Low Sensitivity Speakers Compression Issue.


Main stream Hi-end producers, audio reviewers and sellers push low sensitivity tower speakers and high power amplifier solution to audiophile society as "reference" low distortion solution.

My experience shows me that this solution doesn't work.
I had myself Dynaudio Audience 60 speakers. These speakers was my worth purchase in audio.
These speakers sound horribly compressed. They sounded completely different on different volumes. 

I have listened many different low sensitivity tower speakers and high power amplifier in audio shows, dealers and people houses. This kind speakers ALWAYS sound compressed even with 1000w amplifiers.

I heard about TD in speakers. This distortion are not linear, because they have a time inertia.
Here is I found a very good article written by Derek Alexander-Wilson that enplanes this issue:   
https://www.linkedin.com/pulse/power-compression-vs-thermal-distortion-loudspeaker-alexander-wilson/   


"Power compression Vs Thermal distortion

Power compression is the common term for one of the audible effects of voice coil heating / overheating, i.e. when the voice coil reaches its maximum power handling limit trying to ram more power in is sonic train crash…It does not get any louder just compressed and grossly distorted.

This effect is better described as “Thermal Distortion” (TD) and is much more serious than just a maximum power handling limitation or side effect…..TD is a huge problem, overlooked or deliberately ignored by most manufacturers as there is no easy (low cost) solution and TD is audible and measurable most of the time at most power levels…!

TD is caused by the conductive metal (aluminium, copper or silver) voice coil getting hotter when you pass electrical energy through it….The more power you pass through it the hotter the metal gets….The hotter the metal gets the more the electrical resistance increases….The efficiency goes down and you need to ram in more and more power for smaller and smaller increases in SPL….a very vicious circle.

The vast majority of loudspeaker drivers are severely “under voice coiled” and nearly all suffer from massive audible and measurable TD even at medium power handling levels, long before maximum power where audible “power compression” kicks in.

Taking a look at the relationship between voice coil power handling and voice coil heating reveals that long before audible power compression becomes an issue, TD is a serious problem….So widespread that the entire audio industry accepts it as standard and never questions it.

TD creeps into sound reproduction just like harmonic or inter-modulation distortion from low power levels and gets worse the higher the SPL’s….Far more serious than the simple “loss of sensitivity” at max power as loudspeaker driver describe it.

TD causes severe program dependent tonal imbalances in multi-way active speakers and the effect is even worse in passive loudspeakers.

Room temp to 120 degrees C

Taking an example of a voice coil using 10 meters of 0.2mm Copper with a typical resistance of 5.5 Ohm at 20 degrees C.

Apply power (music signal) & the voice coil jumps to 120°C and the resistance rises to approx. 7.5 Ohm….Now the resulting current flow through the voice coil drops by approx. 35% ….So does the SPL….This is a 35% distortion!

Orders of magnitude worse than the usual suspects of THD or IMD which loudspeaker manufacturers like to quote.

With passive crossovers in multi way systems thermal power compression will shift XO frequencies when the voice coil resistance changes – resulting in summation errors between the two drivers involved, depending on program and listening level.

The same holds true for notching out resonances or compensating impedance….The more complex the passive crossover the worse the TD effect becomes.

Why 120 degrees C?

This is a safe or medium range voice coil temperature, many voice coils hit double or triple this temperature on a regular basis.

The energy involved is as follows:

The weight of our VC wire is slightly below 3g if we calculate length multiplied by cross sectional area multiplied by specific mass of copper.

We require approx. 1 Watt for 1 sec for each Kelvin temperature increase (no cooling assumed here), looking at the thermal material constants for copper

To heat up this < 3g of copper wire our voice coil is made from, we only need 10 sec of 10W input to finally arrive at a 100°C increase – or – if we have fortissimo playing – a very short 1 sec of 100W input.

Well within the realms of average….

Now if you want to enjoy your fortissimo at 115dB SPL and given your speaker would be 95dB / W / m - well - after one (!) single second you have reached 35% of distortion – given your speaker to be 105dB sensitive you can enjoy your fortissimo of 115db SPL a fantastic long 10 (!) seconds until you end up at reaching 35% distortion.

Obviously the higher the loudspeaker sensitivity the lower the TD effects, and the better the voice coil cooling the lower the TD effects. Most “raw” audiophile loudspeaker drivers struggle to exceed 89dB over their working range, Pro audio drivers average out around 94 dB to 95dB SPL for 1 watt at 1 meter over their working bandwidth.

