SPL voltage verses power
In my experience, both voltage and power reference units are informative. Sure, if you know impedance, you can calculate either knowing the other. But for comparisons, it always seems that one is a little more handy than the other. It usually depends on what I'm measuring, and what I'm comparing with.
For example, if I'm matching levels between a woofer and a tweeter, knowing the voltage sensitivity is most important to me. The amplifier voltage doesn't change at the crossover point even if the load varies, provided the load isn't too close to the internal source resistance of the amp. So knowing what each driver does at some reference voltage is most helpful.
On the other hand, if I'm comparing speakers at high power levels, I almost always want equal power measurements. If you set a fixed voltage level, then speakers with different impedances will measure different SPLs even if their efficiency is about the same. So the impedance difference has to be compensated for, either by post-processing or by setting the drive voltage.
I've come to a conclusion that the best way to do this is to find the minimum impedance of each speaker in the passband. An impedance sweep is run and minimum impedance is used to calculate drive voltage for the intended power level. This gives accuracy within 1dB, and nobody has to post-process the data to match levels.
This is really important, I think, for giving confidence in SPL shootouts and things like that. By setting drive levels beforehand, where all participants see the process and agree, there is no need for manipulating the data afterward to compensate for impedance. This give a greater degree of confidence in the data, because it ensures that nobdody has the upper hand. It's a minor thing, not a technical issue at all, but it does tend to remove any potential for skepticism in the measurement results.
Some have said average impedance or even advertised impedance might be a better value to use than minimum impedance. I wouldn't be opposed to that, except to find average impedance requires some mechanism to find the area under the impedance curve in a measured dataset. While not impossible, it isn't trivial in the field. However, finding Zmin is easy. And if you do the math, there's only about 0.25dB to 0.5dB difference in SPL between speakers driven to power calculated using Zmin verses Zavg.
Here's an example, an actual loudspeaker measured at a Prosound Shootout. This is the impedance chart of a loudspeaker measured there:
Zmin is approximately 9 ohms. From the forumula E = sqrt(PZ), we find the drive voltage required to dissipate 100 watts in a 9 ohms load is 30 volts.
What would you say average impedance is? Should it be 8 ohms, because that's the closest multiple of 8? Should it be 12 ohms, since that's the next higher multiple of 4? Or should it be closer to 10 ohms, since averaging the area under the curve up to 150Hz puts it closer to that?
Let's run the numbers and see what each of the drive voltages would work out to:
28.3v into 8 ohms dissipates 100 watts
30.0v into 9 ohms dissipates 100 watts
31.6v into 10 ohms dissipates 100 watts
34.6v into 12 ohms dissipates 100 watts
Now let's calculate the difference between each of these in decibels:
The 28.3v (8 ohm) level is 0.94x the 30.0v level, which is -0.25dB.
The 31.6v (10 ohm) level is 1.05x the 30.0v level, which is +0.23dB.
The 34.6v (12 ohm) level is 1.15x the 30.0v level, which is +0.62dB.
We're only talking about a half decibel max difference between Zmin and any sort of realistic average impedance. You actually could eye-ball an average value and be pretty close, like probably less than +/-0.25dB. But the minimum impedance method uses a tangible number that nobody has to argue about. It's what you see on the chart.
That's why I like the minimum impedance method for calculating the drive voltage used to set power levels for testing. I like using fixed voltage drive too. Seems to me a useful battery of tests is to measure at 2.83v, 1W, 10W, 100W, etc. It lets you see not only what the speaker does at low drive levels, but also at high drive levels. And it makes interpretation much easier, without having to post-process the data.
In my experience, both voltage and power reference units are informative. Sure, if you know impedance, you can calculate either knowing the other. But for comparisons, it always seems that one is a little more handy than the other. It usually depends on what I'm measuring, and what I'm comparing with.
For example, if I'm matching levels between a woofer and a tweeter, knowing the voltage sensitivity is most important to me. The amplifier voltage doesn't change at the crossover point even if the load varies, provided the load isn't too close to the internal source resistance of the amp. So knowing what each driver does at some reference voltage is most helpful.
