Distortion in loudspeakers

[watchnerd]

On the general theme of speaker distortion, I have seen mentioned elsewhere mechanical "hysteresis" and "memory" effects of the surrounds and spider. These have been raised as possible reasons why amplifier damping factor is important, not just for frequency domain reasons, but actual distortion. Does anyone have a view on this?

It doesn't seem theoretically preposterous, but I haven't seen any data about it.

Nelson Pass has made some statements about damping factoring being over-rated once you get above 20 or so.

 

[oivavoi]

It doesn't seem theoretically preposterous, but I haven't seen any data about it.

Nelson Pass has made some statements about damping factoring being over-rated once you get above 20 or so.

I just bought a pair of hypex amps for my next speaker project, which will be externally bi-amped (actively of course). Not the Ncores, but the predecessor, UCD180hg HxR. Pretty sure that's good enough performance when you feed them a benign load. I think the damping factor is about 500!

 

[watchnerd]

I just bought a pair of hypex amps for my next speaker project, which will be externally bi-amped (actively of course). Not the Ncores, but the predecessor, UCD180hg HxR. Pretty sure that's good enough performance when you feed them a benign load. I think the damping factor is about 500!

Yeah, Pass likes to let the driver flop around more, thinks it sounds warmer / fuller.

 

[oivavoi]

Yeah, Pass likes to let the driver flop around more, thinks it sounds warmer / fuller.

I see. It seems to me like mr. Pass and me have different ideals in audio.

 

[watchnerd]

I see. It seems to me like mr. Pass and me have different ideals in audio.

Brilliant designer whom I'm fully convinced can make any amplifier anyone wants, but said in an interview he leans towards making his own-brand amps sound "lush".

 

[hvbias]

Maybe, but not so pure as the beer/kahlua bottle.

So this is why this makes good thinking music

 

[DonH56]

Why even talk about THD when IMD is the real bad guy?

Note HD and IMD are mathematically related, e.g. for the same p-p voltage output the third IMD products are 9.54 dB above the third HD products. More or less and IIRC. The derivation is straight-forward but tedious.

 

[DonH56]

On the general theme of speaker distortion, I have seen mentioned elsewhere mechanical "hysteresis" and "memory" effects of the surrounds and spider. These have been raised as possible reasons why amplifier damping factor is important, not just for frequency domain reasons, but actual distortion. Does anyone have a view on this?

I do not believe it is a significant factor (as tactfully stated as I can make it). There is a "break-in" period but most of it happens in a few minutes to a few hours, and beyond that should be pretty "broken in". Of course anything (including surround and spider) that constrains movement will lead to distortion as its limits are exceeded, but I think amplifier damping and speakers is primarily related to voice coil/magnet interaction that is one of the sources of the driver's electrical impedance, along with the crossover and impact of the cabinet.

That said, and bearing in mind my experience in this area is very limited, a lot of attention has been paid to the material used to make those pieces because they do play a role in controlling cone movement. As for the amplifier, a servo loop can help bring those under control, but lots of speakers sound fine without that.

 

[DonH56]

There have been papers on why really damping factor doesn't really matter. Of course a lot depends upon the speakers the amp is driving...

For a DF of 20 with an 8 ohm load, the driving impedance is 8/20 = 0.4 ohms. Seems a bit high to me, and likely few SS amps are that high, though quite a few tube amps are that high or higher. I seem to remember reading DF = 50 somewhere but I am sure opinions vary widely. Note a 10' run of 14 AWG cable adds about 0.05 ohms, and an amplifier with DF = 100 has a 0.08 ohm output, so the wire almost halves the effective damping factor. If the amplifier were ideal (0-ohm output) that 0.05 ohms of wire equates to a DF = 160. If we used 10' of 16 AWG cable the cable alone adds about 0.08 ohms to the driving impedance seen by the speaker.

A speaker with significant impedance variations across the audio band coupled with low impedances is most likely to benefit from lower amplifier output impedance (and thus higher damping factor).

I strongly suspect output impedance most for the woofers, which are the largest and hardest to control drivers, and that is a good thing because damping factor typically falls off significantly at higher frequencies.

 

[fas42]

I do not believe it is a significant factor (as tactfully stated as I can make it). There is a "break-in" period but most of it happens in a few minutes to a few hours, and beyond that should be pretty "broken in". Of course anything (including surround and spider) that constrains movement will lead to distortion as its limits are exceeded, but I think amplifier damping and speakers is primarily related to voice coil/magnet interaction that is one of the sources of the driver's electrical impedance, along with the crossover and impact of the cabinet.

IME, the distortion appears to be that the surround and spider combination lose flexibility when they haven't been exercised for some, relatively short period. The result is, that low level detail is lost - anything which requires only very minute movements of the cone, which includes listening at low volume, is less well reproduced. My solution is to exercise the suspension hard, at switch on; effectively, the suspension has to be "broken in" every day - this is a big nuisance, obviously, but it appears that there is little interest in improving matters - I see almost nothing mentioned about this factor.

 

[watchnerd]

Note HD and IMD are mathematically related, e.g. for the same p-p voltage output the third IMD products are 9.54 dB above the third HD products. More or less and IIRC. The derivation is straight-forward but tedious.

That's on the electrical side in a linear situation. But as things approach limits, you can have all sorts of electromechanical IMD that is introduced from cone break-up, thermal compression, back EMF, etc, that must be observed as opposed to predicted.

 

[watchnerd]

IME, the distortion appears to be that the surround and spider combination lose flexibility when they haven't been exercised for some, relatively short period. The result is, that low level detail is lost - anything which requires only very minute movements of the cone, which includes listening at low volume, is less well reproduced. My solution is to exercise the suspension hard, at switch on; effectively, the suspension has to be "broken in" every day - this is a big nuisance, obviously, but it appears that there is little interest in improving matters - I see almost nothing mentioned about this factor.

