or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Watts and Current
New Posts  All Forums:Forum Nav:

Watts and Current - Page 5

post #121 of 227
Nevermind. I see thats been covered. Sorry.
post #122 of 227
Quote:
Originally Posted by Heinrich S View Post

No, you need to reread the kinds of explanations that have been presented in this thread. Your comments are contradicting the other comments in this thread. You say 10 watts is 10 watts, clipping or not. Other people are suggesting 10 watts clipping can produce more than 10 watts. Are you going to deny this, or are you going to try and BS me? I simply can't believe how any reasonable person could be expected to follow any of this. mad.gif EE or not, the information presented is not clear. AT ALL.

I've read the explanations, they're very clear, as I said, you don't want to understand the basics, so you're having a really hard time understanding this simple concept.

You seem to have ignored this part of my post:
Quote:
An amplifiers power rating is ideally the amount of clean, minimally distorted, power it can supply.

Have you understood Ohms law yet?
Quote:
Are you going to deny this, or are you going to try and BS me?

Don't put words in my mouth and then get angry because you don't understand.
post #123 of 227
Quote:
Originally Posted by Heinrich S View Post

Hi guys,

Would like to know from those more experienced, what the nitty gritty is with current and wattage. I've been told that some amplifiers have lots of wattage but not much current, others have low wattage and high current. What is important in an amplifier : High wattage or high current? Thanks.
I can design an amp which will provide 100 watts into 8 ohms. To do this, it needs to be capable of 40 volts and 5 amperes output.

I can protect the output transistors with circuitry which prevents the output current from exceeding say, 5.1 amps, so if somebody shorts the output it will not blow up.

If I purchase a speaker which has an impedance of 2 ohms in the low bass region, my amplifier will only be able to provide 10 volts to that load, or 25 watts clean....because of a design choice.

The concept of "current is important" is real enough, but most amplifier designers will consider the load variation and accommodate it. In the case of the amp I "designed", I would allow the output transistors to conduct 10 amperes maximum. That would of course, require I use more expensive transistors in the amp, as they could be stressed more. But designing in twice as much current capability allows me to put rated power into a lower impedance load.

When you purchase an amp, it is important to also know what you will ask it to do.




Tweeters historically have two ways to be destroyed. Thermal, and by ultrasonic fatigue by resonance.

Thermal is simple enough. Just loan your system to a son or a friend for a party, they will turn it up amazingly loud despite the horrible sound of overload and clipping, and when you take the tweeter apart you find a charred mess.

Resonance is less understood for tweeter failures. It's like the Tacoma narrows bridge or the Kansas City Hyatt regency walkway, where small forces of wind turbulence or crowd stomping of feet-to-music matched the natural frequency of the structures. With this type of failure in tweeters, take it apart to find no burning, just a wire break where it is unsupported.

I detailed this to Selenium when I found it on their D205TI tweeters. They fixed the problem for subsequent replacement coils.

You should have felt the upper deck at Shea Stadium during the song "We didn't start the fire" during "The last play at Shea".. The tempo matched resonance, I darn near soiled my pants..
post #124 of 227
Quote:
Originally Posted by A9X-308 View Post

As for the rest of it, the amps clipping blows tweeters meme relates back a long way when tweets were rated low and were weak and can safely be ignored in practice for sensible sober adults using a system as intended who are sane enough to turn it down when it distorts enough to be objectionable.
Where's the fun in that???biggrin.gif

Quote:
Originally Posted by arnyk View Post

From what I can see, it is not unusual to run into tweeters that are pretty robust and rated at 60-100 watts. The secret is magnetic fluid that cools the voice coil by conducting heat into the magnet assembly.


Oddly enough, the D205TI from Selenium was ferrofluid cooled. I agree that dissipation is helped by the fluid a bit. The failures I experienced that were resonance were happening in the region where the vc wire exited the vc epoxy, but had yet to leave the magnetic field. The edges of the gap flux actually excited the floating wire at a right angle to the motion of the diaphragm, and it was in a region where the wire was not immersed in fluid. I would have thought viscous damping would have prevented a high Q resonance, but since the fluid was not in the fringe field volume, it didn't happen.
Quote:
Originally Posted by arnyk View Post

The magnets need not be large because only a small magnetic field may be required in a tweeter.
I believe the D205TI had a gap flux of about a tesla. But in terms of total magnetic energy, it is indeed quite a bit smaller than a typical woof.
Quote:
Originally Posted by arnyk View Post

Back in the old days tweeters might be rated at 1-3 watts and that can make them fragile. People revisited tweeter design and power ratings when big SS amps and digital became common.
Agreed. my 205's were rated 50/75 watts, but they were rated crossover freq dependent.. Still, I remember the good ol days as well..

