Higher order passive is more difficult. If you can survive that, you still can't avoid the effect of the more elements in series with the drivers. Active crossover kinda proves that there is nothing wrong with high order filter (high group delay here) as long as the amplifier is in direct control of the voice coil (no series elements). IMO, especially when you talk about Beryllium, great speakers must have less element in series. To do this without having to use shallow slopes, notch filters are mandatory, taking into accounts driver natural roll-off. Notch filters are expensive, but it's funny if an expensive tweeter is used without any notch filter in place.
As of right now, I'm using SB26ADC tweeters with the high end rise copied from TW29R in the DSP. Sounds the same, but better (lower distortion).
I like the clean sound of the SB26, but when I see things like the BlieSMa 25mm Be dome, or the new Viawave waveguided drivers, I'm scratching my head trying to figure out if any of these will be an upgrade, subjectively and/or measured.
If anybody has heard all of these drivers please let me hear your thoughts on them, comparitively. I'm looking for the lowest distorting driver.
I like the clean sound of the SB26, but when I see things like the BlieSMa 25mm Be dome, or the new Viawave waveguided drivers, I'm scratching my head trying to figure out if any of these will be an upgrade, subjectively and/or measured.
If anybody has heard all of these drivers please let me hear your thoughts on them, comparitively. I'm looking for the lowest distorting driver.
If you look around, you can see that there are many loudspeaker designs that are using 2nd order slopes around 2kHz. For example Wilson Audio from commercial, Troels Gravesen (also a fan of Wilson, you can see
from DIY sector etc...They are using tweeters with Fs around 500-700Hz.
One thing is sure, a higher order filter is typically protecting better the tweeter but 2nd order is usually protect enough to live with. Of course it depends on the (ab)usage of the speaker and the actual construction.
Last edited:
I have heard Scan and SB BE domes in a speaker design I'm very familiar with as the designer has brought numerous incarnation of tweeters (and filters) for me to audition during the last years. I also own a pair with non-BE tweeters. They have sounded different but not in one occasion did I feel I was going to kill for BE. Other things seem to matter more.
//
//
Far as I know (and this is not related to Be) the TW29R has that rise to allow tilting of the tweeter for correct time alignment, the response basically flattens out when the tweeter is tilted back by 10 degrees that allows correct mating to the MW13 (maybe the 16, not sure) and you get prefect time alignment for LR2.
The rise in the Satori dome is far more pronounced because of breakup (I think) but again that allows for woofer on top placement with a high order crossover, perfect for the 6 and 7.5" LF drivers. I'm in the build stage of a MW19+TW29DN with woofer on top and the sims look very promising.
The Be dome is a bit different in the top octave. A bit of a head-scratcher, though I guess flat baffle and a delay of some sort, either active or passive, is the way to go. I would say stepped baffle also, but I don't like those.
The ADC is a lovely tweeter. I have those with the SB17NAC and the speaker is excellent.
Robbintip: funny you should mention the SB26ADC... I have decided to start my project with either the SB26ADC or the SB26CDC. Both seem really impressive for the modest price. I want to get comfortable with the Hypex Amp and designing biquad filters before I start working with high-cost drivers.
My thought is that I will start with SB17CAC35-8 and SB26CDC. Once I have made something really nice with those, I will move on to the new SB Textreme midwoofer and the Satori TW29 Beryllium. I suppose I am not the first person to notice that SB Acoustics has a really impressive line up of drivers.
I have been away from speaker building for 26 years, and a lot has changed, but a lot has stayed the same. My other interests are furniture building and boat building, so I am well-equipped with wood working skills and tools.
My thought is that I will start with SB17CAC35-8 and SB26CDC. Once I have made something really nice with those, I will move on to the new SB Textreme midwoofer and the Satori TW29 Beryllium. I suppose I am not the first person to notice that SB Acoustics has a really impressive line up of drivers.
I have been away from speaker building for 26 years, and a lot has changed, but a lot has stayed the same. My other interests are furniture building and boat building, so I am well-equipped with wood working skills and tools.
Last edited:
You mean 'design it competently'?
Assuming common sense and finite limits, you don't need an active crossover to prove that higher order crossovers can be useful; you can demonstrate it just as easily with passive. That said, if a design requires extremely steep filters (say, faster than LR8), then I would gently suggest there is something fundamentally wrong with it.
Why? Perhaps I'm a little old-fashioned, but I'd rather use what is needed to achieve the design goals, rather than deliberately restricting what I can do with them.
