Damping Factor - Actual listening tests?

Jason Lavoie

Junior
Jan 13, 2011
459
0
16
Ottawa
Has anyone done any research or subjective listening tests relating to Damping factor?

Aside from power loss in a long run of cable DF seems to be the next item on the hit list.
However I have a few thoughts
Since DF matters most for subs and low frequencies, is there a chart (that maybe doesn't exist yet) of acceptable DF vs HPF of the signal?

There are times when a larger amplifier might be cheaper or easier than heavier cable
I'm talking about fixed installs where the speaker cable might be 200-300' long and we're trying to use existing conduits or something like that.

I envision a listening test where long speaker cables of various gauges are switched in while simultaneously adding makeup gain to keep the volume the same and then listening for any difference.

Would love it if someone has already figured this out, but if not then I'm accepting suggestions of testing methods :)

Jason
 
Re: Damping Factor - Actual listening tests?

Has anyone done any research or subjective listening tests relating to Damping factor?

Aside from power loss in a long run of cable DF seems to be the next item on the hit list.
However I have a few thoughts
Since DF matters most for subs and low frequencies, is there a chart (that maybe doesn't exist yet) of acceptable DF vs HPF of the signal?

There are times when a larger amplifier might be cheaper or easier than heavier cable
I'm talking about fixed installs where the speaker cable might be 200-300' long and we're trying to use existing conduits or something like that.

I envision a listening test where long speaker cables of various gauges are switched in while simultaneously adding makeup gain to keep the volume the same and then listening for any difference.

Would love it if someone has already figured this out, but if not then I'm accepting suggestions of testing methods :)

Jason

Marty McCann Peavey's clinician used to do a demo where he artificially lowered the damping factor for some speakers on sticks to show what low damping factor sounded like.

I attended one such session and as expected the differences were subtle and not even heard by all in attendance.

Since I haven't said this lately, DF is an old archaic specification left over from tube amplifier days, when amps had higher output impedance. It takes a lot of wire to make that much source impedance. Then the audibility depends on the speaker, the less flat the speaker impedance vs frequency, the more audible any errors generated.

If you want to do your own experiments you can do a crude null by setting up two identical speakers driven by identical (or the same amp channel) through different wire lengths.

Then bridge a mono headphone across the two loudspeaker hots, or maybe one can between the hots and another between the grounds, while the null will be better if you use short cables for both returns and just the long cable for the one hot.

For fixed install running speakers at 70V/100V constant voltage with step down transformers helps manage wire losses. For subs you are generally better off locating the amps proximate to the subs since LF will not play nicely with step up/down transformers.

JR
 
Re: Damping Factor - Actual listening tests?

Has anyone done any research or subjective listening tests relating to Damping factor?

I envision a listening test where long speaker cables of various gauges are switched in while simultaneously adding makeup gain to keep the volume the same and then listening for any difference.

Would love it if someone has already figured this out, but if not then I'm accepting suggestions of testing methods :)

Jason
Bennett figured it out and wrote about it:

http://www.bennettprescott.com/downloads/dampingfactor.pdf

Long cable runs with low impedance loads make the lack of damping factor noticeable, increasing power does not fix the problem.
 
Re: Damping Factor - Actual listening tests?

Once you factor in ALL of the losses (connectors on the amp-the speaker-the speaker cable-the cable itself etc, the damping factor starts to become almost a non issue. The cable size/length alone often greatly outweighs any advantage the amp would offer.

I have done side by side testing of amps with high damping factor and those with much lower damping factor. Some of the lower DF had the "tighter" sound.

In my opinion (others are free to correct or disagree), this comes from the early days/decades of solid state amps.

The amps that had the tightest sound also had the highest damping factor-and were the most expensive. So the two got associated with each other.

Times have changed however.

The idea is valid, but (again in my opinion), there are much more important things to worry about.
 
Re: Damping Factor - Actual listening tests?

Once you factor in ALL of the losses (connectors on the amp-the speaker-the speaker cable-the cable itself etc, the damping factor starts to become almost a non issue. The cable size/length alone often greatly outweighs any advantage the amp would offer.