Quote:
Originally Posted by amirm View Post

It can be hard to follow JN's arguments smile.gif.

Whuh??
Quote:
Originally Posted by amirm View Post

In addition, I made the argument that the harmonics generated as a result of clipping do add to the total power. The trick is to look at only the tweeter, and not trying to compute that percentage as the total power delivered to the speaker. The tweeter only sees the high frequencies and there, the harmonics do contribute to extra power. So as a point of math that describes what the waveforum looks like, there is extra energy generated. If you are playing the tweeter near its limit and then you pile on extra harmonics, you could push it over the edge and destroy it.

What I've not seen is a good explanation of how the clipping power is described.

If one compares a sine with peak +/-40 volts and a square wave peaks +/- 40 volts into 8 ohms, one sees 100 watts for sine and 200 for square.

But the additional 100 watts of the square is mainly odd harmonics. If for example, I heavily clip a 1 KHz program content sine, the harmonics will be 3, 5, 7....kHz. That's a lot of power being produced in the tweeter range, it's gonna burn. So much for the lead guitar solo of Zep's Heartbreaker..

In a club setting, I could run a song like Sandstorm by Darude so far into clipping during the crescendo, that the clip led's never turned off, I was pushing 600 to 700 watts of PURE distortion (the rmx1450 is rated 300 at 8 ohms) into the cabinets and yet, the song pretty much sounded the same...twas a very busy song with lots of electronic synth.. But yet, songs like Get Lucky by Pharrel/Daft Punk are so clean and bass heavy, that any clipping is extremely noticeable.

jn
post #125 of 227
Hmmm... The power series for a square wave has the harmonics going down relative to the fundamental as 1/n where n is the harmonic number. In other words, the 3rd harmonic is 1/3 the power of the fundamental, the 5th harmonic is 1/5 the power, etc. My only slight caveat would be that the signals in music or movies most likely to induce clipping are LF signals, which tend to be much larger tha HF signals, so I suspect the likelihood of clipping damaging a tweeter in practice is less than expected. However, the HF's get clipped along with the low since they are riding on the LF, so it still sounds bad...
post #126 of 227
Quote:
Originally Posted by DonH50 View Post

Hmmm... The power series for a square wave has the harmonics going down relative to the fundamental as 1/n where n is the harmonic number. In other words, the 3rd harmonic is 1/3 the power of the fundamental, the 5th harmonic is 1/5 the power, etc. My only slight caveat would be that the signals in music or movies most likely to induce clipping are LF signals, which tend to be much larger tha HF signals, so I suspect the likelihood of clipping damaging a tweeter in practice is less than expected. However, the HF's get clipped along with the low since they are riding on the LF, so it still sounds bad...

Yup. So If I crank Led Zepplin's Heartbreaker like it says to do on the album cover (turn it up waaaay past 11), almost all the harmonic power will fall into the tweeter. As you point out per my 100 watt example, that would be 33 watts of 3Khz, 20 watts of 5 KHz, 14 watts of 7 KHz, 11 watts of 9 KHz, 9 watts of 11 KHz.. So far, 87 watts into the tweeter. Note that I am speaking about the lead guitar solo...

Granted, it is extreme. My example with Darude's "Sandstorm" is more likely for me, and no, I've no way of calculating exactly what made it to the tweeter. I do note that I've never burned one. That only happens when I loan out a system...me, I subtly resonate the suckers to death by using the clip indicator as a BPM monitor.

jn
Edited by jneutron - 7/2/13 at 12:11pm
post #127 of 227
Quote:
Originally Posted by jneutron View Post

Quote:
Originally Posted by DonH50 View Post

Yup. So If I crank Led Zepplin's Heartbreaker like it says to do on the album cover (turn it up waaaay past 11), almost all the harmonic power will fall into the tweeter.

Unlikely.

Most energy in music is in the first few octaves. Even if you clip it pretty badly, the most energetic harmonics get shifted up an octave or 2, but remain in the band that is served by the woofer.
Quote:
As you point out per my 100 watt example, that would be 33 watts of 3Khz, 20 watts of 5 KHz, 14 watts of 7 KHz, 11 watts of 9 KHz, 9 watts of 11 KHz.. So far, 87 watts into the tweeter. Note that I am speaking about the lead guitar solo...

Again, unlikely.

Music isn't steady sine waves but is constantly varying multiple tones. The music could possibly peak at 100 watts, but on the average it is probably more like 10 watts.

Just for grins you ought to measure the actual voltages that are received by your speakers, and then use that information to scale the relevant frequency bands in the music.