Real world voice coil cooling inside loudspeaker cabinets is very limited and varies from almost zero in a small sealed box with small magnet / motor to reasonable in large vented cabinets with huge heavy ferrite magnets.

But still cooling down from a 100Ws input will take much longer than the near instantaneous heating up….

Using a single layer 50mm voice coil with a height of 13mm with no isolating former (Kapton or Nomex...) is the most optimistic scenario.

This will provide a heat dissipation area of around 40 square cm. A two layer voice coils will half the effective radiating area & a four layer voice coil reduces effective radiating area by a factor of 4!

Again, being optimistic & assuming the surrounding magnet structure remains at 20°C while the voice coil already has reached 120°C the heat dissipation will be approx. 3W.

This means the decay from our 35% distortion fortissimo would last around 30 sec. Unfortunately this does not hold true - because – if the voice coil is cooled down to 70°C (half way down) then there is only about 1W cooling through radiation left (again optimistically assuming the surrounding magnet structure still to be at 20°C), further stretching the time of distortion decay.

Using an aluminium former would help to enlarge the radiating area of the coil itself but the need to minimise moving mass forces designers to use ultra-thin aluminium which cannot support massive heat transport

Forced air / convection cooling will help but pianissimo follows fortissimo - you only have forced cooling due to voice coil and membrane movement during fortissimo…

The bottom line is TD is a massive problem….The elephant in the room.

Currently the best ways to minimise the effects are:

(1)  Use an active crossover.

(2)  Select loudspeaker drivers with the largest voice coils…Bigger is better.

(3)  Single layer voice coils are best, two layers bad, 4 layers worst.

(4)  Aluminium voice coil formers may help, but they have other sonic downsides compared to Kapton or Nomex."

 


alexberger

Showing 11 responses by alexberger

Thank you. @roxy54 ,

I'm pleased to know that other people knew about this problem.

The actual article was written by Derek Alexander-Wilson.
I like high sensitivity speakers. IMHO they have a lot of advantages:
1. They sound alive, musical.
2. The soundstage has a real size.
3. You can listen them on high volume without fatigue.
4. You can use highest quality most natural and transparent amplifiers
like tube SET amplifiers (300B, 2a3, 45, ad1,...) or First Watt SIT transistor amplifiers.

The drawback of high sensitivity speakers is a very deep bass.
I know that people who want to have dynamic sounded system with a deep bass use two solutions:
1. Array with a big number of bass drivers. A big power is distributed between a number of drivers that decreases TD.
2. Sub bass horns.

Both these solution are too expensive.
So I chose for myself a real dynamic over deep bass. 

Regards,
Alex.
Hi @viridian ,

I heard first about TD in speakers and transistor amplifiers from Russian audio designer and guru - Anatoly Markovich Lichnitsky. This person liked to listen mono records (mostly classical from 78RPM era) and didn’t care about all audiophile stuff like soundstage, imagined, bass slam,...
Many years ago my friends and me couple time a week went to listen systems of different audiophiles in our region. We listened different box, electrostatic and planar speakers.
And ones we went to one person who had big horn system with Lowther PM4 speakers and 15 inch EV mid bass speakers. It was a culture shock for us. We went back home and can’t listened our home systems during a long period of time. Since this visit, low sensitivity speakers did not interest me.

Regards,
Alex.
Hi @atmasphere ,

All transistor amplifiers (maybe, except pure class A) have thermal distortion.
Worst at all - thermal distortion have inertia and they are nonlinear.
Our brain is very sensitive to this kind of distortions because they are not natural.

Regards,
Alex. 
I don't say all high efficiency speakers are good and all low sensitivity speakers are bad.
For example, I like Harbeth and Spendor Classic speakers.
What is interesting they sound much more dynamic compared to similar efficiency speakers (probably because these companies don't lie in their product specifications).
But I talk about dynamic. 
I'm getting mad when some audio sellers, reviewers and producers talk that 85dB speakers with 1000 Watt amplifier can give a real, uncompressed dynamics. It is simply not true.
The same according tone. 
Yes, there are a lot of natural sounding low efficiency speakers and you can find many examples of high efficiency speakers with not natural sound.
But I can't say that for sensitivity you have to sacrifice a natural ton. 
In most cases, paper cone give more natural tone. And most paper cone drivers are more sensitive than kevlar and plastic drivers.

