The Significance of Speaker Impedance – Avantgarde’s Change from 8 to 18 Ohms

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
When I first started dabbling with loudspeaker design and my own DIY efforts, virtually all hi-fi speakers were rated at 15 ohms.

Then rather rapidly the industry moved to 8 ohm speakers.

Why? Was it the transition from tube to solid state amplification that prompted this wholesale change in speaker design?

And why have a few brands more recently gone the other way?

Example Avantgarde horn speakers. Avantgarde didn’t exist in the old 15-ohm days, but their Unos, Duos and Trios were all introduced in the late 1990s at the then standard 8 ohm nominal impedance. Then in about 2006, their new Omega range saw the drivers changed to nominal 18 ohms.

So much was claimed of these new drivers that many Duo owners parted with significant sums of hard-earned cash to change the drivers in their speakers to Omega ones - and everyone crowed over how huge the improvement was. Perhaps they had to say that after spending so much, but let’s be generous and agree on this as a significant improvement in sound quality.

Perhaps the new drivers were simply better irrespective of their new impedance, but Avantgarde speakers remain at this 18 ohm impedance to this day.

Why is this? Will other brands change to use higher impedance speakers or will Avantgarde remain the odd man out? Is it because these high sensitivity horn speakers are so often powered by low powered SET tube amps? This seems a plausible explanation until you consider that Avantgarde’s own amps are all solid state.

I’d be very interested to learn more about the significance of speaker impedance, especially as I’m about to change my old 8 ohm Duos for new 18 ohm ones. Thanks. Peter
 
  • Like
Reactions: jeff1225

Alrainbow

Well-Known Member
Dec 11, 2013
3,189
1,387
450
High impedance drivers were really there to suit tooby amps , OTL ‘s loves them for eg. lower impedance drivers were really better suited for high current driven SS amps, Drivers for the amps of the day for sure ...
 

Robh3606

Well-Known Member
Aug 24, 2010
1,471
462
1,155
Destiny
I don't think there are any obvious differences between say 4-8 or 16 ohm drivers. There are drivers available with multiple impedance options. If you want to parallel up drivers you would go with 16's. If you wanted to squeeze as much power as you can from an amp for sub-woofer duty 4 would be the best option. Depends on what you are doing.

Rob :)
 

Brad Lunde

Member
Sep 18, 2020
55
65
23
Las Vegas
www.lonemountainaudio.com
Keep in mind speakers are not one single steady state impedance, impedance varies with frequency. Sometimes drivers can vary way more than you know. What is spec'd as an "8 ohm nominal impedance driver" can be anything but across its bandwidth.
Brad
 

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
Keep in mind speakers are not one single steady state impedance, impedance varies with frequency. Sometimes drivers can vary way more than you know. What is spec'd as an "8 ohm nominal impedance driver" can be anything but across its bandwidth.
Brad
Yes and no. Yes speaker impedence will vary with frequency, but a NOMINAL 8 ohm speaker will vary a bit around that 8 ohms, but an 18 ohm one will be in a very different impedence range despite its own variations. Don't you agree? The earlier 8 ohm Avantgarde drivers are far different from the present 18 ohm ones. I'm trying to establish why they have changed. They seem to be one of very few speaker builders to move back to 1960s preferred impedence.
 

MPS

Well-Known Member
Jun 20, 2016
112
84
160
Finland
I don't know Avantgarde's reasoning behind 18 ohms impedance but generally most (?) amplifiers produce less distortion to 8 ohms or higher loads compared to 4 ohms or lower.
I'm currently running my AA Duo XDs from 8 ohm taps on my EAR 861, it sounds cleaner and more accurate this way (maybe slightly more "hifi" as well) than when connected to 16 ohm taps.
Then again I preferred to run ML CLX from 8 ohm taps compared to 4 or 16 ohm taps on my ARC VT200mkII.
Anyways it's very easy to try different combinations and see what works for you.

P.S.
If I remember correctly, actual minimum impedance shouldn't drop to less than 50% of the nominal impedance. So 16 ohm speaker should not present lower than 8 ohm load at any frequency to an amplifier.
 