Then you have to forget the myth that speakers are 8 ohm resistors and use the actual impedance that the speakers are in the respective frequency bands which is occasionally less than 8 ohms, but often more.
post #128 of 227
Quote:
Originally Posted by arnyk View Post

Unlikely.

Most energy in music is in the first few octaves. Even if you clip it pretty badly, the most energetic harmonics get shifted up an octave or 2, but remain in the band that is served by the woofer.

Please explain how a harmonic gets "shifted". As has been pointed out, it is odd harmonics of the fundamental.

As to remaining in the band that is served by the woofer, you are going to have to explain how a lead guitar solo, running midband content, will have harmonics which are served to the woofer.

Please re-read the example, you're not on the mark here.
Quote:
Originally Posted by arnyk View Post

Music isn't steady sine waves but is constantly varying multiple tones. The music could possibly peak at 100 watts, but on the average it is probably more like 10 watts.

You clearly have never listened to Zep... certainly not the way it was intended.

What you are speaking about is typical orchestra, jazz, chill. Programs which certainly have very high peak to rms ratios.

I'm speaking about hard core rock and roll, techno, dubstep, house... It's not all about church music..



jn
post #129 of 227
Quote:
Originally Posted by jneutron View Post


Oddly enough, the D205TI from Selenium was ferrofluid cooled. I agree that dissipation is helped by the fluid a bit.

http://www.ferrotec-europe.de/pdf/audio_ferrofluids.pdf

"Increased Thermal Power Handling: Ferrofluid is roughly 5 times more thermally conductive than the air it displaces from the gap. The fluid provides a much lower thermal resistance between the coil and pole/top plate, lowering the voice coil operating temperature under both transient and steady state conditions. This increases thermal power handling capabilities."

I would call an 5 X improvement in power dissipation capability more than "a bit". Consider an automobile with a radiator that is shrunk to 1/5.
Quote:
The failures I experienced that were resonance were happening in the region where the vc wire exited the vc epoxy, but had yet to leave the magnetic field. The edges of the gap flux actually excited the floating wire at a right angle to the motion of the diaphragm, and it was in a region where the wire was not immersed in fluid. I would have thought viscous damping would have prevented a high Q resonance, but since the fluid was not in the fringe field volume, it didn't happen.

That agrees with my experience. Since ferrofluid has become common a high proportion of the tweeter failures that I have experienced were lead wires, not cooked voice coils. Woofers generally are not fillled with ferrofluid and their voice coils can cook and fail. However the wire insulation handles high temperatures much better, which again helps quite a bit.
post #130 of 227
Quote:
Originally Posted by jneutron View Post

Quote:
Originally Posted by arnyk View Post

Unlikely.

Most energy in music is in the first few octaves. Even if you clip it pretty badly, the most energetic harmonics get shifted up an octave or 2, but remain in the band that is served by the woofer.

Please explain how a harmonic gets "shifted". As has been pointed out, it is odd harmonics of the fundamental.

You are apparently just picking at words that whose meaning should be and probably is or at least was some years ago quite clear to you JN.
Quote:
As to remaining in the band that is served by the woofer, you are going to have to explain how a lead guitar solo, running midband content, will have harmonics which are served to the woofer.

2 way speaker system with a crossover around 3 KHz. I'm sure that you have seen those JN. Perhaps that slipped from your memory as well....
Quote:
Quote:
Originally Posted by arnyk View Post

Music isn't steady sine waves but is constantly varying multiple tones. The music could possibly peak at 100 watts, but on the average it is probably more like 10 watts.

You clearly have never listened to Zep... certainly not the way it was intended.

It is possible that my audio system is much more difficult to push into clipping than yours. I decline to apologize for that. ;-)

No, it is quite clear JN that you have never seen an spectral analysis of Zep music or are suffering from loss of memory from when you did. I have done both and just lately. FWIW I was listening to Led Zepplin 4 (CD) yesterday.
post #131 of 227
Quote:
Originally Posted by arnyk View Post

http://www.ferrotec-europe.de/pdf/audio_ferrofluids.pdf

"Increased Thermal Power Handling: Ferrofluid is roughly 5 times more thermally conductive than the air it displaces from the gap. The fluid provides a much lower thermal resistance between the coil and pole/top plate, lowering the voice coil operating temperature under both transient and steady state conditions. This increases thermal power handling capabilities."

I would call an 5 X improvement in power dissipation capability more than "a bit". Consider an automobile with a radiator that is shrunk to 1/5.


Nice link. It's weird that they provide specifications for viscosity, saturation magnetization, operation and peak temperatures, but they do not provide any data whatsoever on the thermal conductivity of the material. Did you see anything on that? Since they are pushing that as a selling feature, they should provide hard data.

