[germanium]

Damping Factor, What is it & how does it effects drivers & drive systems?
 
Damping factor relate to the output impedance of the amplifier, The lower the output impedance the higher the damping factor. The main way of getting a high damping factor is to have an amp with very high negative feedback. In the olden days high feedback amps were frowned apon as most had really bad sound. This was blamed on the feedback & for the most part this was in fact true but it wasn't the only factor. The biggest issue was having high negative feedback combined with poor phase margin. You see high negative feedback combined with poor phasemargin caused intermodulation distortion as well as in really bad cases oscilation because as you go higher in frequency & you get beyond a cirtain frequency the negative feedback actually will turn into positive feedback which can cause oscilation. Modern high feedback amps tend to have a better phase margin & less likely to devolop these issues so many sound way better than early designs.
 
High feed back desigms did not really come into being until transistor amps were deveoped. It was discovered that high feedback amps could take tighter control of speaker cone motion. What most people don't tell you is that the extra control costs in terms of power available as it actually takes some power to shunt the driver back electromotive force which it does through the power supply. This wis whi many amps that just barely deliver thier rated power into a 8 ohm resister may actually start to clip at significantly lower power into a speaker because some of the available power is being used to shunt the back electromotive force of the driver. It's like driving your car with the emergency brake on.It takes more power to drive it to the same volume. This is why the claim that tubes play louder for a given power output than transistor gear. Usually by about 3db & sometimes even more.
 
During the early days of transistor amps they came up with a speaker design that would make use of the increased damping factor. That type of speaker is known as the acoustic suspension speaker which the first commercialy available to my knowledge was from Acoustic Research. These speakers have a sealed speaker cabinate & the bass extension was determined by the drive system resonance with a 12db/oct rolloff after you reach this resonance on the way down in frequency. Many of these speaker were designed with a drive total Q of greater than .707, .707 being optimally flat. Higher Q's gave a bump in the response at the resonance of the drive system. These speakers if hooked to a low feedback vacume tube amp would sound to varying degrees a little sloppy in the bass due to the natural bump in the response that already existed in these speaker at resonance. What would happen is the impedance peak of the drive system would interact with the high output impedance causing the voltage to rise along with the impedance of the driver thus essentuating the peak in the drive system. Transistor amp having higher feedback & lower output impedance did not do this making them perfect for this design of speaker. You could have a reasonably small cabinate combine with a high compliance driver & still get decent bass extension.
 
This of coarse is not true of vented spakers as they have thier peak response generally at the port resonance which actually is at the bottom of the impedance curve , not the top like it is with the sealed speaker. For this reason the vented speaker responds quite well to amps with a somewhat higher output impedance as in many cases the response will be smoother & slightly more extended I might add with the somewhat higher output impedance. Bear in minde these speaker designs were designed in the hayday of higher output impedance vacume tube amps. Horn speakers were likewise designed during this era & many of these designs sound horrid on low output impedance transistor gear.
 
 Crossover design must also take into account the output impedance of the amp that is driving them as well as the speaker impedance which is why you rarely see balaced crossover designs anymore as speaker & amp impedances are no longer matched as they were on the best tube gear of old. Here you could easily build a balanced cossover (using same parts for woofer & tweeter crossovers) & not end up with a hole in the response, Any more the woofer & tweeter parts are different as there would be a hole in the response if not with low output impedance amps.
 
While I've been talking speakers here similar applies to earphones as well, some designs work well with high output impedance & some don't & of coarse some don't really seem to care either way. It is foolish to say there is only one proper way to drive a speaker of any type. You have to drive them the way they are designed to be driven. Etymotic earphones tend to work beat with a higher output impedance.

 

[headfinoob]

Thanks for the write up- this hobby is like the never-ending science experiment.
 
I don't want to derail the topic, but I am curious if there are issues driving modern 8-ohm passive home speakers with modern class AB car audio amplifiers that are "meant" to drive 4-ohm loads.  I would expect that an amplifier could be "optimized" for a certain nominal impedance, but I assume that speaker impedance curves vary so much that this may not be the case.

 

[germanium]

Quote:

Thanks for the write up- this hobby is like the never-ending science experiment.
 
I don't want to derail the topic, but I am curious if there are issues driving modern 8-ohm passive home speakers with modern class AB car audio amplifiers that are "meant" to drive 4-ohm loads.  I would expect that an amplifier could be "optimized" for a certain nominal impedance, but I assume that speaker impedance curves vary so much that this may not be the case.