On the other hand, if I'm comparing speakers at high power levels, I almost always want equal power measurements. If you set a fixed voltage level, then speakers with different impedances will measure different SPLs even if their efficiency is about the same. So the impedance difference has to be compensated for, either by post-processing or by setting the drive voltage.
I've come to a conclusion that the best way to do this is to find the minimum impedance of each speaker in the passband. An impedance sweep is run and minimum impedance is used to calculate drive voltage for the intended power level. This gives accuracy within 1dB, and nobody has to post-process the data to match levels.
This is really important, I think, for giving confidence in SPL shootouts and things like that. By setting drive levels beforehand, where all participants see the process and agree, there is no need for manipulating the data afterward to compensate for impedance. This give a greater degree of confidence in the data, because it ensures that nobdody has the upper hand. It's a minor thing, not a technical issue at all, but it does tend to remove any potential for skepticism in the measurement results.
Some have said average impedance or even advertised impedance might be a better value to use than minimum impedance. I wouldn't be opposed to that, except to find average impedance requires some mechanism to find the area under the impedance curve in a measured dataset. While not impossible, it isn't trivial in the field. However, finding Zmin is easy. And if you do the math, there's only about 0.25dB to 0.5dB difference in SPL between speakers driven to power calculated using Zmin verses Zavg.
Here's an example, an actual loudspeaker measured at a Prosound Shootout. This is the impedance chart of a loudspeaker measured there:
Zmin is approximately 9 ohms. From the forumula E = sqrt(PZ), we find the drive voltage required to dissipate 100 watts in a 9 ohms load is 30 volts.
What would you say average impedance is? Should it be 8 ohms, because that's the closest multiple of 8? Should it be 12 ohms, since that's the next higher multiple of 4? Or should it be closer to 10 ohms, since averaging the area under the curve up to 150Hz puts it closer to that?
Let's run the numbers and see what each of the drive voltages would work out to:
28.3v into 8 ohms dissipates 100 watts
30.0v into 9 ohms dissipates 100 watts
31.6v into 10 ohms dissipates 100 watts
34.6v into 12 ohms dissipates 100 watts
Now let's calculate the difference between each of these in decibels:
The 28.3v (8 ohm) level is 0.94x the 30.0v level, which is -0.25dB.
The 31.6v (10 ohm) level is 1.05x the 30.0v level, which is +0.23dB.
The 34.6v (12 ohm) level is 1.15x the 30.0v level, which is +0.62dB.
We're only talking about a half decibel max difference between Zmin and any sort of realistic average impedance. You actually could eye-ball an average value and be pretty close, like probably less than +/-0.25dB. But the minimum impedance method uses a tangible number that nobody has to argue about. It's what you see on the chart.
That's why I like the minimum impedance method for calculating the drive voltage used to set power levels for testing. I like using fixed voltage drive too. Seems to me a useful battery of tests is to measure at 2.83v, 1W, 10W, 100W, etc. It lets you see not only what the speaker does at low drive levels, but also at high drive levels. And it makes interpretation much easier, without having to post-process the data.
If you want to impress someone with your horn speaker setup, play Instrumentals, jazz, brass instruments, and specially percussion and drums. Thats easy, and can be reproduced with life like dinamics. you wont even find out, if your horns honk terribly. I have some 25 years old direct to disc LP's from Jeton, Charly Antolini , a swiss drummer, Knock Out, Countdown, and Menue. Playing these with horns is just plain crazy.... Robert, have you heard voices, specially soprano, classical music , and piano ? thats difficult. If it pasts the test with this kind of music, than the setup might be really good....
Angelo
Angelo,
Yes, that's where the swings had problems, not with percussion.
Just with the colors of more complex compositions.
Mostly in higher octaves, yes soprano more than alto, and alto more than lower, and in symphonic works.
Acoustic guitar was very nice though, even at very low leval.
(About the same range).
Yes, that's where the swings had problems, not with percussion.
Just with the colors of more complex compositions.