I have done side by side testing of amps with high damping factor and those with much lower damping factor. Some of the lower DF had the "tighter" sound.

In my opinion (others are free to correct or disagree), this comes from the early days/decades of solid state amps.

The amps that had the tightest sound also had the highest damping factor-and were the most expensive. So the two got associated with each other.

Times have changed however.

The idea is valid, but (again in my opinion), there are much more important things to worry about.

Not sure what you are even saying...
---Some of the lower DF had the "tighter" sound.
---The amps that had the tightest sound also had the highest damping factor

Or what "tight" sound is.

This is a very old subject IMO made moot wrt amp design decades ago by modern solid state amps.

Silly high published amp DF specs are a stupidity tax on users that don't understand the need to add wire resistance to any real world DF calculation.

Jason is asking about wire related DF effects not amp DF. I already suggested a way for him to listen to just the wiring caused error using cans.

JR
 
Re: Damping Factor - Actual listening tests?

Bennett figured it out and wrote about it:

http://www.bennettprescott.com/downloads/dampingfactor.pdf

Long cable runs with low impedance loads make the lack of damping factor noticeable, increasing power does not fix the problem.

I've read it, and others like it, and while they illustrate how easily overloading a long line can lead to low DF, no one seems to comment on what the result of low DF sounds like, or at what point it becomes measurable/audible/annoying.
I've seen a few references that say to keep it above 10, but other than that it seems like just opinion plus a general (and obvious) consensus that we need thicker cable for subs.


Here's the example that got me wondering:
2 channel bridgeable amplifier, 300' of #14 2-pair, 4x8ohm speakers in a cluster, or at least in an area where all 4 require the same signal
(assume the wiring and all equipment is fixed and we have to use it)

Option 1:
run the speakers in 2 pairs down to 4 ohms each and connect to the amplifier

Option 2:
run the speakers in series-parallel (net 8ohms), double up the wire (about the equivalent of #11) and bridge the amp

ignoring the effects of the wire length this should result in the exact same power applied to the speakers
Except once we take the wire into account the DF goes from 2.6 to 10.2 if we choose option 2
So I asked myself: is the unconventional wiring, and the chance that a failure in one speaker might take 2 of them out of commission offset by the increase in Damping Factor?

And of course the answer is: It depends

So the follow up question is: How much better will a DF of 10.2 sound than 2.6? and I have no idea.
Will it be noticeable? would its detrimental effects only be noticeable at frequencies outside the pass band of these speakers?


I'm not opposed to hooking up a few spools of wire and testing, but at best that will only give me a few data points, and then it'll only be subjective data.

Jason
 
Re: Damping Factor - Actual listening tests?

If you want to do your own experiments you can do a crude null by setting up two identical speakers driven by identical (or the same amp channel) through different wire lengths.

Then bridge a mono headphone across the two loudspeaker hots, or maybe one can between the hots and another between the grounds, while the null will be better if you use short cables for both returns and just the long cable for the one hot.

That might work, although I'd have to account for the level difference somehow.
Maybe I could pad them both and record it so that I can listen to the difference more critically

For fixed install running speakers at 70V/100V constant voltage with step down transformers helps manage wire losses. For subs you are generally better off locating the amps proximate to the subs since LF will not play nicely with step up/down transformers.
JR

Right, so that comes back around to asking if the performance hit from using 70V transformers is better or worse than just accepting the lower DF?
Trying to balance two subjective details.

Jason
 
Re: Damping Factor - Actual listening tests?

For the rare case where there is sufficient cable resistance to appreciably affect the response of a woofer near the box resonance frequency there are amplifiers available with adjustable negative output resistance. The Powersoft K series can do this for frequencies below 400 Hz. Bear in mind that by using negative output resistance you're flirting with instability. I run my Powersofts at zero output resistance (guaranteed stability) and use reasonably short, fat cables on subs. Thankfully, in situations where I've needed longer speaker runs it's been voice-only and high-passing pretty much avoids concerns about bottom octave response due to cable resistance.