While stating "roughly 5 times more", actual data floats my boat.. My biggest concern would be that number is a consequence of gap turbulence in lower frequency drivers, whereas a tweeter gap has little if any turbulence nor laminar flow.
Quote:
Originally Posted by arnyk View Post

That agrees with my experience. Since ferrofluid has become common a high proportion of the tweeter failures that I have experienced were lead wires, not cooked voice coils. Woofers generally are not fillled with ferrofluid and their voice coils can cook and fail. However the wire insulation handles high temperatures much better, which again helps quite a bit.

When you examined the tweeters to find that it was the lead wires which failed, did you have the opportunity to examine the break under a metallurgical scope? You may recall I posted e-scope photo's of the classic slip plane dislocation failure consistent with high cycle fatigue caused by ultrasonic resonant failure.

jn
post #132 of 227
Quote:
Originally Posted by arnyk View Post

You are apparently just picking at words that whose meaning should be and probably is or at least was some years ago quite clear to you JN.

Shifting of harmonics is your statement, not mine. They don't shift, terrible choice of words. Recall however, that the example had 5th at 20 watts, and 7th at 14w...not chump change, especially if the tweet isn't designed to handle it.
Quote:
Originally Posted by arnyk View Post

2 way speaker system with a crossover around 3 KHz. I'm sure that you have seen those JN. Perhaps that slipped from your memory as well....
Eminence delta pro 12a, selenium D205TI, woof crossed over 3Khz first order, tweet crossed at 5K third order. Cab sized for a Q of .7 And still, twice both tweets were toasted as a result of loaning the system out. Despite the fact that they were ferrofluid cooled.
Quote:
Originally Posted by arnyk View Post

It is possible that my audio system is much more difficult to push into clipping than yours. I decline to apologize for that. ;-)
It's all in who controls the knobs. I use no compressor. I hate compression...
Quote:
Originally Posted by arnyk View Post

it is quite clear JN that you have never seen an spectral analysis of Zep music or are suffering from loss of memory from when you did. I have done both and just lately. FWIW I was listening to Led Zepplin 4 (CD) yesterday.

So, you are claiming that the lead guitar is bass???

4???? 4???

Get back to me when you listen to some real Zep.wink.gif.. I gave up on them after 3.

Zep 2, track 5. and crank it..

jn
Edited by jneutron - 7/2/13 at 1:29pm
post #133 of 227
And lest you think otherwise, I'm already aware of the typical lead guitar spectral range, as well as the harmonics. around 500, with harmonic spectra out to over 3k.

Trumpet, piano, synth (as per my Darude example), as well as female vocals have fundamentals which exceed lead g, climbing well over 1k.

Not to mention a fuzzbox..

Odd that you'd mention my old speakers as a counter. Weird indeed. I set the tweet hp to 5k 3rd order to keep the nut at the controls from toasting the tweets. And the woof lp to 3k because the driver specs showed a peak at 3k. On axis beaming of course, but half the audience is on axis..

But as anybody can tell ya, that's not how typical speakers are designed. Normally, they make the high end less robust. A sacrifice they make so that they have a flatter response curve. Not a problem for a mobile system in a nightclub where most of the listeners are drinking.

And, lots of people don't use their system as you would. Many (yours truly included) tend to crank it up at times. Mainly because the music is just begging for it.



jn
Edited by jneutron - 7/2/13 at 1:59pm
post #134 of 227
My bad, I was thinking "power series" as in the power-series expansion of a square wave and carried that word through the rest of my post. Long week. It is the voltage that decreases by 1/3, 1/5, 1/7, etc. The power naturally goes as the ratio squared, so 1/9, 1/25, etc.
post #135 of 227
I got into this discussion a little late, but here are my thoughts regarding speaker failure caused by clipping. If an amplifier is driven into clipping, it produces high levels of harmonic distortion. These harmonics are directed by the crossover network into the tweeter, which typically has a lower power rating than a woofer. For the following example, assume that the speaker presents a pure 8 Ohm resistive load. Also assume that the amplifier clips at an output voltage of 40V peak. Consider a 2 way speaker with an ideal (infinite slopes) crossover network with a crossover frequency of 2000 Hz. Drive this speaker with a 40V peak 1000 Hz sine wave, and the tweeter receives no input. The power delivered to the 8 Ohm speaker is 100 Watts, all of which is delivered to the woofer. Now drive this speaker into hard clipping so that it produces a 40V peak to peak square wave. The amplifier is now producing a 1000 Hz sine wave plus an infinite series of odd harmonics. The power delivered to the woofer under these conditions is 162 watts. The power delivered to the tweeter is 38 watts, even though it is receiving no power due to the 1000 Hz input signal! It is easy to see that with a high power amplifier driven into clipping, a tweeter could be damaged.
post #136 of 227
Quote:
Originally Posted by KralNoj View Post

