RMS power is typically specified with regard to heating and burn out. Form a performance point of view what matters is excursion and linearity which is function of the applied voltage AND frequency, as the excursion will decrease roughly as 1/f^2. The Neo3 is rated at 20w RMS continuous power. SL's test is at and RMS level of 10.6 W into the 4 ohm load of the Neo3. So from an RMS power point of view the tweeter could run all day and not burn out. The peak power is 56w into 4 ohms (15V peak not 30 Vpp), at the limit of the Neo3. But with regard to the performance of the Neo3 when appropriately used the test is inconclusive and should not be taken to reflect the true performance of the tweeter. It's like expecting to get 100 dB at 30 Hz from a 5" midrange. Apply that same peak voltage swing across the tweeter at a higher frequency and there may not be a problem. The tweeter won't fit SL's needs because of his design constraint of a low crossover point.
I just used:
Looks like I put in 30 instead of 15! LOL.....half asleep....doh.....
Still get 32W. Whether it's amplitude modulated or not, the peak instantaneous current is still close to overload territory. Since most power ratings aren't based on modulated input - I don't think the power density really matters too much - overload is overload.
Looks like I put in 30 instead of 15! LOL.....half asleep....doh.....
Still get 32W. Whether it's amplitude modulated or not, the peak instantaneous current is still close to overload territory. Since most power ratings aren't based on modulated input - I don't think the power density really matters too much - overload is overload.
It seems pretty obvious Linkwitz was measuring in this case to prove a conclusion he had already reached prior to getting started. A more reasoned approach would have been to take the 20 W continuous power rating and 3.5 ohm impedance (vs 80 W and 6 ohm) into consideration before arbitrarily applying a pretty high voltage and current to both units. As we all know, power ratings are current based. We can apply all the force (voltage) we want. If little or no current is drawn as a function of resistance, you get little or no power dissipated. BG recommends using a pair of Neo 3s in a tight vertical array when running them below 2khz. As a matter of fact, Neo3's produce noticeably less distortion at the bottom of their band - so long as current and excursion limits aren't exceeded. I'd have more interest in Sig's test if he compared two Neo 3's in a small array against the SEAS driver. Given the obvious mismatch of ratings, I see very little value in this particular test.
I certainly had a hunch that the Neo3W would not work for the Orion, but I read about its apparently very successful application in several DIY dipole speaker designs. So I was curious and bought a unit. This was just in time for an upcoming test of a prototype tweeter, which I combined with a thorough evaluation of the external trigger capabilty of ARTA. That capability was not available or working properly before and became my stumbling block for switching from MLSSA to ARTA.
So my write up was about the new capability of ARTA and I used the two tweeters to illustrate what can be learned from high level testing. It was never meant to be a complete test of either tweeter, but the drive signal voltage and power level would have been realistic in my applications.
Apparently I caused some misunderstandings and concerns. So I added several sentences to clarify. Spatial distortion -
If your particular interest is increased understanding the Neo3W's performance and limitations, then I encourage you to add a burst test to your measurement suite.
SL
So my write up was about the new capability of ARTA and I used the two tweeters to illustrate what can be learned from high level testing. It was never meant to be a complete test of either tweeter, but the drive signal voltage and power level would have been realistic in my applications.
Apparently I caused some misunderstandings and concerns. So I added several sentences to clarify. Spatial distortion -
If your particular interest is increased understanding the Neo3W's performance and limitations, then I encourage you to add a burst test to your measurement suite.
SL
Thanks for speaking up, SL. That was my impression about your article, that it was more about the testing method than the particular tweeters.
I have an ARTA license as well. It would be nice if we could persuade Ivo to provide burst testing without having to use an external generator. All the pieces are pretty much in place if he could just hook them together. I think Praxis can do it now but it's 8-10x the price.
Have you tried his Burst Decay post-processing of the impulse response? He claims it's mathematically equivalent to your burst decay envelopes except it plots the whole frequency range at 1/3 or 1/6 octave resolution on a waterfall-style plot. The time axis is in periods rather than seconds. He even refers your paper in the manual. 😉
Edit: all he'd have to do is make the triggered scope so it will work work with internal signals and add the tone bursts to the available signal types -- specify frequency, number of cycles and number of repeats.
