Psst... Ready to join TalkBass and start posting, make new friends, sell your gear, and more?  Register your free account in 30 seconds.

Power Compression in Loudspeakers

Discussion in 'Amps and Cabs [BG]' started by ESP-LTD, Aug 24, 2005.


  1. ESP-LTD

    ESP-LTD

    Sep 9, 2001
    Idaho
    I have read a lot of threads recently in which folks are using the electrical power rating of speakers as a guideline for matching cabs to heads. To my way of thinking, not only do you need headroom on your amplifier, but also on your speakers to avoid losses due to power compression-

    (see this Loudspeaker FAQ for description).

    To quote their excellent FAQ:


    Power Compression The Watts that are not turned into sound get transformed into heat. This heat increases the electrical resistance of the speaker, making it harder for the amplifier to drive. The harder you drive the speaker, the more heat is generated.

    To return to the motoring analogy: This is the "aerodynamics of the speaker": The faster you drive, the more wind resistance there is. Thus, even more power is required to make up for this. The lower the power compression figure is, the better.


    Unfortunately, I have not seen any drivers quote a spec for power compression (I'm not really sure how that is measured). A friend who seems knowledgeable about such things once told me that for a specific driver rated at 600w RMS, there was no real change in volume in driving it at 300w vs 600w due to changes in speaker specs caused by power compression.

    Can anyone shed light on this topic? This sort of thing wasn't much of a problem 'in the old days' when we had 100w amps and 200w speakers, but now drivers can withstand a kilowatt and folks are making 'watts' their primary shopping consideration.
     
  2. Selenium and JBL publish power compression measurements in the data sheets for their drivers.

    The trick to avoiding power compression is wicking away the heat from the voice coil. Wayne Parham is doing some leading edge work with heat sinks/wicks for this purpose. He has demonstrated significant heat reduction (by half).

    Sealed chambers are the most difficult to cool, i.e. sealed boxes and the rear chambers of bass horns. One novel way of dealing with this in bass horns such as the Tuba, is reversing the driver and hanging its ass in the breeze, instead of inside the compression chamber. The popular LAB horns have a pair of 12" drivers heating up a small compression chamber. Heat is an issue, and removal of that heat is critical to power compression.

    "A cool tool, is a happy tool".

    :D
     
  3. ESP-LTD

    ESP-LTD

    Sep 9, 2001
    Idaho
    Quoting from specs at Parts Express for a very nice 15" driver rated at 500w RMS:


    Selenium WPU 1507

    Power Compression @ 0 db (nom. power) .......... 3.9db
    Power Compression @ -3 db (nom. power)/2 ...... 2.8db
    Power Compression @ -10 db (nom. power)/10 ... 1.0db


    If I am reading this right, then by doubling the power from 250w to 500w you add 3db of acoustic power, but power compression consumes 2.8 db of that?
     
  4. Bob Lee (QSC)

    Bob Lee (QSC) In case you missed it, I work for QSC Audio! Gold Supporting Member Commercial User

    Jul 3, 2001
    Costa Mesa, Calif.
    Technical Communications Developer, QSC Audio
    Power compression is related to voice coil temperature, as you mentioned. Dynamic loudspeakers are only a few percent efficient, meaning that nearly all the power going into them turnd into heat instead of acoustic energy. The voice coil has mass and specific heat, and it has cooling by convection (into the air in the gap) and radiation (into the magnet structure), so there is a lag between putting more power into the loudspeaker and the voice coil heating up, and there likewise is a lag between decreasing the power and the temperature dropping.

    The resistance of a metal increases with temperature, which is why the voice coil's resistance increases when it is hot. And for a given signal voltage from the amp, the higher resistance raises the loudspeaker's impedance, so it draws less current. Less current at the same voltage means less power and reduced acoustic output.

    I don't pretend to be a loudspeaker engineer, but I know that one of the big challenges facing designers is cooling that voice coil. A few years ago Community, a manufacturer in Pennsylvania, introduced their AirForce line of loudspeaker systems that used forced air blown through ducts into the back of the driver's magnet to circulate air through the gap and improve cooling. The better the cooling, the lesser the effect of power compression will be, and the lesser the risk of burning out the driver will be.

    To your example of a loudspeaker driven with 300 watts and with 600: there will be a 3 dB difference in loudness, at least for a short while. If the higher power level is sustained, the voice coil will heat up and gradually introduce power compression. Often in PA rigs, an inexperienced operator will experience both power compression in the loudspeakers and temporary threshold shift in his or her own auditory system and conclude that the system needs to be turned up, a procedure response that will tend to only make things worse.

    You can avoid the phenomenon of power compression if you have enough loudspeaker(s) and amp power to operate your system with headroom.
     