I got into this discussion a little late, but here are my thoughts regarding speaker failure caused by clipping. If an amplifier is driven into clipping, it produces high levels of harmonic distortion. These harmonics are directed by the crossover network into the tweeter, which typically has a lower power rating than a woofer. For the following example, assume that the speaker presents a pure 8 Ohm resistive load. Also assume that the amplifier clips at an output voltage of 40V peak. Consider a 2 way speaker with an ideal (infinite slopes) crossover network with a crossover frequency of 2000 Hz. Drive this speaker with a 40V peak 1000 Hz sine wave, and the tweeter receives no input. The power delivered to the 8 Ohm speaker is 100 Watts, all of which is delivered to the woofer. Now drive this speaker into hard clipping so that it produces a 40V peak to peak square wave. The amplifier is now producing a 1000 Hz sine wave plus an infinite series of odd harmonics. The power delivered to the woofer under these conditions is 162 watts. The power delivered to the tweeter is 38 watts, even though it is receiving no power due to the 1000 Hz input signal! It is easy to see that with a high power amplifier driven into clipping, a tweeter could be damaged.

The third harmonic will swamp the other harmonics in terms of power (third harmonic at 1/3 the voltage of the fundamental, fifth harmnoci at 1/5 the voltage of the fundamental, etc in a square wave). in a square wave the third harmnic adds about two dB to the fundamental. But you're way over 30 percent distortion.

More to the point, if your amp can output 200 watts (forty volts into eight ohms according to the calculator I use) you simply could not drive the outptut section into a square wave at 40 volts. Just look at an amp power versus THD plot. There's lots of power available above one percent distortion. You might be at 3 or four percent distortion at 300 watts. THat's be forty eight volts. I've never seen an amp test that shows what happens above 10 percent or so distortion. You'll get higher vboltage along with the higher distortion. The only way to get a square wave at the same voltage output would be to clip the sgnal coming into the amp. Any amp I've seen tested will stay way too clean, relatively speaking, well above its rated power
post #137 of 227
Quote:
Originally Posted by KralNoj View Post

I got into this discussion a little late, but here are my thoughts regarding speaker failure caused by clipping. If an amplifier is driven into clipping, it produces high levels of harmonic distortion. These harmonics are directed by the crossover network into the tweeter, which typically has a lower power rating than a woofer. For the following example, assume that the speaker presents a pure 8 Ohm resistive load. Also assume that the amplifier clips at an output voltage of 40V peak. Consider a 2 way speaker with an ideal (infinite slopes) crossover network with a crossover frequency of 2000 Hz. Drive this speaker with a 40V peak 1000 Hz sine wave, and the tweeter receives no input.

(1) That won't happen with a regular recording of music, so you are basing any following arguments on a fallacy. Music, even solo instruments don't produce pure sine waves. Many instruments produce more energy in the first harmonic than the fundamental.

(2) Reality is that some music actually has more high frequency spectral content than a square wave. Loudspeaker manufacturers know this and started redesigning their speaker systems accordingly, stimulated by the popularization of digital audio back in the early 1980s. Prior to that they could count of the poor high frequency dynamic range of analog media partciularly the LP to limit the high frequency power that was delivered to loudspeakers.

(3) Today many speaker systems have tweeters whose power ratings are the same or greater as that of the woofers. This is facilitated by the fact that tweeters can use ferrofluid for cooling. I've previously documented that this can provide a 5x increase in tweeter power capacity.

(4) Also, there is no need for tweeter voice coils to be excessively lightweight to produce high frequencies because a heavier voice coil decreases efficiency at all frequencies above its fundamental resonance which it usually below the lowest frequency it reproduces.