Here are my fresh memories from Montreal audio show this year.
I listened a number of very high efficiency speakers systems there.
There are different solutions for different price and with different kind of sound:
Tune Audio Marvel - https://www.youtube.com/watch?v=JCLuH_NeVgI
Avantgarde Uno - https://www.youtube.com/watch?v=4s8AeLSdk2M
Avantgarde Duo - https://www.youtube.com/watch?v=-cjmSP574Y4
Tri-Audio - https://www.youtube.com/watch?v=oed2nHBWVJY
I liked Tune Audio Marvel system a lot. It was very good in both tone and dynamic. IMHO the best system in Montreal Audio Fest this year.
Avantgarde systems has a huge sound, with very good dynamic and a lot of details. But I don’t like tone and the quality of bass.
Tri-Audio was a nice system and very good value for money. It sounded good but their mid-range drivers have some coloration, typical for full range drivers with a visor.
Hi @georgehifi ,
Yes for me this has always been the downfall of high sensitivity speakers, they are usually too coloured for me.
If you can live with that fine, but I’ve always chased the opposite, I’d sacrifice a little sensitivity for neutrality and transparency.

I gave here a good sounded high sensitivity speaker example -  Tune Audio Marvel.
Most of Tannoy speakers are natural in term of tone.
Audio Note  speakers are very good in term of tone.
I also like my vintage Altec 604E. They are not as extended in high frequencies and sub bass as best modern speakers. But I think they overperform most of modern low-mid sensitivity speakers in term of tone. They are also transparent and have good soundstage. They are great for vocals and any kind of jazz music. 

I also like Harbeth and Spendor (mid-sensitive speakers) in term of tone and musicality.

I listen mostly jazz and classical music. I listen a little classical rock. I don't listen any type of pop and electronic music.

But all people are different, have different test and like to listen different kind of music.

Regards,
Alex

I don't understand why do you contrapose high sensitivity and natural tone of speakers?
There is not trade off between high sensitivity and natural tone.  
But there is a very clear correlation between low sensitivity and compression.
Hi @glupson,

All those links worked for me. I do not know if there are some "requirements" in design of such speakers. Of course, talking strictly about sound, there may be a lot of discussion. However, it is also understandable if majority of people do not give them much attention and decide not to buy them. All of the speakers you linked to are large and imposing. Not exactly living room friendly for most. That may play against them on the market. I know, that is a whole another topic, but just noticing.

I don’t care what use most of people. The big part of audiophiles today listen junk pop and electronic music that we can hear at most audio show and YouTube videos.

I used my Altec 604 in 200 sq feet room in the corners.
Bass wasn’t boomy at all, because of very strong bass drivers magnets and as result very low Qts.
I think speakers like Tune Audio Marvel can be used in very small room.
In 50-60x big speakers weren’t issue.

I think small speakers become the fashion because transistor amplifiers, WAF, cost reduction and marketing but not for sound quality.
Why do you need to produce expansive speakers with big boxes, big paper cones with strong alnico magnets, if you can produce cheaper small drivers with plastic cones and weak magnets and your marketing people will brainwash folks that you do it not because cost reduction but for sound quality?

By the way, that video with Tune Audio Marvel is puzzling. Are those record covers on the floor? Are there records in them? Records on the floor? One of the first things I have ever read about audio reproduction. No records on the floor. Dust. Horizontal only on the turntable.

It was their show room in Montreal Audio Fest.
Hi @georgehifi ,

I think horn coloration issue was exaggerate by modern audiophiles.
Why most recording studious have been using horns for monitoring, from old Tannoy, Altec and JBL to modern Ocean Ways, Tad and Kinoshita?
IMHO beryllium tweeters sound more unnatural that good designed horns. 
When I'm listening speakers with beryllium tweeters I have fealling the someone is crumpling sheet of paper.
If you don't like horns, there are very good in term of tone, high sensitive Audio Note speakers.

Regards,
Alex.
Hi @mijostyn ,

Horns can be made with very little coloration and certainly get the nod in terms of efficiency. If you are in love with low powered amps there is no better solution. However, I don't like them. I do not like them because they are point source radiators and do not broadcast a realistic sound stage. I think that most well designed drivers run within their established parameters are not going to compress anything. 
I think you have an experience with bad designed  low powered amps.
1. Many of them use too small transformers (that have bass issues) for cost reduction. 
2. A big pat of 300B SETs use very week drivers like 6SN7 (to high output impedance and low anode current).
Because these design mistakes these amplifiers sound "small".

I have 300B amplifier with 6F6 driver in triode. I have a big and 3D soundstage with my Altec604.

Regards,
Alex.