Actually the only compromise by driving a higher impedance speaker with a car amp is that you won't get the rated power as generally it is rated into 4 ohms not 8 ohms. However because it is a transistor amp the damping fact should be rather high in most cases. There are a few that use no feedback but one that I heard that had the option of no feedback sounded horrid at low volumes but sounded greal at high volumes This was due to not having sufficient bias on the output transistors. By the way if you turned on the feedback on this amp it sounded flat & lifeless. If this amp had better low level response without feedback this amp would have been the amp to beat for sound quality, such was it's sound with sufficient volume when the negative feedback was turned off. Sound was full of life without the feedback on this amp.
 
Power output into 8 ohms is about 1/2 that of it's power into 4 ohms assuming a tight power supply. With a looser power supply (underbuilt power supply) you could actually end up with near equal power.
 

 

[headfinoob]

Quote:

Actually the only compromise by driving a higher impedance speaker with a car amp is that you won't get the rated power as generally it is rated into 4 ohms not 8 ohms. However because it is a transistor amp the damping fact should be rather high in most cases. There are a few that use no feedback but one that I heard that had the option of no feedback sounded horrid at low volumes but sounded greal at high volumes This was due to not having sufficient bias on the output transistors. By the way if you turned on the feedback on this amp it sounded flat & lifeless. If this amp had better low level response without feedback this amp would have been the amp to beat for sound quality, such was it's sound with sufficient volume when the negative feedback was turned off. Sound was full of life without the feedback on this amp.
 
Power output into 8 ohms is about 1/2 that of it's power into 4 ohms assuming a tight power supply.
 

 
Interesting, I used to think amps were amps, but lately it seems that a particular amp should be chosen to match your speakers.  With all the amplifier classes available, and factors like different damping factor and distortion at different volume levels, it gets confusing quickly.  Not to mention impedance curves, baffle type, efficiency and driver types of the speakers.

  It seems like one should basically pick an amplifier to meet their listening preferences after having chosen speakers.  Or conversely, chose speakers based upon an amplifier that you already have.
 
Hopefully after some learning, I can make a positive contribution this thread...
 
 

 

[JRG1990]

Feedback is not evil I don't know about the vintage stuff but in modern dacs and amps the right amount of feedback only reduces the disortion it doesn't colour the sound or anything else.
Amps are amps you just need to match the impendance don't buy a amp thats only rated for 8ohm to drive 4ohm speakers, a modern amp will be transperant sounding unless it's faulty or badly designed, it's the tube / vintage amps which may affect the sound and be fussy with speakers.
 
Also according to this http://sound.westhost.com/biamp-vs-passive.htm , with passive speakers the damping factor is 1-2 so passive speakers don't really have any damping so i'm not sure if what you said about the amp needing extra power to take control of the drivers is correct.
 
 

 

[MaciekN]

I understand how using negative feedback lowers output impedance but I don't get the connection between deep feedback and output impedance. Even if we use negative feedback to achieve a gain of 2 we already have the near-zero output impedance, no? What does deeper feedback have to do with it?

 

[Mauricio]

Quote:

 
efficiency and driver types of the speakers.

  It seems like one should basically pick an amplifier to meet their listening preferences after having chosen speakers.  Or conversely, chose speakers based upon an amplifier that you already have.
 
 

who is in the best position to determine, for a given price point, the best amplifier(s) for the drivers?  The engineers designing the speaker.  That's one of the premises of active speakers in which the amp guessing game is eliminated because the speaker already comes with amps precisely matched--for frequency range, impedance, damping and efficiency--to the driver.
 

 

[firev1]

I find your article pretty insightful but essentially flawed. Transistor amps were built at first to combat the "evils" of tube gear, which generally require maintenance(tube replacement) and more importantly, help with sound engineers with music production(more precise/transparent). Sure you could get a very nice musical listening session with low/no feedback and high impedance designs but it will never match the sheer transparency of the higher feedback, lower impedance designs.
 
Of course to the average user, high damping factor is not extremely important, but to some, staying true to the source counts more for some of us and specs just helps(not confirm) us to do that. Also if you would like, a lower damping factor can be achieved by the simple addition of impedance adaptor/resistor of choice which by TOTAL COINCIDENCE is precisely what Etymotic did too with their ER4PT and ER4S models. Another factor that I may highlight with Etymotics(especially ER-4), higher Z = better response accuracy is as I speculate(opinion) is to be compatible with high Z stage gear outputs like mixing boards etc. 
 
Lastly I'm sure all of you guys know the fluctuations of THD with higher output impedance with some headphones, a good lower Z SS amp(no need for excess damping) will ultimately play very well with most headphones if not all, which is why SS amps are met with very widespread popularity too.