For steady-state it should be possible to correct for cable resistance with EQ (within limits). It's more complicated for transients and I'd have to crawl off and think about it.

Yet another problem that will go away someday when ALL speakers are powered.

--Frank
 
Re: Damping Factor - Actual listening tests?

That might work, although I'd have to account for the level difference somehow.
Maybe I could pad them both and record it so that I can listen to the difference more critically
Pad? The headphone is a floating differential device that will express only the difference between the two speakers. Driving both speakers from a single channel there will be a flat wide band error for wire R, and a non-flat difference caused by the loudspeakers non-flat impedance. If you drive the speakers from two different amp channels you can use the gain on one to trim out the wide band component and only get the non-flat error. I might try to predict what this error will sound like from your speaker's impedance plot.
Right, so that comes back around to asking if the performance hit from using 70V transformers is better or worse than just accepting the lower DF?
Trying to balance two subjective details.

Jason

I don't know why you are fixated on DF. I wrote about this in my Audio Mythology column back in the '80s. It's an old subject we hashed to death over at PSW too.

Stop thinking about DF as some mysterious phenomenon. The wire and speaker are forming a simple voltage divider. So the top R is the wire DCR (well mostly DC), and the bottom R is the loudspeaker's resistance (hint not a simple DC resistance). For hundreds of feet of wire and low impedance loudspeaker that can vary a few ohms between 20Hz an 20kHz you can use math to calculate a projected frequency response error.

The problem with running low bass through step-up and step-down transformers is the core size has to double for every halving of the LF cut off. So a 20 Hz transformer has to be twice as large as a 40 Hz transformer to handle the same power. For perspective the audio transformers will be 2-3x the size of the mains power transformer in the amp.

You need to look at the impedance plots for your actual speakers and determine how sensitive they will be to the wire resistance you expect to use in series.

JR

PS: I recall one old exchange over at PSW with a fan of funny speaker wire who claimed he experienced some small gauge funny wire that sounded better to him that old school heavy gauge zip cord. After a little discussion we determined that the combination of his high resistance wire, and speaker impedance that varied vs frequency was creating an audible HF boost (actually LF was cut relative to the HF).
 
Re: Damping Factor - Actual listening tests?

We had a pair of subs, I believe Danley DBH218LC's, with Powersoft DSP power with damping factor compensation. One time we moved the subs, and one sub accidentally had a different cable plugged in (75' 12AWG rather than 5' 10AWG, or something like that, it's been a few years). Both the client and I noticed one sub, the one with the long cable, sounded a bit muddy. We tracked down the cables, adjusted the damping factor compensation, and they sounded the same again.

I'm sold on damping factor as a wiring design factor, but not as an amp spec factor. How noticeable it is depends on the subs, how compromised the wire length/diameter is, and how critical the listener is to accurate bass.

We had a church install with 175' or so subwoofer runs, so we used the 4ohm rather than 2ohm mode (Danley DBH), 10AWG wire, Powersoft DSP amps with damping compensation, and called it good. The client wanted the best bass possible, short of full room acoustic treatment, so that was a relatively inexpensive way to cross one variable off the list.
 
Re: Damping Factor - Actual listening tests?

Not sure what you are even saying...
---Some of the lower DF had the "tighter" sound.
---The amps that had the tightest sound also had the highest damping factor

Or what "tight" sound is.


JR
I was commenting that with the more modern amps (at the time in the late 90s) some of the ones with the highest damping factor had a "less controlled" sound, or more "flopping around" of the drivers.

The last comment you quoted was based on a decade or two earlier amps.

This was done with amps that were pretty close in power ratings, and all the gains were matched.

We used one of our front loaded 2x18" subs. The source material was recorded kick drums, low freq impulses and kick heavy music.

We did it in a blind test (not double blind) but close enough.

The same cable and speaker was used in all tests.

We listened and watched the speaker cones for movement. We did not measure the movement-so it was subjective.

The "tightness" was judged by the "free extra sounds" we could hear after the sound stopped from the source.

