Hi,
I don't know if you can get this tweeter in the US, but the Audax TW 025 A16 has one of the best low-end performance I know. Frequency response with -3dB@1kHz, first dome resonance >20kHz, slight problem around 18kHz (nothing bad, the response just drops a little), harmonic distortion <0.6%@1kHz@90dB and <0.3%@90dB above 2kHz. Very wide dispersion, therefore sounds very detailed and bright.
About the damping factor: the dynamic behaviour of two systems with identical frequency response - amplitude and phase - is also identical, regardless if there's a series resistor or not.
Bye
Baseballbat
I don't know if you can get this tweeter in the US, but the Audax TW 025 A16 has one of the best low-end performance I know. Frequency response with -3dB@1kHz, first dome resonance >20kHz, slight problem around 18kHz (nothing bad, the response just drops a little), harmonic distortion <0.6%@1kHz@90dB and <0.3%@90dB above 2kHz. Very wide dispersion, therefore sounds very detailed and bright.
About the damping factor: the dynamic behaviour of two systems with identical frequency response - amplitude and phase - is also identical, regardless if there's a series resistor or not.
Bye
Baseballbat
Hi Hartono,
you've explained before:
That's no explanation, that's an assertion.
It would've been an explanation if you've told us why it sounds different.
If you put a resistor in series to an XO and do not EQ it it _will_ sound different, because its dynamic behaviour has changed. But if you EQ to the same response as before its dynamic behaviour is identical as without a series resistor.
That does _not_ mean that it sounds identical, that is because there are some interesting secondary effects:
imagine you have a closed box with a Qtc=0.707, that is you -3dB at the resonant frequency. Then you put a resistor in series to the box that the Qtc rises to 1.0, so at the resonance frequency you have 0dB. To EQ bck to Qtc=0.7 you have to reduce the level at fs to -3dB. Nothing special about this, but the fun follows:
to get from Qtc=0.707 to Qtc=1.0 the sum of the series resistance (Rs) and the DC-resistance of the speaker (Rdc) must be 1.414*Rdc (Qms (Rms) neglected). Thus, the SPL of the speaker is -3dB below the speaker w/o Rs, but you reduced the input power by factor 4, because you reduced the input voltage level by 3dB (to get back to Qtc=0.707) and the flowing current by 3dB (because 1.414*Rdc).
In fact, you won 3dB power, because SPL decreased by 3dB, but input power decreased by 6dB.
So, what are the effects on the sound?
First, you reduced the flowing current. This leads to lower flux modulation. Second, you decreased the input power by 3dB, and even more the power at the voice coil, because the series resistor has to dissipate some of it. This reduces the power compression at high levels.
The only flaw is that you need a resistor which can handle enough power. In the design above (1.414*Rdc) it has to dissipate 1/5th of the input power.
Bye
Baseballbat
you've explained before:
Hartono said:
That's no explanation, that's an assertion.
It would've been an explanation if you've told us why it sounds different.
If you put a resistor in series to an XO and do not EQ it it _will_ sound different, because its dynamic behaviour has changed. But if you EQ to the same response as before its dynamic behaviour is identical as without a series resistor.
That does _not_ mean that it sounds identical, that is because there are some interesting secondary effects:
imagine you have a closed box with a Qtc=0.707, that is you -3dB at the resonant frequency. Then you put a resistor in series to the box that the Qtc rises to 1.0, so at the resonance frequency you have 0dB. To EQ bck to Qtc=0.7 you have to reduce the level at fs to -3dB. Nothing special about this, but the fun follows:
to get from Qtc=0.707 to Qtc=1.0 the sum of the series resistance (Rs) and the DC-resistance of the speaker (Rdc) must be 1.414*Rdc (Qms (Rms) neglected). Thus, the SPL of the speaker is -3dB below the speaker w/o Rs, but you reduced the input power by factor 4, because you reduced the input voltage level by 3dB (to get back to Qtc=0.707) and the flowing current by 3dB (because 1.414*Rdc).
In fact, you won 3dB power, because SPL decreased by 3dB, but input power decreased by 6dB.
So, what are the effects on the sound?
First, you reduced the flowing current. This leads to lower flux modulation. Second, you decreased the input power by 3dB, and even more the power at the voice coil, because the series resistor has to dissipate some of it. This reduces the power compression at high levels.
The only flaw is that you need a resistor which can handle enough power. In the design above (1.414*Rdc) it has to dissipate 1/5th of the input power.
Bye
Baseballbat
Member
Joined 2003
Hi Baseballbat and Hartono,
I truly know that both of you are knowledgeable enough to work out accurate information about this resistor issue and correct any misunderstanding between you. But could you please do this at another thread or in another way? I want people to talk about my tweeter conundrum.
Thanks,
Jay
I truly know that both of you are knowledgeable enough to work out accurate information about this resistor issue and correct any misunderstanding between you. But could you please do this at another thread or in another way? I want people to talk about my tweeter conundrum.