4 audiophile myths in one post!
post #138 of 227
For those who fear the evil square wave, I suggest taking a gander at this video. http://video.search.yahoo.com/video/play;_ylt=A0S00MmyIdRR1ggAPMH7w8QF;_ylu=X3oDMTB2Y3ZuNzNzBHNlYwNzcgRzbGsDdmlkBHZ0aWQDVjE0NgRncG9zAzI-?p=oscilloscope+show+distortion&vid=ffedb967a8a121e9099fc4a18da9650f&l=5%3A24&turl=http%3A%2F%2Fts2.mm.bing.net%2Fth%3Fid%3DV.4722004089569593%26pid%3D15.1&rurl=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D7dWRqnJZjgA&tit=SMD+DD-1+-+OldSchoolStereo.com+Test+Bench+Update&c=1&sigr=11aponcip&age=0&fr=yfp-t-900&tt=b

It's a preamp level test, but distortion is distortion. At about 3:22 his DD1 detects distortion at one percent - - the "usual" clipping level. What's relevant to this discussion is that the sine wave on the oscilloscope looks almost unchanged. It is certainly miles and miles closer to a sine wave than a square wave. The point, for this discussion, being that at "clipping" in home theater terms, what you have is nothing at all like a swquare wave. Nothing at all like a square wave.
post #139 of 227
Even clipped by 10% it is still mostly a sine wave. http://www.whatsbestforum.com/showthread.php?8484-Clipping-101

However, considered in terms of headroom, if somebody is pushing their amp to the limit and then sends a signal just 3 dB over the max output (the real max, not max rated, but 3 dB over the clipping point) then that is twice the output power and the signal would be heavily clipped. Such peaks tend to be pretty short-lived, and as has been said most of us would not stand the sound at that point as it would be heavily distorted. Parties excepted... We have much more powerful amps (and tweeters) than decades ago when this was a significant concern. Teens and deaf old farts excepted...
post #140 of 227
Quote:
Originally Posted by arnyk View Post

(1) That won't happen with a regular recording of music, so you are basing any following arguments on a fallacy. Music, even solo instruments don't produce pure sine waves. Many instruments produce more energy in the first harmonic than the fundamental.

It is not a requirement that a solo instrument produce a pure sine wave. You have setup a strawman. In reality, an instrument's harmonics can help mask system clipping, leaving the user unaware of the problem... I pointed this out with respect to a specific song by a specific artist.

first "mythbuster" proven incorrect..
Quote:
Originally Posted by arnyk View Post

(2) Reality is that some music actually has more high frequency spectral content than a square wave. Loudspeaker manufacturers know this and started redesigning their speaker systems accordingly, stimulated by the popularization of digital audio back in the early 1980s. Prior to that they could count of the poor high frequency dynamic range of analog media partciularly the LP to limit the high frequency power that was delivered to loudspeakers.

Second "mythbuster" proven....nada. I'm not sure why you classified this as a myth arny.. It's very much true, even I agree with it..eek.gif
Quote:
Originally Posted by arnyk View Post

(3) Today many speaker systems have tweeters whose power ratings are the same or greater as that of the woofers. This is facilitated by the fact that tweeters can use ferrofluid for cooling. I've previously documented that this can provide a 5x increase in tweeter power capacity.

Actually, you are guilty here of creating a brand new, entirely unsupported "myth".
Your linked vendor provides ZERO specification nor measurement of the thermal conductivity of their ferrofluid product. They state WITHOUT ANY PRESENTED FACTS that 5x (actually, it was "up to", you've promoted it).

They also did not provide any evidence by way of actual measurements of a voice coil temperature to support this "up to 5x" statement.

In point of fact, the magnetic problem is far more complex than you can imagine. As this is within my realm of expertise, I'll elaborate (unexpected, no??cool.gif )

Introduction of ferrofluid into the gap changes the magnetic environment.

1. It reduces the path reluctance. As a consequence, high frequency fields from the vc are able to produce higher eddy currents into the gap faces. This effect will be seen in the measurement Ls/Rs, with Rs being the dominant parameter. Note that this test MUST be performed with a blocked voice coil, as movement of the coil will hugely affect the measurement of both L and R.

Higher Rs introduces additional dissipation, although it is not dissipation in the vc copper, but directly into the magnet structure.

2. It introduces assymetrical particles into the gap flux. During excitation, particles which are closest to the vc and which will orient to the field, will rotate as a result of local field distortions. This movement introduces additional dissipation directly into the ferrofluid.

3. Movement of the vc, albeit rather small, will drag the magnetic particles along for the ride, so to speak. As a consequence, the vc will cut fewer field lines during motion because the field lines will move with the particles.

The introduction of higher thermal conductivity fluid into the gap does not scale the thermal path conductivity as the ratio of it's conductivity to air. While a layman might think this, reality is far different and much more complex.

It would be better to have actual measurements of actual drivers.

As to your statement that ""many speaker systems have tweeters whose power ratings are the same or greater as that of the woofers "", can you provide any links in support of that? I wonder if it is rms dissipation being cited, or are they using a system power rating for the tweeter and an rms thermal rating for the woofer. It's always difficult figuring out tweeter ratings when the vendors mix and match ratings like that.
Quote:
Originally Posted by arnyk View Post

(4) Also, there is no need for tweeter voice coils to be excessively lightweight to produce high frequencies because a heavier voice coil decreases efficiency at all frequencies above its fundamental resonance which it usually below the lowest frequency it reproduces.