  5. ESP-LTD

    ESP-LTD

    Sep 9, 2001
    Idaho
    It seems then, that any given system reaches an equilibrium at which additional power doesn't really add much additional acoustic power. In the case of the Selenium driver I quoted specs from (chosen pretty much at random) that point seems pretty close to 50% of maximum RMS power.

    Does degradation of the magnet field strength due to temperature impact this as well? Or is that pretty insignificant and 'self-healing' at the temperatures we are talking about?

    Personally, I like to derate my speakers by about 50% and my amp as well. I think it sounds better, and I KNOW it will sound that way for a lot longer.
     
  6. Bob Lee (QSC)

    Bob Lee (QSC) In case you missed it, I work for QSC Audio! Gold Supporting Member Commercial User

    Jul 3, 2001
    Costa Mesa, Calif.
    Technical Communications Developer, QSC Audio
    If your average power is reasonably lower than the loudspeaker's power rating and you use the power peaks for clean reproduction of transients, you're going to get good dynamic range even if you occasionally push the loudspeaker to near its limits. But if you really heat up the voice coil with high average power, pushing more power into it won't produce comparable acoustical increases.

    Yes, magnets lose strength at higher temperatures, which makes loudspeakers less sensitive. Magnets do have the advantage of having contact with a substantial metal structure that can carry away heat.
     
  7. billfitzmaurice

    billfitzmaurice Commercial User

    Sep 15, 2004
    New Hampshire
    Owner, Bill Fitzmaurice Loudspeaker Design
    Power compression also has a mechanical aspect. In vented boxes acoustic pressure in the vent can build to the extent that the output level relative to the cone output drops. In horns mechanical compression can result when the motor strength is no longer adequate to overcome the acoustic impedance of the throat. At best a linear relationship between power input and acoustic output can't be counted on past 50% of the drivers rated power. I've seen drivers operate relatively linear up to only 1/3 power and literally crap out above that. This bears directly on other threads where players are concerned about getting every possible watt out of their amps when in most cases it won't make things a bit louder anyway.
     
  8. geshel

    geshel

    Oct 2, 2001
    Seattle
    Not quite. What this means is:

    - operating the driver at max power continuously will cause a 3.9dB drop in output eventually

    - operating the driver at half power continuously will cause a 2.8dB drop in output eventually

    - operating the driver at 10% power continuously will cause a 1.0dB drop in output eventually

    So: If the driver is rated at 90dB 1w/1m, with a max continuous power rating of 200w, and you run it at 200w, it'll start out at 113dB but then gradually drop to 109.1dB.

    (the 113dB is the baseline of 90dB at 1w plus 10dB (going from 1w to 10w) + 10dB (going from 10w to 100w) + 3dB (going from 100w to 200w)).
     
  9. ESP-LTD

    ESP-LTD

    Sep 9, 2001
    Idaho
    A thanx to all for your contributions. It certainly seems like power compression has a pretty significant impact on total volume. I hope folks will consider that when trying to cram as much power as possible into a single box.
     
  10. Monomer

    Monomer

    Jul 22, 2005
    I was talking with Wayne (when I was order my speakers) and he talked me into making of of these "Heatsinks"

    Well the speaker arrived (Mag12HO) and I dont have (ahem) "balls" enough to go at them with a scraper (taking the vent cover) and sandpaper (to get rid of the paint) The speakers are just to nice, and pricey.

    Although I do have a delta12pro (half as cheap) that could use one....


    -I'll keep an update on it, either here or on wayne's/bills forum.
     
  11. fdeck

    fdeck Supporting Member Commercial User

    Mar 20, 2004
    Madison WI
    HPF Technology LLC
    I think Bob Lee's comment is well taken that you are OK if you are using the speaker's dynamic range to track occasional transients -- possibly even once per note -- rather than full time. I am now playing around with ways to measure the power content of a "typical bass signal," i.e., my bass when I play it. What I am finding is that the signal spends a significant fraction of its time at amplitudes far below the peak that occurs at the front end of each note.

    Right now I am applying this information to power amp design -- the aforementioned thread -- rather than to my speaker. But I can see how it is equally applicable in both cases.

    I have a working analysis, I think. My plan is to bring home a MiniDisc full of bass signal from my next gig, and start crunching numbers.
     
  12. Monomer, if your Magnum is for a Tuba, reverse the driver and hang it's ass-end out in the breeze instead of facing the compression chamber. Performance is the same either way, but cooling is much improved.
     
  13. That is why the optimum vent area should equal Sd. Radiation at the tuning frequency is almost entirely from the port and cone movement is nil at that frequency.

    Using a larger diameter vent means using a longer duct. This introduces "organ pipe" resonances into the system where the duct length is an integral of the higher wavelength. Always a compromise.