Not if the response is sufficiently linear they aren't. In most cases, they aren't, and some form of EQ is required, but that doesn't make it mandatory as exceptions exist, even if they are rare. One would hope every crossover design irrespective of order accounts for (or at least considers) the driver's natural on-baffle acoustical rolloff of course.
Not all of them. It depends on the frequency, BW and therefore the component values. Small component values = less cost. I completed a design last week with a parallel notch. Cost of the three components was £7.84 per channel; air-core inductor, non-inductive resistor & quality film cap.
Why, if none is needed?
Last edited:
It just so happens that the TW29 response almost matches the BBC loudness curve in the top end. Matter of taste maybe. Time alignment is done in DSP here.
TW29 Be defeats the point if you're asking me, because of the huge surround. It makes it a soft dome/Be hybrid practically. SB26 Be would be welcomed, but BlieSMa has some good options too, especially considering directivity.
SB26ADC is a top performer, regardless of price. If they ever launch the announced waveguided version, I will definately buy it. DXT motor structure is outdated and it's impossible to get the DXT Beryllium as found in Grimm Audio.
My next build will be MW19, MR13 and a tweeter in a waveguide that matches the DXT's directivity. MR13 and DXT are a perfect combo regarding directivity. I need 4MW19's per side to keep up with the subwoofers. Although I could do with less now I've moved into the tiniest appartment you could imagine.
Will swap the drivers for the new textreme ones, once they come out. The old MW19's will then be used in the surround speakers. I might try coating the MR13's surround inner side partially myself, like in the new Textreme ones alledgedly. So far MR13's biggest flaw is that surround. It is in a more awkward spot than the MW16's.
Regarding Hypex, good choice. So far beats the nested LM3886 amps even in the distortion department. Not so fond of the software interface, but it probably can handle my modified Linkwitsz crossovers unlike the DCX2496's that I'm using right now.
TW29 Be defeats the point if you're asking me, because of the huge surround. It makes it a soft dome/Be hybrid practically. SB26 Be would be welcomed, but BlieSMa has some good options too, especially considering directivity.
SB26ADC is a top performer, regardless of price. If they ever launch the announced waveguided version, I will definately buy it. DXT motor structure is outdated and it's impossible to get the DXT Beryllium as found in Grimm Audio.
My next build will be MW19, MR13 and a tweeter in a waveguide that matches the DXT's directivity. MR13 and DXT are a perfect combo regarding directivity. I need 4MW19's per side to keep up with the subwoofers. Although I could do with less now I've moved into the tiniest appartment you could imagine.
Will swap the drivers for the new textreme ones, once they come out. The old MW19's will then be used in the surround speakers. I might try coating the MR13's surround inner side partially myself, like in the new Textreme ones alledgedly. So far MR13's biggest flaw is that surround. It is in a more awkward spot than the MW16's.
Regarding Hypex, good choice. So far beats the nested LM3886 amps even in the distortion department. Not so fond of the software interface, but it probably can handle my modified Linkwitsz crossovers unlike the DCX2496's that I'm using right now.
In the USA Selah Audio has an attractive design called Essenza which uses two Satori WM19P-8 woofers in a 2cuft ported cabinet for -F3~43Hz -- before room gain effects.
Attachments
That looks nice indeed. Cabinet size looks the same as envisioned, even the styling is close, color-wise. There are however a lot of critical notes that come to mind. Luckily we're not limited to buying pre-made speakers. Thanks for sharing!
On a side note, but actually on topic:
I've tested coating the inside of a spare softdome tweeter, a Vifa D26tg-35-06 modified by Procus. 'Kurt Muller' coating the inside of the dome and surround made the 3rd harmonics dropped by 20dB. Total output dropped by circa 1.5dB. FR is smoothened too.
I don't know if it would be possible to DIY coat the inside of the TW29 Be style tweeters, in the first place because I remember my SB29's diaphragms are glued onto the motor, and secondly because of sudden death by Berillyum poisoning. Other option: just coat the outside if you need to, or cover up the surround with a waveguide.
On a side note, but actually on topic:
I've tested coating the inside of a spare softdome tweeter, a Vifa D26tg-35-06 modified by Procus. 'Kurt Muller' coating the inside of the dome and surround made the 3rd harmonics dropped by 20dB. Total output dropped by circa 1.5dB. FR is smoothened too.