 

[germanium]

Quote:

I find your article pretty insightful but essentially flawed. Transistor amps were built at first to combat the "evils" of tube gear, which generally require maintenance(tube replacement) and more importantly, help with sound engineers with music production(more precise/transparent). Sure you could get a very nice musical listening session with low/no feedback and high impedance designs but it will never match the sheer transparency of the higher feedback, lower impedance designs.
 
Of course to the average user, high damping factor is not extremely important, but to some, staying true to the source counts more for some of us and specs just helps(not confirm) us to do that. Also if you would like, a lower damping factor can be achieved by the simple addition of impedance adaptor/resistor of choice which by TOTAL COINCIDENCE is precisely what Etymotic did too with their ER4PT and ER4S models. Another factor that I may highlight with Etymotics(especially ER-4), higher Z = better response accuracy is as I speculate(opinion) is to be compatible with high Z stage gear outputs like mixing boards etc. 
 
Lastly I'm sure all of you guys know the fluctuations of THD with higher output impedance with some headphones, a good lower Z SS amp(no need for excess damping) will ultimately play very well with most headphones if not all, which is why SS amps are met with very widespread popularity too.

Actually many studios went back to tube gear after using the early transistor gear because as my engineer friend put it it sounded like they threw the baby out with the bath water. Music was flat & lifeless not to mention lacked any kind of smoothness with the speakers of the day (mostly horn or really big vented boxes that were very highly efficient, better than 98db/watt/m). These speakers did not need a high damping factor & it actually worked against these speakers sonically. A properly designed horn  can almost completely inihilate an impedance peak in the driver in excess of 120 ohms in an 8 ohm speaker due to the emmense air load that the horn provides. Horn drivers use extremely powerfull magnets & any visable movement of the driver will develop huge back EMF making it common that horn drivers in free air to have impedance peaks greater than 10 times the rated driver impedance. This peak mostly disappears once attached to a properly designed horn. Damping the driver electrically is not needed as the air load will damp it. The driver is said to be locked down by the air load.
 
Modern Solid State gear is considerably better than the old gear but still does not work well with horns as well a true proffesional vacumm tube amp with no feedback that has a matching output impedance. There is just no comparison Horns really like to be turned loose to sound thier best.
 

 

[firev1]

Quote:

Actually many studios went back to tube gear after using the early transistor gear because as my engineer friend put it it sounded like they threw the baby out with the bath water. Music was flat & lifeless not to mention lacked any kind of smoothness with the speakers of the day (mostly horn or really big vented boxes that were very highly efficient, better than 98db/watt/m). These speakers did not need a high damping factor & it actually worked against these speakers sonically. A properly designed horn  can almost completely inihilate an impedance peak in the driver in excess of 120 ohms in an 8 ohm speaker due to the emmense air load that the horn provides. Horn drivers use extremely powerfull magnets & any visable movement of the driver will develop huge back EMF making it common that horn drivers in free air to have impedance peaks greater than 10 times the rated driver impedance. This peak mostly disappears once attached to a properly designed horn. Damping the driver electrically is not needed as the air load will damp it. The driver is said to be locked down by the air load.
 
Modern Solid State gear is considerably better than the old gear but still does not work well with horns as well a true proffesional vacumm tube amp with no feedback that has a matching output impedance. There is just no comparison Horns really like to be turned loose to sound thier best.
 

Sorry Germanium, I get your point there about using vacuum tubes with horns since horns have different methods of damping(air damping if I understand correctly) but your article also pointed at the damping of some headphone driver designs which to my limited understanding do not(I think) have the same air load damping mechanisms as speakers do. I would like to hear your take on that.

 

[germanium]

Quote:

Sorry Germanium, I get your point there about using vacuum tubes with horns since horns have different methods of damping(air damping if I understand correctly) but your article also pointed at the damping of some headphone driver designs which to my limited understanding do not(I think) have the same air load damping mechanisms as speakers do. I would like to hear your take on that.

Dynamic driver headphones generally have massive damping pads attached to the back of the driver that no amount of electrical damping would be able to control as well as these damping pads do considering the size of the magnets on the earphones. Plus even if the magnets were larger you would have to provide the extra damping because of the air space they fire in to is much smaller than this type of driver was originally designed to work with.
 