  • Like
Reactions: christoph

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
I don't know Avantgarde's reasoning behind 18 ohms impedance but generally most (?) amplifiers produce less distortion to 8 ohms or higher loads compared to 4 ohms or lower.
I'm currently running my AA Duo XDs from 8 ohm taps on my EAR 861, it sounds cleaner and more accurate this way (maybe slightly more "hifi" as well) than when connected to 16 ohm taps.
Then again I preferred to run ML CLX from 8 ohm taps compared to 4 or 16 ohm taps on my ARC VT200mkII.
Anyways it's very easy to try different combinations and see what works for you.

P.S.
If I remember correctly, actual minimum impedance shouldn't drop to less than 50% of the nominal impedance. So 16 ohm speaker should not present lower than 8 ohm load at any frequency to an amplifier.
Hi MPS. Good to hear from a fellow Duo owner. In fact I no longer use tube amps, although I still have 845-based monos, so I have no choice of amplifier tap. I think the higher the impedence, the "easier" the load is on the amp.

I also have Martin Logan 13As that I'm trying to sell and these need an amp that's happy with loads that drop as low as 2 ohms at some frequencies. I powered these with Sanders or GamuT amps - both now sold.
 
  • Like
Reactions: MPS

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
An interesting explanation. Thanks for sharing. When I started getting into hi-fi, speakers' efficiency was often expressed as a percentage. So if 10 watts were put into a speaker and 9 watts ended up as heat, the sound output was 1 watt, so 10% efficient. Lowther managed up to about 55% whereas Wharfedale (whose Airedale syatem I home-built from Gilbert Briggs's drawaings) were about 25%. Efficiency has given way to sensitivity, rightly or wrongly, and it's interesting to learn that the method of measurement has changed as per the video you posted.

I've not attempted to work it out per the formulae in the video, but my Duo XDs are rated at 107 dB and are nominally 18 ohm, so what would their sensitivity be if the drivers were 8 ohms? Higher or lower or the same?
 

audiobomber

Well-Known Member
Oct 13, 2020
243
227
115
Sudbury ON, Canada
I've not attempted to work it out per the formulae in the video, but my Duo XDs are rated at 107 dB and are nominally 18 ohm, so what would their sensitivity be if the drivers were 8 ohms? Higher or lower or the same?

If they were 8-ohm speakers instead of 18 ohm, they would be much more sensitive. There's no indication of how Avantgarde measured this number, therefore I believe the 107dB figure is simply marketing puffery. This Hi-Fi News test of the Duo Primo XD seems more realistic:

"This is just the second occasion on which we have recorded a genuine 100dB-plus sensitivity from a loudspeaker – and without any recourse to stratagems like low impedance. Just horns, and both from Avantgarde [see HFN Feb 19]! Measured at 2m on the tweeter axis and corrected to 1m, averaged pink noise sensitivity for the two channels was 102.5dB SPL for 2.83V input – not the 107dB (for 1W) claimed by Avantgarde but still an astonishing figure, particularly when the Duo Primo XD's impedance drops to a (high) minimum of 7.5ohm at a high 7.9kHz! Coupled with low impedance phase angles this results in the minimum EPDR being a high 4.9ohm at 3.7kHz (at a very narrow dip), which makes this speaker one of the easiest to drive that we have encountered, as well as the most sensitive."
https://www.hifinews.com/content/avantgarde-duo-primo-xd-loudspeaker-lab-report
 

analogsa

Well-Known Member
Apr 15, 2017
382
122
175
Cascais
One thing is for sure: low impedance speakers do not lead to better sound. Louder, yes, better, no.

This is perhaps less obvious with tube amps, where only the demands towards the output transformer become more vigorous. Provided the transformer is up to the task and can provide the required taps the tubes don't care if the load is 2ohms or 20. But they do care if the impedance varies madly.

Solid state amps are not sensitive towards impedance variations, but consistently exhibit much higher distortion into low impedances.
 

remdeck

Well-Known Member
Apr 10, 2018
71
100
138
At the factory tour in Nov 19, Matthias Ruff told that the 18 ohm is marketing babble. Their drivers have a higher impedance than other speakers, but he orders drivers with different impedance when he wants the speakers to match better. You can use resistors for that, but he prefers using different impedance to match... and this works, as I have never seen a resistor in a Avantgarde speaker.
 