I have an ARTA license as well. It would be nice if we could persuade Ivo to provide burst testing without having to use an external generator. All the pieces are pretty much in place if he could just hook them together. I think Praxis can do it now but it's 8-10x the price.
Have you tried his Burst Decay post-processing of the impulse response? He claims it's mathematically equivalent to your burst decay envelopes except it plots the whole frequency range at 1/3 or 1/6 octave resolution on a waterfall-style plot. The time axis is in periods rather than seconds. He even refers your paper in the manual. 😉
Edit: all he'd have to do is make the triggered scope so it will work work with internal signals and add the tone bursts to the available signal types -- specify frequency, number of cycles and number of repeats.
Last edited:
Check the "real life" peak signals to your tweeter at normal listening levels. I believe that SL's modulated burst is right on the mark as far as mimicking what tweeter will receive when playing music.
Now, playing Neo3 at 1.5kHz does push it quite a lot, crossed at 3kHz it will have 4x less excursion. Typically they do get crossed somewhat lower though.
What that does to distortion levels, we really don't know (yet).
What we do know is that it is unparallel value when true dipole tweeter is concerned.
Hello,
The wavelet transform is valid only in the linear domain, since the inpulse response describes a linear system.
Using real life bursts is different method, however, since it captures nonlinear effects at the same time.
I do usually both. Wavelet transform is a bit faster method to see
everything (linear effects) at once. Real life bursts allow you to actually LISTEN how the speaker sounds, which is very very important and valuable information.
- Elias
The wavelet transform is valid only in the linear domain, since the inpulse response describes a linear system.
Using real life bursts is different method, however, since it captures nonlinear effects at the same time.
I do usually both. Wavelet transform is a bit faster method to see
everything (linear effects) at once. Real life bursts allow you to actually LISTEN how the speaker sounds, which is very very important and valuable information.
- Elias
I just found a workaround to record ARTA's internal signals. You can open two instances of ARTA, one in spectrum analyzer mode as a generator and one in external impulse mode as a triggered-scope recorder.
Thanks for clarifying Sig. As you most likely know, a lot of folks value your informed opinion very highly. And as very often is the case, excerpts of an article frequently get taken out of context and some lesser experienced folks draw conclusions about your perspective that may or may not be entirely accurate. So with your response to this thread, I think those who may have had concerns about the NEO 3's usefulness can put those concerns to rest. This is not a "compression prone" driver as some in this thread seem to have concluded from your tests. If properly used, it is one of the best tweeters available - particularly in light of its cost. While we're on the subject, folks should be reminded that BG Radia produces products primarily intended for line array or line source use - of which the BG Neo 3 is a primary example. In many respects, it is an ideal high frequency driver when used in multiples in dipole arrays.
Yes, I have tried the waterfall burst decay, but I want the capability to normalize the max amplitude to 1 at all frequencies to more clearly see stored energy. I have not been able to move Ivo to add this feature, which should be really simple. And yes, I have suggested a software burst generator to him. He has implemented quite a few of my requests, but it has taken a long time.
SL
I know this is an old thread.
May I remind all posters of the need to match DIRECTIVITY at crossover?
Polar testing - no matter the stimulus or measurement method - is just one necessity before even worrying about frequency response on axis. It looks to me that SL is doing his homework, taking in to account the need for the selected driver to withstand the inevitable long term wear and the short term occasional "mistake".
I am warming to the dipole approach. I have learned a lot from all of you!
Keep up the good work - and its documentation!
May I remind all posters of the need to match DIRECTIVITY at crossover?
Polar testing - no matter the stimulus or measurement method - is just one necessity before even worrying about frequency response on axis. It looks to me that SL is doing his homework, taking in to account the need for the selected driver to withstand the inevitable long term wear and the short term occasional "mistake".
I am warming to the dipole approach. I have learned a lot from all of you!
Keep up the good work - and its documentation!
- Status
- Not open for further replies.