Balanced armature drivers also depend on external damping to reduce resonances in the drive system though the resonances are caused from a different source. Tube resonances rather actual driver resonances. The balanced armature driver has a sealed space behind the driver that limits excessive bass response so damping is not used for that purpose as it is with dynamic driver headphones. In other words the current production of balanced armature drivers don't need damping to get proper bass response, only to reduce tube resonances.Anouther attribute of balanced armature earphones is the driver is sized more properly for the air space they are firing into which is very small.
 
Multidriver IEMs that have crossovers need to be driven with amps that match the impedance of the amps they were origially designed to work with , Otherwise the crossovers will not not work correctly. The amp they were originally designed to work with may not nessearily have a low output impedance approaching zero. Shure has amps that they make for thier in-ear monitors that would be instructive to measure thier output impedance to see what they were designed to work with at least for Shure brand in-ear monitors. I'm betting the output impedance of these amps is not zero. Probably in the 5-10ohm range I'm guessing

 

[Anaxilus]

UE actually told me the same thing about the UE18.  Was not designed to run from a iPod but various packs used on stage.

 

[germanium]

Quote:

I understand how using negative feedback lowers output impedance but I don't get the connection between deep feedback and output impedance. Even if we use negative feedback to achieve a gain of 2 we already have the near-zero output impedance, no? What does deeper feedback have to do with it?

Transistors can have relatively low output impedance without feedback but not the super low output impedances that result in damping factors greater than 100. I believe the pass labs aleph which is zero feedback comes in at a damping factor of about 70. Amps that have feedback can achieve damping factors of over 1000. Granted that once you pass this though a crossover of a passive speaker super high damping factors don't mean much but there still is more than a a damping factor of 5 at the most critical frequency with most passive speakers which is the woofer resonance. By the way the Aleph is a super heave built class a amp that is run single ended but with a constant current source in order to provide a cancelation of a substantial d.c. offset that would be present in a single ended direct couped amp such as the aleph.
 
I have done experiments with weighted driver that have demonstrated the effects of damping factor even through the crossover as still having a substantial effect on driver cone motion stopping the driver far more quickly than if the amp was turned off and then the driver being thumped with the same vigor as when the amp was on.
 
This type of behavior from the amp does deminish available power with amps that have underbuilt power supplies as it takes power to accompish this feat. If you have an amp with regulated power supplies like some of Krell & Mark Levinsons this type behavior  will not reduce power available to drive the speaker as there is huge power reserves sitting behind the regulator that will replenish the power after the regulator as it is being used. Regulators artificially limit the power to a specified voltage level until you reach a load that exceeds the power supplies ability to keep the capactors full that come before the regulators. In most cases that would be between 1-2 ohm loads before you can sufficiently deplete the power supply before the regulators. Krell at one point had a monster amp that could take this all the way down to a 1/2 ohm load & stay with in regulation. At this load level it could provide 16,000 watts of regulated power!!! It required it's own transformer on the power pole & needed 2-60amp 220 volt lines per channel to reach this level of performance. 
 

 

[xnor]

The thing with 0 output impedance amps is that you can go anywhere from there. Add two resistors and you can have any output impedance you desire as long as the amp is not too weak to begin with. Theoretically, these resistors could be added to the headphones right in front of the drivers. This of course doesn't work with headphone amps that have an output impedance of tens or even hundreds of ohms.

 

[germanium]

Quote:

The thing with 0 output impedance amps is that you can go anywhere from there. Add two resistors and you can have any output impedance you desire as long as the amp is not too weak to begin with. Theoretically, these resistors could be added to the headphones right in front of the drivers. This of course doesn't work with headphone amps that have an output impedance of tens or even hundreds of ohms.

You can do that but it still doesn't sound the same as a zero feedback tube amp that has a true matching output impedance. I know because I've had both & made my own impedance matching networks for transistor amps in order to drive horns & while it did improve the sound over the original transistor sound the zero feedback tubes still had the sound to die for even though the output impedance was matched on both with the output network on the transistor amp. Even the output power was matched with the tube amp with the impedance matching network on the tranny amp, 25 Watts /channel to the speaker. By the way the tranny amp still neded to have a resistor network across the terminals at the back of the amp as well as the resistors in series with the speaker to sound it's best with the Altec Lansing Voice Of The Theatre A7-500 that I had at the time. This in part hid the reactance of the speaker from the amp Which was an Adcom GFA 545 which was modified. I had used it to drive many other speakers before, even some as low 2ohms & is sounded great on them but really needed the impedance matching network to sound it's best on the horns even though the load was an easy load for the amp. The easiest load that amp has ever seen with me. 
 
Of coarse this does not apply to headphones, They should be driven with the output impedance that they were designed with originally & not make assumptions that we know best how to drive them without consulting the manufacturer.