Duke LeJeune

[Industry Expert]/Member Sponsor
Jul 22, 2013
747
1,200
435
Princeton, Texas
This video from John Devore is very much on point:

I agree with John.

The convention I follow is to use the word efficiency when talking about wattage (as in 87 dB/watt/meter), and sensitivity when talking about voltage (as in 87 dB/2.83 volts/meter). Unfortunately there does not seem to be a uniform convention in the industry regarding usage of the terms efficiency and sensitivity. They are often used interchangeably, so you may have to dig a bit to figure out which yardstick they are actually measuring their speakers with. It doesn't matter if the speaker's nominal impedance is 8 ohms, but if it's 4 ohms, they may be using the 2.83-volt sensitivity figure, which will be 3 dB higher than the 1-watt efficiency figure.

I’d be very interested to learn more about the significance of speaker impedance, especially as I’m about to change my old 8 ohm Duos for new 18 ohm ones. Thanks. Peter
I can think of two reasons why a speaker designer might choose to use 18 ohm drivers instead of 8 ohm drivers:

1. My understanding is the MOST amplifiers will sound better into the higher impedance load, assuming they are NOT driven into clipping. I was told this by two tube amplifier designers, one an OTL designer and one a SET designer, and both also told me that this holds true for solid state amps as well. My own experience and the experiences of my customers with solid state amps leads me to the same conclusion. I used to make a speaker which was user-configurable as either an 8-ohm load or a 32-ohm load, so this opinion is based on "apples to apples" comparisons. (The reason I say "MOST" instead of "ALL" amplifiers is because I'm not enough of an amplifier guy to know for sure that there are no exceptions).

2. The higher impedance driver may well have better top end extension and better inner detail, at the expense of some low-end extension and some thermal power handling. This is because the most straightforward way to raise the impedance of the voice coil is to use a thinner wire, and thinner wire = reduced moving mass. I've made frequency response measurements of different impedance versions of the same driver and the frequency response range usually shifts upwards slightly for the higher impedance voice coils. I THINK I've heard improved articulation as well but that could be confirmation bias, and I don't do high power testing so that part is inferred rather than measured.

Anyway with a speaker like the Avantgardes, the reduced voltage sensitivity of the higher impedance load is unlikely to be an issue with most amplifiers, and likely to be an advantage with OTL and SET tube amps, assuming the latter have 16-ohm transformer taps. And any reduction in thermal power handling is likely to be of academic interest only.
 
Last edited:

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
I agree with John.

The convention I follow is to use the word efficiency when talking about wattage (as in 87 dB/watt/meter), and sensitivity when talking about voltage (as in 87 dB/2.83 volts/meter). Unfortunately there does not seem to be a uniform convention in the industry regarding usage of the terms efficiency and sensitivity. They are often used interchangeably, so you may have to dig a bit to figure out which yardstick they are actually measuring their speakers with. It doesn't matter if the speaker's nominal impedance is 8 ohms, but if it's 4 ohms, they may be using the 2.83-volt sensitivity figure, which will be 3 dB higher than the 1-watt efficiency figure.


I can think of two reasons why a speaker designer might choose to use 18 ohm drivers instead of 8 ohm drivers:

1. My understanding is the MOST amplifiers will sound better into the higher impedance load, assuming they are NOT driven into clipping. I was told this by two tube amplifier designers, one an OTL designer and one a SET designer, and both also told me that this holds true for solid state amps as well. My own experience and the experiences of my customers with solid state amps leads me to the same conclusion. I used to make a speaker which was user-configurable as either an 8-ohm load or a 32-ohm load, so this opinion is based on "apples to apples" comparisons. (The reason I say "MOST" instead of "ALL" amplifiers is because I'm not enough of an amplifier guy to know for sure that there are no exceptions).