Mostly in higher octaves, yes soprano more than alto, and alto more than lower, and in symphonic works.
Acoustic guitar was very nice though, even at very low leval.
(About the same range).
Re: SPL voltage verses power
Djezzz Wayne
You took the time to write it. I didn't have the time...but who comes first, gets the premium.
Of course what impendance plot you show of what ever this unfiltered driver shows here, requires at lot of impedance notches to be straight pure perfect perfect plot acoustistically, so that static componets filtering the certain unit does have a chance to do it right. Only to say it popurlarely: The filter works as long the Voice Coil doens't move.
You present a showcase here that proff's it all.
We have only this little problem to deal with...oh boy...a massive compromise:
As soon as the voice coil starts to move in a magnetical gap the picture is different and the variables deriving from that action leads to: we have to make filters that varies simultaniously. Who in the world can solve the problem having static componets being suddenly dynamic? Following the dynamics of the TS parameters when the music is making our feet stamp of pleasures? Unsolvable.
You hit a good point here. One way is to drag AC impedance down to being VC DC impedance. This is one way to go. Like Steen Duelund would have done it. But it doesn't change the pricture, as the VC moves and the TS parameters changes anyhow.
But it doesn't work IMO as the filter itself with it's many needed notches introduces so many filter components parasites that following the natural drivers behaviour of a given driver, whatever it is, does not contribute to better sound, but oposite kills the sound naturalness build into the driver itself, if it has it at all. And it has as the statics and the nature dunamics conflicting. Completely.
For me your contrution here makes sence one way or the other. What way you choose to go will anyhow end up in a certain determnined sound.
Nice meeeting you here.
It' bed time and it's the nature of law governing.
Gerner
Wayne Parham said:In my experience, both voltage and power reference units are informative. Sure, if you know impedance, you can calculate either knowing the other. But for comparisons, it always seems that one is a little more handy than the other. It usually depends on what I'm measuring, and what I'm comparing with.
For example, if I'm matching levels between a woofer and a tweeter, knowing the voltage sensitivity is most important to me. The amplifier voltage doesn't change at the crossover point even if the load varies, provided the load isn't too close to the internal source resistance of the amp. So knowing what each driver does at some reference voltage is most helpful.
On the other hand, if I'm comparing speakers at high power levels, I almost always want equal power measurements. If you set a fixed voltage level, then speakers with different impedances will measure different SPLs even if their efficiency is about the same. So the impedance difference has to be compensated for, either by post-processing or by setting the drive voltage.
I've come to a conclusion that the best way to do this is to find the minimum impedance of each speaker in the passband. An impedance sweep is run and minimum impedance is used to calculate drive voltage for the intended power level. This gives accuracy within 1dB, and nobody has to post-process the data to match levels.
This is really important, I think, for giving confidence in SPL shootouts and things like that. By setting drive levels beforehand, where all participants see the process and agree, there is no need for manipulating the data afterward to compensate for impedance. This give a greater degree of confidence in the data, because it ensures that nobdody has the upper hand. It's a minor thing, not a technical issue at all, but it does tend to remove any potential for skepticism in the measurement results.
Some have said average impedance or even advertised impedance might be a better value to use than minimum impedance. I wouldn't be opposed to that, except to find average impedance requires some mechanism to find the area under the impedance curve in a measured dataset. While not impossible, it isn't trivial in the field. However, finding Zmin is easy. And if you do the math, there's only about 0.25dB to 0.5dB difference in SPL between speakers driven to power calculated using Zmin verses Zavg.
Here's an example, an actual loudspeaker measured at a Prosound Shootout. This is the impedance chart of a loudspeaker measured there:
Zmin is approximately 9 ohms. From the forumula E = sqrt(PZ), we find the drive voltage required to dissipate 100 watts in a 9 ohms load is 30 volts.
What would you say average impedance is? Should it be 8 ohms, because that's the closest multiple of 8? Should it be 12 ohms, since that's the next higher multiple of 4? Or should it be closer to 10 ohms, since averaging the area under the curve up to 150Hz puts it closer to that?
