Output Compression due to Heat
Loudspeaker conversion efficiency to start with is very low and band-width limited. What ever electrical energy is not translated into acoustical energy goes to heat. Most of the heat is dissipated by the voice coil. When the heat generated is more than that dissipated, the ambient temperature of the voice coil increases. In the first stage, voice coil resistance is increased, causing a drop in driver output (now less efficient and results in 'compression' of the acoustic output). If allowed to continue you get thermal runaway and a smoked gig. This has nothing to do with the notion of a compression driver or the function of peak lopping signal compression. WHG
Loudspeaker conversion efficiency to start with is very low and band-width limited. What ever electrical energy is not translated into acoustical energy goes to heat. Most of the heat is dissipated by the voice coil. When the heat generated is more than that dissipated, the ambient temperature of the voice coil increases. In the first stage, voice coil resistance is increased, causing a drop in driver output (now less efficient and results in 'compression' of the acoustic output). If allowed to continue you get thermal runaway and a smoked gig. This has nothing to do with the notion of a compression driver or the function of peak lopping signal compression. WHG
Last edited:
I'd say you're right.
Lets say you play just a kick drum signal through a speaker. As you push the fader, the sound will get louder until the driver starts to exceed Xmax and the suspension/motor starts going nonlinear. When that starts happening, the sound might still get louder, but it won't follow the fader any more. If you push for +3dB, it might only give +2dB output.
Kick drums are low duty cycle signals, so the thermal effects will be pretty minimal.
Mechanical nonlinearity is instantaneous, since it's acting on the waveform peaks. Thermal effects have time constants of the order of minutes, so you'll find that if you hammer a system for a while, its voltage sensitivity has gone down.
I found thermal effects to be a problem with some PA speakers I had. I was using some 500w 12" drivers that had a 3" voicecoil and basic motor venting (just a polepiece vent). By the end of a rock 'n' roll gig, they were around 6dB down on the rest of the system. Checking inside the cabinet, the magnet was hot! I'm glad I checked on them, though - I had been tweaking levels and EQs in response to the 100Hz-1kHz range slowly dropping. If it'd carried on like that, I would've turned things up even further, drivers would've got hotter and quieter, round and round until the drivers burn out completely.
These days I use some drivers that also have 3" voicecoils, but the extensive motor venting means I've never got them hot, so the system sounds the same at the end of the night.
Chris
According to an older JAES paper written by someone from JBL power compression can happen with quite low time-constants. They showed their driver with 4" voice coil to perform better in this respect than their competitor's with a 2.5" variant, with both drivers speced for the same rated power.
Voice coil diameter is not the only construction detail determining the power compression behaviour of course.
I don't know any manufacturer showing the time constants involved. But at least in the professional area a lot of them are at least showing the power compression figures for some values of constant input power (quite common values are for -10dB, -3dB and full rated power during 5 minutes).
Beyma for instance publishes graphs for some of their driver's power compression behaviour.
Because the power compression behaviour can differ quite a bit for each driver in multiway Systems which can be leading to power dependant frequency response, some manufacturers show frequency response graphs at different input power. Levels. This famous example here is showing excellent behaviour throughout its whole working range:
http://www.jblpro.com/pub/obsolete/443035.pdf
Because driver nonlinearities do not result in very high order products (which would be very objectionable) which themselves are at the same time mechanically low pass filtered by the driver, they might often sound like limited dynamics rather than distortion. Not unlike some soft-clip limiters.
Regards
Charles
Voice coil diameter is not the only construction detail determining the power compression behaviour of course.
I don't know any manufacturer showing the time constants involved. But at least in the professional area a lot of them are at least showing the power compression figures for some values of constant input power (quite common values are for -10dB, -3dB and full rated power during 5 minutes).
Beyma for instance publishes graphs for some of their driver's power compression behaviour.