Errr....

You're saying, in room temperature (not Arctic cold), that the suspension changes compliance enough to be audibly different within 24 hours of last use?

 

[fas42]

Errr....

You're saying, in room temperature (not Arctic cold), that the suspension changes compliance enough to be audibly different within 24 hours of last use?

Precisely ...

Subjectively, the sound becomes more "interesting", there is more apparent detail, you can hear "deeper into the mix" - pretty easy to test, get two pairs of the same model of speakers, check that they match each other, especially audibly, using all the methods that you care to apply - put one pair aside for 24 hours, and keep the other running at relatively high volume for that period; then, hook up the "cold" set, and compare immediately, before the virgin pair get into their stride.

 

[DonH56]

That's on the electrical side in a linear situation. But as things approach limits, you can have all sorts of electromechanical IMD that is introduced from cone break-up, thermal compression, back EMF, etc, that must be observed as opposed to predicted.

I think you meant to say "nonlinear" situation... I think I mentioned those sources earlier but maybe not. Cone modes and breakup require dynamic mechanical modeling but thermal and EMF issues are well-understood and can be modeled. I have seen simulations of cone breakup so I know they are out there but it is far afield from my area of expertise so, other than marveling at the pretty pictures and seeing the same thing in a cone using high-speed imaging, I couldn't say exactly what is in their models.

 

[oivavoi]

Precisely ...

Subjectively, the sound becomes more "interesting", there is more apparent detail, you can hear "deeper into the mix" - pretty easy to test, get two pairs of the same model of speakers, check that they match each other, especially audibly, using all the methods that you care to apply - put one pair aside for 24 hours, and keep the other running at relatively high volume for that period; then, hook up the "cold" set, and compare immediately, before the virgin pair get into their stride.

Temperature and humidity can indeed change the response of loudspeakers in a room. But I don't think it's for the reason you think. Changes in temperature+humiditiy change how sound waves move through the air, and will likely change the frequency response somewhat if the changes are big enough (paper cone woofers can also be affected by changes in humiditiy). This can be measured, and probably has a larger effect on how things sound than changes in electronics, cables, etc. But I've yet to see audiophiles who implement intricate systems for keeping temperature and humidity at at optimal level at all times...

(for us pianists, btw, this is a no-brainer if one is lucky enough to have a grand piano in the house: them pianos like the climate to be stable. it's like the retirees who move to florida. so an investment in a grand piano also necessitates an investment in humidity control)

 

[fas42]

Temperature and humidity can indeed change the response of loudspeakers in a room. But I don't think it's for the reason you think. Changes in temperature+humiditiy change how sound waves move through the air, and will likely change the frequency response somewhat if the changes are big enough (paper cone woofers can also be affected by changes in humiditiy). This can be measured, and probably has a larger effect on how things sound than changes in electronics, cables, etc. But I've yet to see audiophiles who implement intricate systems for keeping temperature and humidity at at optimal level at all times...l)

Note, I'm not talking about temperature and humidity in the room - I'm talking about the mechanical condition, behaviour of those parts of the drivers which physically support the moving cones. These are almost never talked about, which is like completely ignoring the suspension engineering of a high performance vehicle.

Interesting side note: starting up the engine of a F1 machine is a very involved process - key parts of the engine have to be at a precise temperature before hitting the start button - otherwise, bye, bye, engine!!

 

[oivavoi]

Note, I'm not talking about temperature and humidity in the room - I'm talking about the mechanical condition, behaviour of those parts of the drivers which physically support the moving cones. These are almost never talked about, which is like completely ignoring the suspension engineering of a high performance vehicle.

Interesting side note: starting up the engine of a F1 machine is a very involved process - key parts of the engine have to be at a precise temperature before hitting the start button - otherwise, bye, bye, engine!!

Ok, I mistunderstood you. It's getting late here in Norway

 

[Don Hills]

Precisely ...

Subjectively, the sound becomes more "interesting", there is more apparent detail, you can hear "deeper into the mix" - pretty easy to test, get two pairs of the same model of speakers, check that they match each other, especially audibly, using all the methods that you care to apply - put one pair aside for 24 hours, and keep the other running at relatively high volume for that period; then, hook up the "cold" set, and compare immediately, before the virgin pair get into their stride.

You keep doing things the hard way, because you insist on measuring with your most unreliable test instruments. Any change in the compliance of the suspension is easily measurable with far more accurate instruments. Its effect on the reproduced waveform is also easily measurable.

 

[fas42]

You keep doing things the hard way, because you insist on measuring with your most unreliable test instruments. Any change in the compliance of the suspension is easily measurable with far more accurate instruments. Its effect on the reproduced waveform is also easily measurable.

Yes, those things would be measurable - but I don't see any data around the place which confirms, either way, what I'm saying. I would be much happier to have audio components which are static in their behaviour, with respect to time and previous use - but I don't find that to be the case. So, there is a "problem" - getting measurements is just confirmation of something that should be addressed.

 

[Blumlein 88]

Yes, those things would be measurable - but I don't see any data around the place which confirms, either way, what I'm saying. I would be much happier to have audio components which are static in their behaviour, with respect to time and previous use - but I don't find that to be the case. So, there is a "problem" - getting measurements is just confirmation of something that should be addressed.

So instead of your approach which most of us here find lacking, and since such measurements are not very hard to do why persist in using your most unreliable instrument. Why not do some measures and show us what you come up with to support what it is you say you are hearing/experiencing?