But that was almost 20 years ago-so I might come to different conclusions today. I believe we had 5 different amps that covered the range from cheap MI to "accepted on every rider" amps. So not a wide assortment-but it was all I could get my hands on at the time.
 
Re: Damping Factor - Actual listening tests?

The only time I have heard significant damping factor was at school when our county's tech sent the speaker level signal through 100+ feet of 28 gauge telephone wire.

Sent from my XT1254
 
Re: Damping Factor - Actual listening tests?

The only time I have heard significant damping factor was at school when our county's tech sent the speaker level signal through 100+ feet of 28 gauge telephone wire.

OK, story time: Back in the '80s there was a joint meeting of the AES San Francisco Section and the East Bay Skeptics Society (or something close) to debunk speaker cable myths and have a little fun. We were at an auditorium somewhere in Berkeley and had a pair of very nice, large studio monitors set up for playback and A-B-X testing of the cables. We also had a room full of studio guys, golden ears, crusty old farts, etc. The experimenters were careful to match levels at the speakers after switching cables.

We tried a wide variety of cables versus some standard install stuff like 50 ft. of 12-2 with the gray vinyl jacket. You guys probably know the Belden number. Anyway, the ONLY "cable" that anyone could hear a difference on was, as I recall, 30 ft (one way) of 30 AWG wire-wrap wire. The deep base was a little different. A good time was had by all. (Was anyone else on the forum there?)

--Frank
 
Re: Damping Factor - Actual listening tests?

The damping factor spec is mainly a measure of output impedance, mostly reduced by large amounts of negative feedback but that doesn't tell the whole story about how much current the amp can deliver, when, and how the amp will behave dynamically. The hypothetical possibilities of what could cause audible artifacts in circa '90s MI amps is not on point. By the late '90s decent amps were quite good, but I repeat excessive DF will not tell you much useful about how the amp will behave, more about the physical design for marketing purposes (NF taken right from the jack to eliminate mOhms of DCR from internal point to point wiring).

I repeat a null test performed with real wire and the actual speakers might tell Jason something useful regarding his specific application.

JR

PS: To insure that the amplifier has good current delivery capability, avoid marginal (2 ohm or 4 ohm bridged) loads.
 
Re: Damping Factor - Actual listening tests?

I don't know why you are fixated on DF. I wrote about this in my Audio Mythology column back in the '80s. It's an old subject we hashed to death over at PSW too.

I'm fixated on it because it's a spec that people seem to hold as important (in regards to wire gauge and length) but nobody seems to know anything more than 'higher is better'

Stop thinking about DF as some mysterious phenomenon. The wire and speaker are forming a simple voltage divider. So the top R is the wire DCR (well mostly DC), and the bottom R is the loudspeaker's resistance (hint not a simple DC resistance). For hundreds of feet of wire and low impedance loudspeaker that can vary a few ohms between 20Hz an 20kHz you can use math to calculate a projected frequency response error.

I get the voltage divider, and it is easy to calculate how much power gets lost in the wire vs how much gets delivered to the speaker.
I'm only just now (just this second as I write this) having the light bulb go off re: the voltage divider is different at different frequencies (because of the speaker's impedance curve)

I get it now. so if the speaker's graph shows that the impedance is 8 ohms at 100hz and 80 ohms at 2k then obviously 2k will hardly be affected by a long cable while 100hz will be attenuated a lot more.

Now that i understand this it sounds like something that could be 100% counteracted with EQ. I had visions of the effect being something more dynamic and time dependent the way people talk about DF.

So correct me if I'm wrong here: if the power loss due to the cable is acceptable, and we're going to measure the system with Smaart and shape the response in-situ then the wire gauge (and hence the DF) can be ignored within reason?


Maybe I've had this all backwards. I had pictured it to be more of a consideration for fixed installs where wires are often long, but in reality I guess it is more of a concern for people who are deploying the same system but with different speaker cable lengths every weekend and expecting it to have the same response..

You need to look at the impedance plots for your actual speakers and determine how sensitive they will be to the wire resistance you expect to use in series.

I see that now. thanks for pushing me in the right direction!