Thanks,
Jay
Hi Jay,
no problem, I'll never do it again.
But did you read my first post? Perhaps you can get the Audax. Owner of Audax is Harman, and it is still listed as distributor. But there was a change in their strategy not to deliver the DIY market anymore. A few former engineers of Audax decided to fill this gap and got the licenses for their techniques (like HD-A). I don't know if they sell in the US.
You can see a test of this tweeter here (870kB): http://www.audax-speaker.de/ibase/module/medienarchiv/dateien/PDF-DOWNLOAD/Kit-PRO-17-TDS-PD.pdf
It's a test of a Kit in a german magazine, scroll down to the last page.
Comment on your selection:
i would prefer the Seas, because
- it has very low distortion, even at low frequencies, the Peerless is only little better
- I like metal domes because they usually have a wider dispersion which I prefer
- it's the cheapest of the 3
The frequency response has an early roll-off at the low end, but with series resistor (here it comes again
) you can push him a little (depends on the impedance curve, whether it's with FF or not).
Unfortunately, I've never heard it, but would like to. In fact, I choose every chassis dependend on the measurements, not on my ears, because you can never trust them.
After that, you still have to audit the chassis. Tweeters are perfectly checked first with a single tone near their resonance frequency, some of them tend to "scream" even at low SPL. Another quite good test is with a two-tone, both in the desired frequency range. One at the lower end, and one at the 1.7x frequency, -3dB. This is a good stress test for tweeters, with that you can hear every ****. But be careful with the input voltage, therefor the ears are the better instruments, because you can hear when it's at its limits before any measurement equipment recognizes it. After you got that input voltage you can check the spectrum with an RTA or something and wonder how low distortion (in this case its IMD) you can here.
Bye
Baseballbat
no problem, I'll never do it again.
But did you read my first post? Perhaps you can get the Audax. Owner of Audax is Harman, and it is still listed as distributor. But there was a change in their strategy not to deliver the DIY market anymore. A few former engineers of Audax decided to fill this gap and got the licenses for their techniques (like HD-A). I don't know if they sell in the US.
You can see a test of this tweeter here (870kB): http://www.audax-speaker.de/ibase/module/medienarchiv/dateien/PDF-DOWNLOAD/Kit-PRO-17-TDS-PD.pdf
It's a test of a Kit in a german magazine, scroll down to the last page.
Comment on your selection:
i would prefer the Seas, because
- it has very low distortion, even at low frequencies, the Peerless is only little better
- I like metal domes because they usually have a wider dispersion which I prefer
- it's the cheapest of the 3
The frequency response has an early roll-off at the low end, but with series resistor (here it comes again
) you can push him a little (depends on the impedance curve, whether it's with FF or not). Unfortunately, I've never heard it, but would like to. In fact, I choose every chassis dependend on the measurements, not on my ears, because you can never trust them.
After that, you still have to audit the chassis. Tweeters are perfectly checked first with a single tone near their resonance frequency, some of them tend to "scream" even at low SPL. Another quite good test is with a two-tone, both in the desired frequency range. One at the lower end, and one at the 1.7x frequency, -3dB. This is a good stress test for tweeters, with that you can hear every ****. But be careful with the input voltage, therefor the ears are the better instruments, because you can hear when it's at its limits before any measurement equipment recognizes it. After you got that input voltage you can check the spectrum with an RTA or something and wonder how low distortion (in this case its IMD) you can here.
Bye
Baseballbat
Good advice. The Usher really has a good high frequency response with very low distortion. But there's a reason why I want to cross it at 1.5 kHz. It has a very noticeable dip at 1250 Hz, which cannot be camouflaged by a diffraction bump. I came up with a XO topology using a notch filter that can fill this dip. But this is possible only with exactly 1.5 kHz XO point. I want to be as perfect as possible with this design
Hi Jay,
looking from Zaph site, there's indeed a dip there, but when put in the box with baffle, the front baffle reflection might add this.
Edit : come to think again, maybe no, the tweeter reflection will add this, not the woofer
if looking from partsexpress datasheet, we can not see the dip ???
Hartono
looking from Zaph site, there's indeed a dip there, but when put in the box with baffle, the front baffle reflection might add this.
Edit : come to think again, maybe no, the tweeter reflection will add this, not the woofer
if looking from partsexpress datasheet, we can not see the dip ???
Hartono
I think there's something wrong with PE measurement. Another design with the Usher (by A Feyz P) definitely shows an exactly identical dip to Zaph's (see here: http://members.fortunecity.com/pirimoglu/Design/km-u2/KM-u2.html)
Baffle diffaction of either woofer or tweeter is far from being enough to hide it. I simulated it many times.
Baffle diffaction of either woofer or tweeter is far from being enough to hide it. I simulated it many times.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.