Let's run the numbers and see what each of the drive voltages would work out to:
28.3v into 8 ohms dissipates 100 watts
30.0v into 9 ohms dissipates 100 watts
31.6v into 10 ohms dissipates 100 watts
34.6v into 12 ohms dissipates 100 watts
Now let's calculate the difference between each of these in decibels:
The 28.3v (8 ohm) level is 0.94x the 30.0v level, which is -0.25dB.
The 31.6v (10 ohm) level is 1.05x the 30.0v level, which is +0.23dB.
The 34.6v (12 ohm) level is 1.15x the 30.0v level, which is +0.62dB.
We're only talking about a half decibel max difference between Zmin and any sort of realistic average impedance. You actually could eye-ball an average value and be pretty close, like probably less than +/-0.25dB. But the minimum impedance method uses a tangible number that nobody has to argue about. It's what you see on the chart.
That's why I like the minimum impedance method for calculating the drive voltage used to set power levels for testing. I like using fixed voltage drive too. Seems to me a useful battery of tests is to measure at 2.83v, 1W, 10W, 100W, etc. It lets you see not only what the speaker does at low drive levels, but also at high drive levels. And it makes interpretation much easier, without having to post-process the data.
Djezzz Wayne
You took the time to write it. I didn't have the time...but who comes first, gets the premium.
Of course what impendance plot you show of what ever this unfiltered driver shows here, requires at lot of impedance notches to be straight pure perfect perfect plot acoustistically, so that static componets filtering the certain unit does have a chance to do it right. Only to say it popurlarely: The filter works as long the Voice Coil doens't move.
You present a showcase here that proff's it all.
We have only this little problem to deal with...oh boy...a massive compromise:
As soon as the voice coil starts to move in a magnetical gap the picture is different and the variables deriving from that action leads to: we have to make filters that varies simultaniously. Who in the world can solve the problem having static componets being suddenly dynamic? Following the dynamics of the TS parameters when the music is making our feet stamp of pleasures? Unsolvable.
You hit a good point here. One way is to drag AC impedance down to being VC DC impedance. This is one way to go. Like Steen Duelund would have done it. But it doesn't change the pricture, as the VC moves and the TS parameters changes anyhow.
But it doesn't work IMO as the filter itself with it's many needed notches introduces so many filter components parasites that following the natural drivers behaviour of a given driver, whatever it is, does not contribute to better sound, but oposite kills the sound naturalness build into the driver itself, if it has it at all. And it has as the statics and the nature dunamics conflicting. Completely.
For me your contrution here makes sence one way or the other. What way you choose to go will anyhow end up in a certain determnined sound.
Nice meeeting you here.
It' bed time and it's the nature of law governing.
Gerner
Impedance peaks
The impedance curve shown had a lot of peaks because it is a measurement of an undersized basshorn. We could have used Re as Zmin, probably would have in the case of direct radiators, but horns increase impedance through the passband. They're pretty reactive down low though. That's because they're acoustically small even though they're physically large, and that's what causes the impedance peaks. That particular horn is even more this way than most others.
Woofers are usually connected directly across the amplifier, so the electrical impedance peaks aren't generally a problem, in and of themselves. However, in low power SET amps it sometimes can be, since the output impedance and load impedance form a voltage divider. As you would expect, this makes a frequency sensivitive transfer function, a complex filter really.
Most tweeter horns aren't acoustically small and the mouth termination is better. This gives less response ripple and smaller peaks in the impedance curve. What remains can be damped in the crossover pretty easily, in my experience.
The impedance curve shown had a lot of peaks because it is a measurement of an undersized basshorn. We could have used Re as Zmin, probably would have in the case of direct radiators, but horns increase impedance through the passband. They're pretty reactive down low though. That's because they're acoustically small even though they're physically large, and that's what causes the impedance peaks. That particular horn is even more this way than most others.
Woofers are usually connected directly across the amplifier, so the electrical impedance peaks aren't generally a problem, in and of themselves. However, in low power SET amps it sometimes can be, since the output impedance and load impedance form a voltage divider. As you would expect, this makes a frequency sensivitive transfer function, a complex filter really.