Because the power compression behaviour can differ quite a bit for each driver in multiway Systems which can be leading to power dependant frequency response, some manufacturers show frequency response graphs at different input power. Levels. This famous example here is showing excellent behaviour throughout its whole working range:
http://www.jblpro.com/pub/obsolete/443035.pdf
Because driver nonlinearities do not result in very high order products (which would be very objectionable) which themselves are at the same time mechanically low pass filtered by the driver, they might often sound like limited dynamics rather than distortion. Not unlike some soft-clip limiters.
Regards
Charles
Last edited:
Two
Its heat generation vs heat dissipation. i.e. resistance vs conduction through V.C. former (particularly metal), diaphragm (particularly metal) and surrounding air; convection (air flow over coil and heat sink surfaces), and radiation form coil and heat sink surfaces. WHG
One layer of rectangular section wire, edge-wound on each side of the V.C. former.
Its heat generation vs heat dissipation. i.e. resistance vs conduction through V.C. former (particularly metal), diaphragm (particularly metal) and surrounding air; convection (air flow over coil and heat sink surfaces), and radiation form coil and heat sink surfaces. WHG
Last edited:
Yes ...
... in theory true, but in practice not yet. No mater how you make the driver move, energy in must equal acoustical energy out plus heat generated. If you apply cryogenics to the voice coil the problem disappears, but that also is not likely at the moment.
Note also, that heat is also generated by current flow induced by Back EMF.
WHG
... in theory true, but in practice not yet. No mater how you make the driver move, energy in must equal acoustical energy out plus heat generated. If you apply cryogenics to the voice coil the problem disappears, but that also is not likely at the moment.
Note also, that heat is also generated by current flow induced by Back EMF.
WHG
Look, measured thermal compression time constants and losses for some drivers
https://www.jblpro.com/ProductAttachments/tn_v1n18.pdf
https://www.jblpro.com/ProductAttachments/tn_v1n18.pdf
Thanks! An excerpt:
3dB thermal compression I understand, this occurs when voice coil becomes twice as hot as room temperature, say 600K. (A paper voice coil former or simple glue cannot take this temperature -- it takes sophistication to reach this amount of compression.) But temporal demagnetisation, what is this? I only read of remaining demagnetization of Alnico and neodym.
Permanent magnets reduce there magnatism if they get hot enough. I think that's what they mean.
https://en.m.wikipedia.org/wiki/Curie_temperature
https://en.m.wikipedia.org/wiki/Curie_temperature
Btw, I never understood why, despite its ratio force/weight, neodynium is used : seems very fragile, the Curie point is very low, under 100°C if I remember, more easy to have permanent magnetisation loss and not the one which can be recovered when the magnett becomes cold again ! (well the only reason as designers has pointed out already at DIYAUDIO : less back print area is better for the bac-wave of the cone !)
Last edited:
Just checked, Curie temperatures are very real but way above what any speaker magnet can reach in normal (even heavy duty) use.
Values in Kelvin degrees, substract 273 for Centigrade.
Those still thinking Fahrenheit: it´s the 21st Century guys
Values in Kelvin degrees, substract 273 for Centigrade.
Those still thinking Fahrenheit: it´s the 21st Century guys
So weakest would be Neo at 310C ... still above speaker disintegration temperature.
Thanks to have checked ! Lol, yes use International convention ! Just the Kilogramm is still debatable !
As for the Alnico, often the prmanent loss are more due to the weakest way they where manufactured before ceramics mix swaped it. That's why there are so crapy old Altec, second hand on the markett, their manufacterer about the Alnico was not good enough ! (With a proper Alnico magnett : good luck to demagnetiz it with a hifi home use !) (and a re magnetisation re not arrange the melted mix during the fabrication process)
But sorry it's off topic
! Lol, yes use International convention ! Just the Kilogramm is still debatable !As for the Alnico, often the prmanent loss are more due to the weakest way they where manufactured before ceramics mix swaped it. That's why there are so crapy old Altec, second hand on the markett, their manufacterer about the Alnico was not good enough ! (With a proper Alnico magnett : good luck to demagnetiz it with a hifi home use !) (and a re magnetisation re not arrange the melted mix during the fabrication process)
But sorry it's off topic
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.