I don't know if it would be possible to DIY coat the inside of the TW29 Be style tweeters, in the first place because I remember my SB29's diaphragms are glued onto the motor, and secondly because of sudden death by Berillyum poisoning. Other option: just coat the outside if you need to, or cover up the surround with a waveguide.
It does. Still a very good tweeter though, albeit with the trend for low high order products at the expense of rising [particularly] HD2 < 2KHz that Vifa & Scan started some years back.
I'd suggest it's the cone edge (which the surround is failing to damp properly) rather than the surround per se. You see far less of this issue with rigid cones than softer types. Damping the cone edge & inner edge of the surround can help address that of course.
A solid lump of Be isn't a real issue; beryllium dust or vapour is to be avoided though if you value a healthy life.
Most tweeters don't turn into either so no major issue there, though licking it is probably best avoided (in the same way it's best not to go around licking windows). I'm not sure coating a large roll surround will have the same effect as coating a soft dome however; ditto for coating the dome. Since the dome is unresonant until you get > 40KHz adding a coating (damping) to it isn't going to do a whole lot in or near the audio band other than reduce efficiency. Assuming there are any IM related issues from the sub-harmonics of the dome main bell mode (just assuming) they should be vanishingly low to start off with, so what is there is a product of the suspension & motor design.
Last edited:
Yes. But there's no clear boundary between competent or not. I have sensitive ears so I'm very picky.
You played save with stating LR8 passive. The correct limit is LR4, which is in 'too much' category.
Not a restriction, but an engineering decision based on a concept. If you can make passive speaker without notch AND series elements, then that is great. But we can choose between high order electrical slope without notch filter OR low order electrical slope with notch filter. After realized that LR4 is not for me, I went with minimum phase.
Notch filters dedicated to replace high order filter slope is expensive.
You may be surprised to learn this, but you aren't alone in that...
Since when has LR4 been the 'correct limit' and who made this universal determination without informing the rest of us?
You'll have a mighty struggle, since even the old RL high pass & RC low pass filters require a resistive series element.
Yes. And on occasion, it is possible to have a low order electrical slope without notch filters. Ergo, notch filters are not 'mandatory' since this means they are always required. Which is not the case. Frequently, yes, but there are exceptions. For example: Ellipticor-3
I've just opened up my design folder & am eyeballing a Chebyshev Type II derived filter I designed about 18 months ago, that tracks LR8 acoustical slopes for roughly the first octave of rolloff. Basis for both the high and low pass was a 4th order electrical ladder network, each requiring a single additional capacitive element. For the low pass, that was a 4.7uF capacitor shunting an inductor of roughly similar value to what would have been used had I been tracking a lower order slope e.g. LR4. The extra tweeter cap is larger; 100uF. Total cost per channel using quality MKPs was £24 [just under] plus delivery. Not too pricy.
Last edited:
Use cases for 8th order crossovers can be simulated too I would suppose.
Latest 8th order system I've heard are KH80DSP studio monitors. Couldn't hear the crossover at all, stereo imaging is very, very good.
They traded in system latency in order for it to be minimum phase.
You may be surprised to learn this, but you aren't alone in that...
Reading the above link, there was blind listening of a bandpass network or delay network to inverse phase. I have tried to implement such network but I couldn't make it 'sounded' right. So, how competent could it be designed so that it doesn't ruin the sound? That's why I like to avoid complex series circuit between amplifier and driver.
I'm interested with that filter (especially regarding the smooth ripple and -40dB residue) but how simple the series element could be?
The fact that you weren't able to successfully implement a passive all-pass delay network in a particular design (ditto for Troels) does not make universal applicability. There are many perfectly successful implementations of all-pass delay networks: Troels used one to good effect in another speaker, John 'Zaph' Krutke has a similar implementation in one of his designs, and many others use them also, both in DIY and commercial systems.
In terms of series components, the low pass had two inductors, one of them shunted with the aforementioned 4.7uF cap. The high pass had two capacitors and a single padding resistor. Simple.
Last edited:
The allpass network in Zaph's ZD5 (afair) makes the XT25 sounded awful. I'm also familiar with several dynaudio tweeters where such network was found in the past. Are they still using it?
Is this a universally accepted fact, or are you stating your personal opinion? If the latter, with respect, that is all it is, and not everybody shares it.
I'm also familiar with several dynaudio tweeters where such network was found in the past. Are they still using it?
I've no idea. Although I'd gently point out that you cannot be familiar 'with several Dynaudio tweeters where such network
Be that as it may, this has gone far off topic, so I suggest if you wish to present your views on passive filters, you start a dedicated thread.