go active (more than powered) designs.

[dB Cooper]

I have always been a fan of the aperiodic enclosure design approach, the best known example probably being the Dynaco A-25 bookshelf speaker. These have an impedance variation of about 2:1 as opposed to the more common variations of as much as 8:1 in most speakers. This makes a much easier load to drive, especially for tube amps.

[Eric Strasen]

Anyone seen Parts Express' latest spam? They are now flogging the "ButtKicker", which for mere hundreds of dollars can convert your floor or favorite overstuffed chair into a loudspeaker, sort of, by bolting the 1,900-watt transducer to same.

[cheap-Jack]

Hi.

This will be my first venture back into tubes since the late 1960's, when I stupidly sold my Dyna Stereo 70 and PAS-3 for peanuts.

Yes, it was indeed a shame to sell off the vintage jewels: Dynaco St-70+PAS-3.

Quite resverse, a friend donated me FREE the sampe pair of stock St-70+PAS-3, which sounded hummy, noisy & veiled 6 years back. Thank goodness, I being a DIYer, fixed them & upgraded them bigtime. Now
they both serve as my main amps which sonically blow away the SS DC power amp (DC-200KHz) which now sits idling.

c-J

[Duke]

Some amplifiers approximate constant voltage sources, and some amplifiers approximate constant power sources.   Likewise, some speakers are designed with constant voltage in mind, and some with constant power in mind.  In general, solid state amps behave more like voltage sources, and tube amps behave more like power sources.   

In English, that means some speakers are designed to go with solid state amps, and some are designed to go with tube amps.

A solid state amp will deliver increased wattage into an impedance dip, and decreased wattage into an impedance peak.  Put a tube amp on a speaker designed for solid state, and the frequency response may be messed up.  Unfortunately the tube amp gets blamed because the system was fine before, when really it was just poor system matching.  Likewise a speaker designed for tubes may not sound very good on solid state, but again it's just poor system matching.

For a better explanation than I've offered, here's a well-written paper on the subject.  Regardless of which type of amp you prefer, this is great information:

Voltage Paradigm and Power Paradigm

How do you make a speaker that works with both types of amps?  Keep the impedance curve as smooth as possible, and if it's a ported box, maybe make the bass tuning user-adjustable to work well with a wide range of amplifier output impedances.

[G Georgopoulos]

The amplifier output impedance is constant,this is the case,and can be confirmed when the amplifier is working into a resistive load,the problem with speaker impendance is a problem of the speaker itself,whether the amplifier is tube or solid state is not important,the amplifier works perfectly,most sensitive speakers have smoother frequency response,less sensitive speakers have more peaks and dips,please note that acoustics can make peaks and dips even if the impedance is smooth this is a problem of room acoustics.

[Duke]

... the problem with speaker impendance is a problem of the speaker itself, whether the amplifier is tube or solid state is not important, the amplifier works perfectly...

Unless the speaker's impedance curve is perfectly flat, the system's (speaker's) frequency response is partially a function of the interaction between the speaker's impedance curve and the amplifier's output impedance.  In general, a tube amplifier will interact with the speaker's impedance curve differently from the way a solid state amp will, so the type of amplifier does indeed matter from a system matching standpoint.  Please take a look at the article I linked to in my previous post.

... most sensitive speakers have smoother frequency response, less sensitive speakers have more peaks and dips...

Actually the relationship between speaker sensitivity and frequency response tends to be just the opposite.  Drivers with high internal damping (and thus smoother response) tend to have heavier diaphragms (and thus lower sensitivity).  Also, as a speaker designer, I can smooth out the peaks by means of passive equalization in the crossover (the dips remain), but that will lower system efficiency, not raise it. 

[G Georgopoulos]

Actually the relationship between speaker sensitivity and frequency response tends to be just the opposite.  Drivers with high internal damping (and thus smoother response) tend to have heavier diaphragms (and thus lower sensitivity).  Also, as a speaker designer, I can smooth out the peaks by means of passive equalization in the crossover (the dips remain), but that will lower system efficiency, not raise it.

Hi Duke,thanks for your input,if you had two speakers one with bigger magnet and thus high sensitivity and another with smaller magnet, which would you prefer ???

I ask because the trend now is high sensitivity speakers...

[Duke]

Hi Duke,thanks for your input,if you had two speakers one with bigger magnet and thus high sensitivity and another with smaller magnet, which would you prefer ???

I ask because the trend now is high sensitivity speakers...

Well a lot of things come into play, and imo the key is not going so far pursuing excellence in one area that we end up with a fail in another.   Yes efficiency increases with increasing magnet strength, but low-end response gets correspondingly weaker, though we can often regain much if not all of that low end by increasing box size.  But if we go too far in the direction if high sensitivity, increasing the box size still further won't make up for the loss of low-end response, and our system starts to sound thin unless we add a subwoofer.   And now the price just went up, especially for a high quality subwoofer system that can subjectively "keep up". 

I personally gravitate towards medium-high efficiency woofers, hopefully combining decent low-end extension with good dynamics, even though the combination imposes a larger-than-average box size.   But that is by no means the only viable approach - it's just where a market opportunity seems to collide with something I think I can do reasonably well.

[G Georgopoulos]

Thanks Duke, you're a champion..

[Duke]

Thanks Duke, you're a champion..

Ha!  Thanks, but I just talk a good game, that's all.

[G Georgopoulos]

keep up the good work DUke,thanks