2. The higher impedance driver may well have better top end extension and better inner detail, at the expense of some low-end extension and some thermal power handling. This is because the most straightforward way to raise the impedance of the voice coil is to use a thinner wire, and thinner wire = reduced moving mass. I've made frequency response measurements of different impedance versions of the same driver and the frequency response range usually shifts upwards slightly for the higher impedance voice coils. I THINK I've heard improved articulation as well but that could be confirmation bias, and I don't do high power testing so that part is inferred rather than measured.

Anyway with a speaker like the Avantgardes, the reduced voltage sensitivity of the higher impedance load is unlikely to be an issue with most amplifiers, and likely to be an advantage with OTL and SET tube amps, assuming the latter have 16-ohm transformer taps. And any reduction in thermal power handling is likely to be of academic interest only.
Thanks Duke

I would offer an alternative, more traditional take on speaker “Efficiency”. In days of old when Wharfedale, Goodmans, etc ruled the civilised world, the use of the expression Efficiency was to provide a measure of just that - the percentage of what comes out as sound compared with what went in as electric power. 10-25% was considered pretty average with some designs, notably from Lowther claiming up to about 60% - partly due to their drivers needing no crossover network.

Probably rightly the use of Efficiency as a percentage has given way to Sensitivity as it's easier to measure accurately. However the expression Efficiency has stuck in people's minds and it's still often used in place of what they really mean - Sensitivity.

I agree that tube amps are more sympathetic to high ohm figure speakers as, in the heyday of tube designs, speakers were generally 15 ohm and amps matched this impedance. The coming of transistors encouraged the switch to 8 or 4 ohm speaker designs and tube amp builders responded by taking more taps off their output transformers - probably to their slight detriment.

I guess Avantgarde took into account the high percentage of owners who preferred to drive their speakers with tubes as a good reason to move to 18 ohms, but I was wondering if there may be another reason. It certainly offers no difficult load to any amp of whatever design, so probably a good move on their part. Peter
 
  • Like
Reactions: Duke LeJeune

Robh3606

Well-Known Member
Aug 24, 2010
1,471
462
1,155
Destiny
This is because the most straightforward way to raise the impedance of the voice coil is to use a thinner wire, and thinner wire = reduced moving mass.

Hello Duke

Yes and you can also change the number of turns and not effect the moving mass. Here is an example of a compression driver that is available in both 8 and 16 ohm versions. If you look at the spec sheet it gives you a very clear picture of how the do it.

Rob :)
 

Attachments

  • 2452SL_EDS_Page_1.jpg
    2452SL_EDS_Page_1.jpg
    153.4 KB · Views: 10
  • 2452SL_EDS_Page_4.jpg
    2452SL_EDS_Page_4.jpg
    56.9 KB · Views: 11
  • Like
Reactions: Duke LeJeune

audiobomber

Well-Known Member
Oct 13, 2020
243
227
115
Sudbury ON, Canada
Provided the transformer is up to the task and can provide the required taps the tubes don't care if the load is 2ohms or 20. But they do care if the impedance varies madly.

Solid state amps are not sensitive towards impedance variations, but consistently exhibit much higher distortion into low impedances.
I agree, except I would not say that SS amps are "not sensitive" to impedance variations. Lower impedance increases current demand, higher impedance decreases power output capability. My power amp is rated to supply 100wpc@8ohms, 200wpc@4ohms. Wattage rating for 16 ohms is not given, but would likely decrease to 50wpc, so it will clip at a lower level when driving a 16ohm load.

Normally when people talk about clipping, they mean voltage clipping, where the amp can't supply sufficient voltage into the load. As impedance increases, the ability of the amp to supply power decreases. Voltage clipping causes gross harmonic distortion and sounds ugly.

Current clipping occurs when the amp can't supply sufficient current to supply the load, or cannot supply current quickly enough to meet sudden demand, and can occur on transients even at low volume. That's why a typical 100wpc receiver sounds gutless compared to a 25wpc Pass amp. Current clipping is far more insidious than voltage clipping, causing a collapsed soundstage, gritty sound and weak lows.

Low impedance is bad in that it is more demanding of the amp, but good because it is more efficient. High impedance designs are less demanding of current, but need more power to achieve the same levels, and running at higher power generally increases distortion. IMO, neither is inherently better. Matching the amp and speaker is the way to achieve an optimal result.
 