Jason
 
Re: Damping Factor - Actual listening tests?

I'm fixated on it because it's a spec that people seem to hold as important (in regards to wire gauge and length) but nobody seems to know anything more than 'higher is better'



I get the voltage divider, and it is easy to calculate how much power gets lost in the wire vs how much gets delivered to the speaker.
I'm only just now (just this second as I write this) having the light bulb go off re: the voltage divider is different at different frequencies (because of the speaker's impedance curve)

I get it now. so if the speaker's graph shows that the impedance is 8 ohms at 100hz and 80 ohms at 2k then obviously 2k will hardly be affected by a long cable while 100hz will be attenuated a lot more.

Now that i understand this it sounds like something that could be 100% counteracted with EQ. I had visions of the effect being something more dynamic and time dependent the way people talk about DF.

So correct me if I'm wrong here: if the power loss due to the cable is acceptable, and we're going to measure the system with Smaart and shape the response in-situ then the wire gauge (and hence the DF) can be ignored within reason?


Maybe I've had this all backwards. I had pictured it to be more of a consideration for fixed installs where wires are often long, but in reality I guess it is more of a concern for people who are deploying the same system but with different speaker cable lengths every weekend and expecting it to have the same response..



I see that now. thanks for pushing me in the right direction!

Hi Jason,

An amplifier’s dampening factor is all about its ability to control the movement of the speaker cone. For the amplifier to be able to control the movement of the cone and the back EMF produced by the cone moment, it needs a “stiff” link between the two; and the output impedance of the amplifier also needs to be low.

You cannot correct for dampening factor with EQ or gain.

This paper explains what's happening - http://www.powersoft-audio.com/en/downloads/technical-notes/806-damping-control/file
 
Last edited by a moderator:
Re: Damping Factor - Actual listening tests?

Hi Jason,

An amplifier’s dampening factor is all about its ability to control the movement of the speaker cone. For the amplifier to be able to control the movement of the cone and the back EMF produced by the cone moment, it needs a “stiff” link between the two; and the output impedance of the amplifier also needs to be low.

You cannot correct for dampening factor with EQ or gain.

This paper explains what's happening - http://www.powersoft-audio.com/en/downloads/technical-notes/806-damping-control/file

Not to quibble but if you look at that article's electrical model the voice coil resistance and inductance is in series, so the ability of amplifier output impedance to damp that motor mass could indeed make a difference between high output impedance tube amps, but low Z solid state amps will all be similar and dominated by voice coil impedance. Of course if you use 4-8 ohms worth of wire resistance it should matter.

Peavey (Sondermeyer) got a patent for an amp that used a feedback trick to deliver silly high damping factor. In fact we could actually deliver a negative output impedance, which could compensate for wiring losses outside the amp too. I didn't even try to market this (I was power amp PM at the time) since I feared Peavey customer's heads would explode. I am not smart enough to do this myself but think there might be some tricks possible inside a powered speaker to drive individual drivers from a negative source impedance, but that was even further from Peavey's sweet spot in the market

Peavey (Sondermeyer/Brown) also patented amps with variable damping factor for use inside guitar amps to help solid state guitar amps mimic tube amp's interaction with driver/cabinet resonances. IIRC the peavey guitar amp feature had separate HF damping and LF damping adjustments (while they didn't call it DF). So if you want to hear really low DF maybe check out one of those Peavey guitar amps. :) (Audibly these circuits would enhance speaker voicing resonances, a part of guitar amp design, but now with a "more" knob.)

I didn't read that paper word for word, but since it comes from a power amp company I suspect they are promoting their amps as somehow better in this regard. Crown was notorious for hyping DF well beyond useful amounts.

With the exception of Tom Danley's old servo woofers loudspeaker driver position is not strictly controlled, just accelerated in one direction or the other. Remarkable they work as well as they do.

JR
 
Re: Damping Factor - Actual listening tests?

As an aside, since in the circuit model Rs (source resistance, as due to the cable) is in series with Re (voice coil resistance), cable resistance has exactly the same effect as "power compression" in which Re increases due to heating of the voice coil. -F