Most tweeter horns aren't acoustically small and the mouth termination is better. This gives less response ripple and smaller peaks in the impedance curve. What remains can be damped in the crossover pretty easily, in my experience.
serenechaos said:
Robert...... far over bed time here. .
Yes this is how the Swings are made regarding the top horns and the bass section.
That a out of box BMS in combination of the the horns used is just BS, it has the potentials to shine when you start dealing with those problems. I can't call it at BMS driver anymore, even it's based on the factory choises. It's a completely differnt compresion driver used here in a dedicated horn.
If you wanted to hear it again is not scientific issues, but just to hear what the Swings is capable off!
Gerner
This is becoming rather silly. I guess we can now say e.g. ATC100ASL are in fact about 122dB / 2.83V / 1m. Even if we know that a fully passive brother is only about 89dB/2.83V/m .
Happy HOM listening anyway ! Some love it, me, well - 100+% efficiency or not, AFAIC untreated horns should stay on trucks and boats.
Happy HOM listening anyway ! Some love it, me, well - 100+% efficiency or not, AFAIC untreated horns should stay on trucks and boats.
gerner said:
Hi again John... yes and no.
If the driving amp is sensitive to AC impedance that it dances with, it's is not so.
Those amps are very sensitive to spitting out variating voltage seeing different AC impedances from the speaker load. And that is reflected in the frequency plot enourmously. They live better with a flat Ohm speaker.
So if I know what the amp is capable off, I also know what is the result of voltage efficiency. As well as th V x A outcome.
Half impedance double up in Amperes. U know = Wattage efficiency.
Cheers
Gerner
That the amp can't deliver a constant voltage across the load doesn't change the voltage sensitivity. It just changes the output level in SLP as the as the voltage varies with frequency That is, if flat band response is obtained with constant voltage, Vref, so that the SPL = xx dB/ Verf/M then if the amp if some amp can only deliver Vref/2 at some frequency and Verf at others, then at the frequency where it delivers Vref/2 the SPL will be down 6dB compared to an amp which can deliver Vref over the entire frequency band.
This has nothing to do with voltage or power sensitivity of the speaker but rather the inability of the amp to maintain a constant voltage output vs frequency.
serenechaos said:
angeloitacare said:
guys as i said before you are all talking about the bd-design system, not only about the oris swing horns.
bert believes that his system with dedicated amps driving the swings deliver the most out of that system. as far as i know bert, he prefers a completely realistic sound, devoid of any additional spices and colours and very often that can end in a sound less attractive on short runs that a bit of carefully placed colorations.
i have the swings for almost 10 months in my room and i listened a plenty of amps: altmann BYOB, various T-amps, F3, shitloads of sets (El84, 2A3, various 300b, various 10y), different gainclones and of course bert's crazy a.
truth to be told, the best overall contender is most likely the mentioned crazy A you boys heard at RMAF. it has an outstanding dynamics and excellent extension married with a transparency and complete lack of coloration.
however as my listening paradigm is a bit different, my favourite amps for swings are:
1) 10y SET A1/A2 class, direct coupled (loftin white) - almost paranormal precision, speed & transients, all veils lifted
2) 6C33 SET with power triode employed only as halves therefore low power, 6w, but extremely dynamic fast and explosive sound.
3) F3 - not as open as above but a good overall contentder with a SET-like sound without a psychodelic quality of the best sets.
at least with two from above three amps, i'm sure you will be as far from "brown" colours as possible with the swings. you would certainly lose that pompeus sound of percussions angelo mentione, but you will gain more timbral richness without sacrificing an overall sense of control too much.
the speakers are as perfect as they can be - and in terms of questions related to the efficiency, i can only add that i have no complaints about it, despite the fact i'm using an extremely low gain chain: 0.3V DAC output into TVC (0 gain) into 14db power amp. i really can't say if it is 114db or 110db but whatever it is, it's a monstruosly efficient and i;m sure that only a few speakers would be able to survive the above much. the swings not only survive but they really shine. the last buyer of them gave up on avantgarde trio which i assume is a good compliment
john k... said:
Yes yes...this is all true.... you can of course always look at a speaker by it self, but it's always connected to an amp.