Last edited:

Duke LeJeune

[Industry Expert]/Member Sponsor
Jul 22, 2013
747
1,200
435
Princeton, Texas
I would offer an alternative, more traditional take on speaker “Efficiency”. In days of old when Wharfedale, Goodmans, etc ruled the civilised world, the use of the expression Efficiency was to provide a measure of just that - the percentage of what comes out as sound compared with what went in as electric power.
Yup, that's probably the most correct usage of the term, but it's not a straightforward expression of how loud a speaker will go with a given amount of input power (wattage). What is the one watt/one meter SPL of a .22% efficient loudspeaker? The answer is not obvious, at least not to me. It's 86.7 dB.

I would not say that SS amps are "not sensitive" to impedance variations. Lower impedance increases current demand, higher impedance decreases power output capability.

Agreed. The behavior of a voltage-source amplifier has become the default "standard", but its behavior changes significantly based on the impedance it is seeing. Most loudspeakers are designed with the presumption that the amplifier's power will vary in inverse proportion to the impedance curve, or as a "constant voltage" source, but most tube amps (as well as transformer-coupled solid state amps like McIntosh) behave more like a "constant power " source. (Relatively few amps approximate a "constant current" source, but a few do).

Here's an imo informative (and fairly brief) paper on the subject written by an amplifier designer:

 

marslo

VIP/Donor
May 2, 2014
953
673
605
64
Poland
Some measurement numbers about Duo XD impedance from recent stereo-magazine review
 

Attachments

  • 8F2E5A83-6723-4029-92E8-DF65CDFC0FF0.png
    8F2E5A83-6723-4029-92E8-DF65CDFC0FF0.png
    578.9 KB · Views: 20
  • Like
Reactions: MPS

Hear Here

Well-Known Member
Feb 14, 2020
735
435
155
Portsmouth, UK
Some measurement numbers about Duo XD impedance from recent stereo-magazine review
Thanks for that. I've still not attempted to use the XD software although I've loaded the software, connected the long USB cables and taken a look. Unfortunately with no measurements, I don't know where to start. Once I've made a few mods to the speakers (replacing the upright poles with something more elegent, I plan to get the local distributor to visit with his measurement and tuning kit. Let's hope he knows how to use it! Peter
 

MPS

Well-Known Member
Jun 20, 2016
112
84
160
Finland
An interesting explanation. Thanks for sharing. When I started getting into hi-fi, speakers' efficiency was often expressed as a percentage. So if 10 watts were put into a speaker and 9 watts ended up as heat, the sound output was 1 watt, so 10% efficient. Lowther managed up to about 55% whereas Wharfedale (whose Airedale syatem I home-built from Gilbert Briggs's drawaings) were about 25%. Efficiency has given way to sensitivity, rightly or wrongly, and it's interesting to learn that the method of measurement has changed as per the video you posted.

I've not attempted to work it out per the formulae in the video, but my Duo XDs are rated at 107 dB and are nominally 18 ohm, so what would their sensitivity be if the drivers were 8 ohms? Higher or lower or the same?
It seem to me that AA reports 107 dB as SPL rating to 1 W input power. So when DUO XD can be considered as 18 ohm, it would require a signal of 4.24264069 Volts to reach that 1 W power level.
A 8 ohm speaker will require a signal of 2.82842712 Volts to reach 1 W power level.
DUO XD when driven with 2.82842712 Volts only takes 0.444444443 W of power instead.
If you are looking for voltage sensitivity then DUO XD would have SPL which is ~3.5 dB lower than nominal when driven with 2.82842712 Volts.
So a 8 ohm speaker with sensitivity of 103.5 dB/W would play exactly at the same SPL as DUO XD when connected to the same system and same gain setting. Only difference would be that DUO XD would require less power.
That's all in theory. In practice manufacturers present both sensitivity/efficiency and nominal impedance numbers quite "freely" so take all that with a gain of salt...

Edit: small correction to numerical values
 
Last edited:
  • Like
Reactions: Hear Here

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing