And the U18 is an exceptional sounding driver too! SMOOOOOOOTH!
Wolf
But let us take a look at time domain.
Seas U18RNX
Seas L18RNX
U18 will inevitably leave a "sound signature" because of resonances in the passband. We need it to shut up, and it's doing it too slow.
Seas U18RNX
An externally hosted image should be here but it was not working when we last tested it.
Seas L18RNX
An externally hosted image should be here but it was not working when we last tested it.
U18 will inevitably leave a "sound signature" because of resonances in the passband. We need it to shut up, and it's doing it too slow.
Last edited:
Yes metal cones are impressive in the pass band, but their passband is generally strictly limited to the pistonic band.
You loose one octave and better go active with sharp xover slopes.
No free lunch, no perfect technical solution you have to choose you favorite poison...
Which is exactly why Raidho makes their loudspeakers with 115mm (4.5") midwoofers. With ceramic and diamond membranes break up is past 12KHz which means that HD3 peak will be over 4KHz, and they xover at 3KHz - where 115mm midwoofer is pistonic and almost omni.
I listened Raidho C2.1 (now C2.2) last autumn at Hifi show at Belgrade and that is the cleanest sound i heard so far out of any loudspeaker.
They claim that at break up there is only 3dB peak but i haven't seen any measurements so i don't know if it's true or not. I'd like to see it just for verification.
Last edited:
AFAIK, these resonances can be fully dealt with by using notch filters. It is my poor understanding that, being minimum phase devices, addressing the FR will address the resonance.
Having said that, I don't actually know what that means.
Best,
E
You can suppress the resonance by a notch on axis, but the notch will affect your off-axis responce in an unwanted way. The reflected energy from 30-60 grad would have a sharp dip. Besides, while the notch does deal with the cone resonance, you can still measure and hear a peak in harmonic distortion (usually 3rd order) down in the usable spectrum. This is why such drivers should be crossed very low which imposes limitations on the high frequency section in two ways. One thing to note with metal cones is the resonance is of the cone and as such, even if you suppress it by altering the electrical signal, another driver producing sound with the required frequency and energy will excite the cone of the hard coned driver under consideration. I tested the aluminium SS26W and if used as a bass driver with another unit above, let`s say a 18cm midwoofer, the midfoower would excite the cone of the SS26 no matter what you have done in the crossover. That`s why hard cones are not my gin
What unwanted way ? Do you have an example of this behaviour ?
If you use notch to take care of resonance out of the passband, how can it have measurable effect in the passband ? I use notch filters quite often and never measured anything strange off axis that resembles your description...
Or are you saying that notch filter "doesn't work" off axis ?
If there is a peak in 3rd harmonic, then that is not part of usable range - and yes, you do need high quality tweeter if we are reffering to 5" and 6" two way. It can be found starting from 50$.
...One thing to note with metal cones is the resonance is of the cone and as such, even if you suppress it by altering the electrical signal, another driver producing sound with the required frequency and energy will excite the cone of the hard coned driver under consideration. I tested the aluminium SS26W and if used as a bass driver with another unit above, let`s say a 18cm midwoofer, the midfoower would excite the cone of the SS26 no matter what you have done in the crossover. That`s why hard cones are not my gin
That could happen only if you used both drivers in the same volume, and why on earth would you put bass driver and midwoofer in the same volume and then made xover between the two ?
Last edited:
Hi,
@DogStar ... Your note about the display resolution is correct .... finer than usual ... still the curves show nothing spectacular imho ... compare for example to the old Vifa 17WN225 ... which btw had a more sensible set of TSPs and excursion.
If the resulotion weren't that fine, one could hardly discriminate between the horizontal and vertical off-axis curves.
And if I want to build something sounding exceptionally natural, musical and right and performing with orders lower THD and IMD I'd build an ESL for less
jauu
Calvin
@DogStar ... Your note about the display resolution is correct .... finer than usual ... still the curves show nothing spectacular imho ... compare for example to the old Vifa 17WN225 ... which btw had a more sensible set of TSPs and excursion.
If the resulotion weren't that fine, one could hardly discriminate between the horizontal and vertical off-axis curves.
And if I want to build something sounding exceptionally natural, musical and right and performing with orders lower THD and IMD I'd build an ESL for less
jauu
Calvin
If the drivers were measured in an anechoic chamber, the gate would have been longer. It is not about listening in anechoic chamber, but you have to know where is the problem (if it exists) so it could be treated. First measure drivers, then drivers in cabinets with suitable xover in place and then loudspeakers in room. Every step matters.
First arriving sound is one part of what matters but it's much more complicated than that.
Last edited:
+1!
Ofcourse you can mash up something and do a rough estimate. It will sound, but a. for that the drivers are way too expensive and b. why waste so much potential? If you don't care, you could use a driver that's easily 5th or even a 10th of the price. Developing speakers in this range should really be methodically done.
U18 and L18 performance depends on implementation and personal preferences.
That would be your subjective feeling you're entitled to.
cheers
You are correct. Partly. Yes, you are right, ofcourse it can be used. But you can't do it for the here mentioned resonances. Because the notch filter has to be extremely steep. With a series filter you can't achieve such steep slopes and if you do it parallel, you get an insanely low impedance at the filter frequency.
The notch would cause your offaxis response to have reduced energy in the higher angles. You can measure a hard cone driver yourself and play with a simulation software.
Not just high quality tweeter but one that can be crossed low. Many magnesium and aluminium drivers require low crossover point and that means higher order slopes which have their own contribution to overall sound. A more ellegant solution would be a waveguide but I`ve rarely seen one used with them.
Putting them both in the same enclosure would be very inadequate
Isolated one from another, the midwoofer did excite the cone with no electrical signal applied to the SS26. I have interest into studying this deeper and can post some more info once compiled. Its the same as if you had a glass cabinet in the room, let`s say behind the speaker, which you can hear resonating on the right material.
The whole point of notch filter is to reduce energy on and off axis. When i use notch at 8KHz, and xover at 1600Hz i don't measure anything wrong with my frequency response on or off axis.
When i get to my computer, i'll look into my sim for Seas L15 maybe it tells me something new when i go to extreme angles, but nothing strange showed up to 60 degrees (i usually look up to that angle if it's a direct radiator).
High quality tweeter in my book is the one that can be crossed over low, has low distortion, flat frequency response and resonance free CSD up to at least 25KHz (when metal or ceramic dome).
As for high order slopes, that is debatable if it contributes something to the sound or if it cleans it up by not allowing drivers to overlap that much and playing the stuff that they can't do well. High and low order slopes both work well when implemented well imo.
...Putting them both in the same enclosure would be very inadequate
Than my guess is that it's a matter of decoupling the drivers from the front baffle. If they were in different cabinets then that would imply that you should not use hard cone loudspeaker drivers for midrange and for bass - and that would bum out quite many high end speaker manufacturers
Please announce your results when you finish. I'm quite interested in seeing it. If it's measurable, it should be seen in CSD measurements when sweep is applied.
Last edited:
Okay, just so I get this straight and I'm not blathering on about something...
If we are referring to the L18 and its resonance breakup behavior, which I believe we are; and you are talking of a passive series notch in parallel with the driver of full LCR to compensate the said breakup of the L18; and furthermore saying it can't be done due to the resultant impedance effect brought on by said passive filter; then I still believe it can be compensated passively.
In use of an LC filter with only the inherent resistances of the L and C across the L18, the impedance will of course transition toward the 0 line without any other components present. The standard parallel xover lowpass 12dB arrangement will then counteract the dipping impedance, raising it and attenuating response in the stopband. An additional tank-cap across the LP coil will additionally suppress the breakup in tandem.
I did a simulation of the L18 and the TBFC for a friend a while back who was very pleased with the outcome. Below is the simulation results for what I call the 'Black Knights':
FR sim:
Circuit:
Later,
Wolf
A 17cm driver would beam at above 2Khz, but this would become noticeable at higher frequencies.The resonance behaviour on the amplitude axis would be different in magnitude from the off-axis. When you apply an electrical attenuation, it will be the same for all directions and will yield a different off-axis behaviour closer to an eliptical filter, compared to a low-resonant driver, like a paper or sandwich cone. I have an SP6/108Pro which I work on for a home cinema project which I will post someday here, will use the center speaker for example. 1.6Khz crossover means 4th order slopes, can you post your group delay and the decay of an impulse? How do you like the presence in this area of a 25mm dome? (yes I couldn`t hold on this
)Allowing drivers to overlap for me yields better integration, I`ve never enjoyed anything over a second order but this is very subjective and a matter to personal taste.
I will post results when I do more tests, maybe it could be coupling, don`t know yet.