My point was that is it amp *principle* depending how much SPL you get out in one meter from the Swings 6-9 Ohms load. It general for all speakers and amp. combi's.
I also touched that maybe the measuring amp should not be a T-amp. or maybe some certain non feed-back tuby.
It's very hard to seperate the amp that drives the speaker from the speaker it self, as it simply requires and amp. It's one cirquite then.
Gerner
anubisgrau said:
Anu, I also think that drowning in SPL talks and wattage contra voltage sensitivity is a bit indifferent here.
If the Swings where 110 db/w/m speakers I think I could enjoy them as much as any higher sensitivity whatever thing.
There were some confusion on how it ran up to 115 db in this thread. Inflation? VAT or other taxes?
We have been discussing quite a lot how it can come that high. Explantions were given.
We have solved the mystery on how a 100 db 15" suddenly is becoming 115 db. All the answers is there.
The only thing we cannot solve here in explanations is how they sound to one who heard them.
That be brownish, redish, blueish. Those things is imposssible to rave about as If I don't here it I cannot come with much.
For me it is 110% OK if people hear them this or that way. Are negative or positive. As long as *I'* like them is enough for me as it seems to be the case for you as well.
As Im' now soon trying out the Wavac MD805-m and Solajas Grand Excellence TVC tube pre-amp monster in short time. It will be interesting to see how the Swings respond to that.
Right now I'm using the Crazy A GainClone 30 wattage and even this is a *best buy* for money impresssive amp and you actually cannot point a finger on it's performance, maybe I want a little coloration to enter my door?
I remember very well the impressive amp. under you point 2)...
See you somewhere in town
Gerner
gedlee said:
Fact is, the Swings are a good design but you probably don't like them because there no bit of foam in the top horn.
Still can't believe you got a patent for a chunk of foam rammed down a wave guide, they'll obviously give any idea that bit of paper these days. Oh wait was I rude? Sorry, I was following your lead.
gedlee said:
Oh boy...also on this forum people quarrel a bit.
To day one can get a patent for anything whatever. That's the patents authorities bread and butter, so they are willing enough to spread pattents to God and everyman.
Can I get a pattent on my ears respond and their telling the brain ...oh..this is fantastic or oposite...well I think I can. It's my patent then as it doesn't conflict with how other people hear.
Shin and gedlee... what is a dash of foam in a horn. Something that TAD introduced many years ago, but didn't find worthwhile to patent.
If this is a DIY only forum, what can a patent do to those who just copy it? I'm just asking. Will he be sued...neeeh I don't think so.
Take the Swings as an example. They are stuffed with potential patents and did BD-Design choose to patent anything?...no, because not even one is able to copy it.
Everything outside inside that speaker of natural causes connot be copied. Bert Doppenberg protected him self so well re. the secrets in those speakers. He really doesn't need to do more...no one can get into the speakers basics.
And to try to mould some horns buy a BMS driver and a 15"'er, amp whatever, is a mission impossible if a DIY or a manefacturer want's to find what secrets are inside there.
If one likes the Swings, there is only one way to go: Buy a pair.
And even being an owner of a pair cannot help if you to make a copy of many reasons. It's a closed book.
Thanks for the attention, and let's all stay calm here.
Gerner
gedlee said:
It's all fine gedlee...
I have not studied you applicated patent in your waveguide, and for the moment I even don't know what it is about. So before commenting further I have to study you Home Page. And what is your patent is about...not judging now.
If I have surprices there ..I'll respond, but in another thread. If it's topic relevant. Otherwise I'll PM you.
At this moment i have not read one o/ooo of all the posts.
Maybe your patent has been dealth with 1000 times?
We are still walking around here in a RMAF thread. And there is a tendency that what started a thread is ending up on the other side of the world. It can start with how a certain marmelade jam is, and end up in molyculare bio chemistry....
Gerner Christensen
- Status
- Not open for further replies.