Another very true statement. I'm not sure why you lumped this into a myth category, but IIRC, some speaker vendors may have touted that low mass schtick, so I'll give you that bust...wink.gif

You never answered my question. Given a failed tweeter in your hand, did you use a metallurgical scope to examine the morphology of the failure?? To the trained eye, the distinction between IR heating failure and high cycle fatigue failure is easily spotted. But it requires a good scope.

jn
Edited by jneutron - 7/3/13 at 7:16am
post #141 of 227
Quote:
Originally Posted by DonH50 View Post

My bad, I was thinking "power series" as in the power-series expansion of a square wave and carried that word through the rest of my post. Long week. It is the voltage that decreases by 1/3, 1/5, 1/7, etc. The power naturally goes as the ratio squared, so 1/9, 1/25, etc.

Yup..which is why the numbers went to 87% with only 5 harmonics accounted for.

You need to be more careful...



Me too...frown.giftongue.gif

jn
post #142 of 227
Quote:
Originally Posted by JHAz View Post

The third harmonic will swamp the other harmonics in terms of power (third harmonic at 1/3 the voltage of the fundamental, fifth harmnoci at 1/5 the voltage of the fundamental, etc in a square wave). in a square wave the third harmnic adds about two dB to the fundamental. But you're way over 30 percent distortion.

More to the point, if your amp can output 200 watts (forty volts into eight ohms according to the calculator I use) you simply could not drive the outptut section into a square wave at 40 volts. Just look at an amp power versus THD plot. There's lots of power available above one percent distortion. You might be at 3 or four percent distortion at 300 watts. THat's be forty eight volts. I've never seen an amp test that shows what happens above 10 percent or so distortion. You'll get higher vboltage along with the higher distortion. The only way to get a square wave at the same voltage output would be to clip the sgnal coming into the amp. Any amp I've seen tested will stay way too clean, relatively speaking, well above its rated power

Your numbers do not seem consistent.

100 watts sine is 40 volts peak.
200 watts square is 40 volts peak, clipping beyond belief...

How did you get 300 watts and 48 volts?

What are you using as a distortion metric? Integrated voltage "area under the curve", or summation of harmonic power divided by fundamental power?

jn
post #143 of 227
In my previous post, I meant to say "40V peak" or "80V Peak to Peak" square wave.
post #144 of 227
It is true that hard clipping that produces a 40V peak to peak square wave is severe. It is also true that real instruments don’t produce pure sine waves. The example I gave was idealized to illustrate the point that clipping will generate harmonic distortion products that will be directed to the tweeter. As for the power calculations, a 40 volt peak sine wave represents approx. 28.2V RMS. P = V^2/R = 28.2^2/8 = 100W. I calculated the power delivered to the tweeter under clipping conditions by subtracting the power generated by the 1000 Hz component from the total power delivered to the speaker. The peak 1000 Hz component is 40 X (4/pi )=50.93Vpeak = 36V RMS V RMS This produces 36^2/8 = 162 Watts. The total power delivered by a symmetrical 80V peak to peak square wave is the same as the DC power delivered by 40VDC. This power is 40^2/8 =200W. The difference is 200-162 = 38W, all of which goes to the tweeter.
post #145 of 227
Quote:
Originally Posted by jneutron View Post

Your numbers do not seem consistent.

100 watts sine is 40 volts peak.
200 watts square is 40 volts peak, clipping beyond belief...

How did you get 300 watts and 48 volts?

What are you using as a distortion metric? Integrated voltage "area under the curve", or summation of harmonic power divided by fundamental power?

jn

there's this thing called ohm's law which is, uh, a law of physics. so in order to have more power into the same impedance you nave to have more voltage and more current. It's the law. Unforgiving. Unbreakable. And you will never ever make an amp that is clean at 40 volts clip at 40 volts, let alone produce a square waveish output. I just explained this yesterdqy in this thread with citations to web pages you can go to and actually see how it works if you want to know. Go look. Every amp has a power versus distortion curve. As power increases (at least above a certin point) distortion increases. So you can make more power at a higher distortion leve, ie with voltage (and because of that dang Ohm's law, current) increasing. Absent some kind of a limiter on the system, the amp WILL make more power than it can at say a 1 percent THD nominal clipping level. To turn a sine wave into a square wave you need over 36 percent distortion. Which is, over 36 times more distortion than the nominal clipping level at 1 percent. Look at the stuff I posted. Use your head. Read about ohm's law, too.
post #146 of 227
Quote:
Originally Posted by JHAz View Post