    The solution is running lots of drivers at lesser power. A whole herd of 'em will do the trick. Everybody rows a little, instead of just two rowing real hard.
     
  14. The problem isn't cramming too much power into the box, the problem is trying to get too much sound out of the box.

    What you say is true, but power compression only approaches eventually the effect on volume that simply having less power does RIGHT NOW. And robs you of power that could be used for transients that would not be long enough to cause power compression.

    Using less power to avoid the effects of power compression is like getting the flu to avoid the discomfort of a flu shot.

    If your 300 W speaker driving by a 600W amp (avg power still >300W to avoid melting the voice coil, extra power for headroom) is experiencing power compression, the solution to eliminate power compression is not to reduce the amp to a 300W amp. The problem is you're trying to get 10 pounds of sound out of a 5 pound bag of speakers.

    You need another 300w speaker with another 600w amp so that you aren't running the speaker so close to its max SPL.

    Your amp needs headroom (extra watts) to adequately drive the speaker, and the speakers need "headroom" (extra cone area/excursion available) to adequately fill the room.

    Adequate refers to "with excess capacity to deal with short transients cleanly" in both cases.

    Having more watts than the speaker needs and more speaker than the room requres is the answer to power compression.


    Randy
     
  15. billfitzmaurice

    billfitzmaurice Commercial User

    Sep 15, 2004
    New Hampshire
    Owner, Bill Fitzmaurice Loudspeaker Design
    True, that keeps things linear, but also makes for damn big boxes or awfully long ducts or both. I think 1/3 Sd is adequate except in extreme cases. BTW, pre-T/S it was commonly believed that ports had to be equal to Sd or larger, one reason why boxes were so big and standing waves actually were a problem. It wasn't unusual to have internal dimensions a quarter wavelength long to 100 Hz or lower.
     
  16. Monomer

    Monomer

    Jul 22, 2005

    It's for a 30SLIM


    so plate 1 is already reversed.
     
  17. I'm laying out Tuba 30 KindaSlim (28" width) so pairs will fit into our standard 5x8 cargo trailer. I'm still reversing the driver so the magnet hangs in the breeze for cooling purposes.

    The passive radiator is the solution for small boxes that require large vent diameters. This avoids the standing wave resonance problems, but introduces new issues particular to passive radiators. Always a compromise.

    The point most overlooked is very few drivers will reach full rated power without first exceeding Xmax. This is almost the exclusive domain of JBL and Beyma drivers. Most of the Eminence drivers exceed Xmax far lower than their maximum power ratings.
     
  18. fdeck

    fdeck Supporting Member Commercial User

    Mar 20, 2004
    Madison WI
    HPF Technology LLC
    Just a couple questions... about vent pressure, is this problem adequately handled by looking at the graphs of vent air velocity? I have seen rules-of-thumb that this velocity should be kept small compared to the speed of sound.

    Also, about rated power, isn't that just the thermal power rating? Even for an ideal driver, the maximum amplitude for linear response depends on the amplitude of cone excursion. There are nonlinear components to the suspension and box compliance, but are these significant effects?
     
  19. billfitzmaurice

    billfitzmaurice Commercial User

    Sep 15, 2004
    New Hampshire
    Owner, Bill Fitzmaurice Loudspeaker Design
    What the air velocity warning tells you is whether you can expect port noise, but if you've got enough pressure there to make noise it's a safe bet there's enough flow restriction to cause loss of output.

    The power rating is usually thermal. Many pro-sound drivers have a built-in element of safety when it comes to excursion, which Bruce hit on: they limit the xmax to well below the xmech. You may be able to exceed linear xmax at full power but you won't be able to bottom out the coil. The good news is that this built-in power compression feature saves you a lot of money on drivers, but the bad is that you aren't going to have linear response to full power. Personally I'd rather not have to deal with blown drivers and I can always get that last 6dB of output by just using a second cabinet.
     
  20. Petebass

    Petebass

    Dec 22, 2002
    QLD Australia
    Exactly! I also scratch my head in wonder with the amount of threads where it’s stated that you should feed a speaker twice its rated power. I don't subscribe to that at all. Amp headroom is one thing, and we all want to avoid clipping, but what about the poor speaker?

    Furthermore, this thread confirms what my ears have been telling me for years. I've always preferred to use speakers that are rated for a lot more watts that the amp that drives them. According to a lot of folks that's "underpowering" and I'm going to blow the speakers. Funnily enough, I've never blown a speaker. The truth is I'm minimising the effects of power compression.

    There's another reason I do this and it's been mentioned by Bgavin:-

    Well I own cabs loaded with JBL and Beyma drivers and I can confirm this is definitely the case. They certainly outperform my expensive Eden speakers in this department.