there's this thing called ohm's law which is, uh, a law of physics. so in order to have more power into the same impedance you nave to have more voltage and more current. It's the law. Unforgiving. Unbreakable. And you will never ever make an amp that is clean at 40 volts clip at 40 volts, let alone produce a square waveish output. I just explained this yesterdqy in this thread with citations to web pages you can go to and actually see how it works if you want to know. Go look. Every amp has a power versus distortion curve. As power increases (at least above a certin point) distortion increases. So you can make more power at a higher distortion leve, ie with voltage (and because of that dang Ohm's law, current) increasing. Absent some kind of a limiter on the system, the amp WILL make more power than it can at say a 1 percent THD nominal clipping level. To turn a sine wave into a square wave you need over 36 percent distortion. Which is, over 36 times more distortion than the nominal clipping level at 1 percent. Look at the stuff I posted. Use your head. Read about ohm's law, too.

Ohms law....no wonder...eek.gif

As I said, your numbers do not appear consistent. Rather than bash me, please go back and read your post, it is not internally self consistent. You have mixed arbitrarily peak voltages and rms voltages, so it is not possible to follow the argument you presented. That is why I asked about the 300 watt number, your writing lacked the content necessary to understand what it was you were trying to say.

It is a trivial thing to make an amp clip hard one Vce sat below the rail voltage, nowadays it's typically an RDS on value as it were.. It sometimes takes finess to make it soft. I learned that back in the early 80's when I made some power amps, and IR was only starting to make lots of hexfets, few of them P channel.

edit: It was actually more difficult forcing the output transistors out of saturation. At high slew rates, this can cause cross conduction rail to rail, goodbye outputs..

jn
Edited by jneutron - 7/3/13 at 10:30am
post #147 of 227
Quote:
Originally Posted by KralNoj View Post

It is true that hard clipping that produces a 40V peak to peak square wave is severe. It is also true that real instruments don’t produce pure sine waves. The example I gave was idealized to illustrate the point that clipping will generate harmonic distortion products that will be directed to the tweeter. As for the power calculations, a 40 volt peak sine wave represents approx. 28.2V RMS. P = V^2/R = 28.2^2/8 = 100W. I calculated the power delivered to the tweeter under clipping conditions by subtracting the power generated by the 1000 Hz component from the total power delivered to the speaker. The peak 1000 Hz component is 40 X (4/pi )=50.93Vpeak = 36V RMS V RMS This produces 36^2/8 = 162 Watts. The total power delivered by a symmetrical 80V peak to peak square wave is the same as the DC power delivered by 40VDC. This power is 40^2/8 =200W. The difference is 200-162 = 38W, all of which goes to the tweeter.

To read this plot.... at any specific frequency, the value is the sum of all harmonic power from the first harmonic generated (3Khz) to the frequency of interest. To wit, the total harmonic power delivered from 3 KHz up to 10 KHz is about 29 watts. From 3Khz to 20 KHz, it's about 33 watts. This asymptotically approaches 38 watts, as you correctly point out.

edit:my apologies, it's pretty fuzzy. frequency is hz, left is 1K, right is 1 Mhz, frequency is logarithmic scale. Vertical power is zero bottom, 40 watts top, linear scale. If you zoom in with the browser, it almost becomes clear. sigh



jn
Edited by jneutron - 7/3/13 at 12:40pm
post #148 of 227
Nice chart. Of course we have to worry about the amp's power bandwidth, loop bandwidth and stability, symmetry of the clipping (can add DC), etc. Clipping is such fun, years since I did a lot of work on it.

Sorry about the earlier mix-up; math is easy, English is hard... smile.gif
post #149 of 227
Quote:
Originally Posted by DonH50 View Post

Nice chart. Of course we have to worry about the amp's power bandwidth, loop bandwidth and stability, symmetry of the clipping (can add DC), etc. Clipping is such fun, years since I did a lot of work on it.

Sorry about the earlier mix-up; math is easy, English is hard... smile.gif

Same for me. Except for the math...that's hard too. Especially for someone like me who needs to learn ohms law..

For a 1Khz at hard clipping, the amp bandwidth isn't too important. Even if it's limited to 20 KHz, that's within 5 watts of the theoretical maximum.

I had considered normalizing both frequency and power, but decided to keep it as created, as it supports KralNoj's post.

edit: I also took it out to 50 thousand harmonics, but after 300, there really was no further contribution to power that was visible. It did bog down excel however. go figure.

jn
Edited by jneutron - 7/3/13 at 11:54am
post #150 of 227
Nice job on the chart.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › Watts and Current