A question for those enamored with full range purity
I have a pair of full range Audio Nirvana Super 12" drivers (ferrite magnets). I adore their outstanding soundstage, coherence, and vocal presence, but am not too thrilled by the sound signature of the bass output. Furthermore, some music I listen to requires powerful sub-bass, which the drivers can never reproduce with authority. And so I was thinking of using a pair of hifi bass drivers that I had laying around. I think I will run them parallel to the woofers, crossed in around 80 Hz and biamped.
Some specs of the Audio Nirvanas. I've attached the specsheet (and VituixCAD files) for your convenience.
Nom. imp: 8 Ω
Fs: 37 Hz
Re: 7.2 Ω
Qts: 0.558
Qms: 5.977
Qes 0.615
Vas: 230,5 L (0.812 ft³)
I am looking for ways to cross the full range drivers at around 80Hz to relieve them of bass duties.
- I have tried some simple crossover networks (p.e. a simple cap in series, LC, RC) but I found that vocals are somehow affected... I suspect it's phase shift bleeding into the midrange. I don't remember which caps I used, I borrowed them from a friend. Besides, I find it difficult to model a proper crossover anyway since the drivers' impedance starts ramping up from 100 Hz and downward (up to 160 Ω at Fs! )
- I have considered an LLPXO (line-level passive crossover) right before the preamp's output resistor. This position for a crossover would dodge the impedance rise of the woofers. Though I'm afraid it would cause the same phase shift in the midrange. Haven't tried it yet though
- Acoustical roll-off... I assume the drivers will sound compressed and constrained when they are positioned in a completely stuffed small (20L) enclosure. What are your experiences with high Vas drivers in small boxes? Bad idea, no?
- Maybe acoustical roll-off by some form of purposeful open baffle phase cancellation?
I think I need some guidance in rolling off the full range drivers in a 'pure' way (i.e. midrange is not affected). Would love to hear you 2 cents on this
Stay safe, kind regards, cheers, Raoul
I have a pair of full range Audio Nirvana Super 12" drivers (ferrite magnets). I adore their outstanding soundstage, coherence, and vocal presence, but am not too thrilled by the sound signature of the bass output. Furthermore, some music I listen to requires powerful sub-bass, which the drivers can never reproduce with authority. And so I was thinking of using a pair of hifi bass drivers that I had laying around. I think I will run them parallel to the woofers, crossed in around 80 Hz and biamped.
Some specs of the Audio Nirvanas. I've attached the specsheet (and VituixCAD files) for your convenience.
Nom. imp: 8 Ω
Fs: 37 Hz
Re: 7.2 Ω
Qts: 0.558
Qms: 5.977
Qes 0.615
Vas: 230,5 L (0.812 ft³)
I am looking for ways to cross the full range drivers at around 80Hz to relieve them of bass duties.
- I have tried some simple crossover networks (p.e. a simple cap in series, LC, RC) but I found that vocals are somehow affected... I suspect it's phase shift bleeding into the midrange. I don't remember which caps I used, I borrowed them from a friend. Besides, I find it difficult to model a proper crossover anyway since the drivers' impedance starts ramping up from 100 Hz and downward (up to 160 Ω at Fs! )
- I have considered an LLPXO (line-level passive crossover) right before the preamp's output resistor. This position for a crossover would dodge the impedance rise of the woofers. Though I'm afraid it would cause the same phase shift in the midrange. Haven't tried it yet though
- Acoustical roll-off... I assume the drivers will sound compressed and constrained when they are positioned in a completely stuffed small (20L) enclosure. What are your experiences with high Vas drivers in small boxes? Bad idea, no?
- Maybe acoustical roll-off by some form of purposeful open baffle phase cancellation?
I think I need some guidance in rolling off the full range drivers in a 'pure' way (i.e. midrange is not affected). Would love to hear you 2 cents on this
Stay safe, kind regards, cheers, Raoul
Attachments
An aperiodic vented small box could work, otherwise you will have a bump in the lower frequency. And then couple to a sub without big resonances up high so you can use a lower order crossover to match the roll off. That subwoofer could be a Faital 18HF500 or a Beyma 18PWB1000Fe. Both will need a big enclosure to work low. Most hifi drivers won't really work with a passive crossover as sensivity is to low (the AN12 has a real sensivity of about 94dB). And running them in the same range is alays disturbing the sound. So for passive crossovers you need high sensitive drivers (cfr, pro audio drivers) to match the AN. If you want hifi drivers, then an active crossover and 2 amps are required, one for the AN (lower power) and one for the sub (high power to match the volume).
Can you tell what kind of cabinet you currently have with Super 12 ?
Are they connected directly to amp ?
Are you using any kind of equalisation to smooth their response ?
Which bass drivers do you have around ?
To answer your question, if you put Super 12 in a closed box with Q=0,7 then youll get acoustic rolloff equivalent to Butterworth 2nd order. But that wont relieve them from bass beacuse bass signal would still be at the terminals. And box would be fairly big too.
Idea about open baffle phase cancellation dont sound good to me beacuse below cutoff frequency, acoustic cancellation will result in a fast rise of cone excursion. And you need fairly big baffle too, to set cutoff to 80Hz.
As I understood, everything is fine except bass, so try simple and proven 1,5 way solution, connect Super 12 directly to amp, and connect the bass drivers thru series inductor to limit their response to, say 100-150Hz. That should improve the bass without affecting FR drivers performance.
Are they connected directly to amp ?
Are you using any kind of equalisation to smooth their response ?
Which bass drivers do you have around ?
To answer your question, if you put Super 12 in a closed box with Q=0,7 then youll get acoustic rolloff equivalent to Butterworth 2nd order. But that wont relieve them from bass beacuse bass signal would still be at the terminals. And box would be fairly big too.
Idea about open baffle phase cancellation dont sound good to me beacuse below cutoff frequency, acoustic cancellation will result in a fast rise of cone excursion. And you need fairly big baffle too, to set cutoff to 80Hz.
As I understood, everything is fine except bass, so try simple and proven 1,5 way solution, connect Super 12 directly to amp, and connect the bass drivers thru series inductor to limit their response to, say 100-150Hz. That should improve the bass without affecting FR drivers performance.
X-Kitz LR4 electronic crossovers. Even a little Rolls crossover is better than many give it credit for.
We do have a very small selection of active crossovers. I would love to see someone like JDS or Schiit do one.
I ran a Behringer PA crossover for years and can't say for sure it actually degraded the sound.
We do have a very small selection of active crossovers. I would love to see someone like JDS or Schiit do one.
I ran a Behringer PA crossover for years and can't say for sure it actually degraded the sound.
I know it's trendy nowadays to use active x-overs, but if you want to try passive then try a Series x-over and 1st order at that.
That will take care of the bump in the freq. response around Fs of the Audio Nirvana that you'll otherwise get.
Will also sum quite nicely the whole lot.
Very important as others suggested your woof will need to be at least 3db more sensitive than the AN, or as a compromise you can pad the AN down with a simple resistor in series.
That will take care of the bump in the freq. response around Fs of the Audio Nirvana that you'll otherwise get.
Will also sum quite nicely the whole lot.
Very important as others suggested your woof will need to be at least 3db more sensitive than the AN, or as a compromise you can pad the AN down with a simple resistor in series.
Never heard of an aperiodic vent... I googled it and it seems quite perfect. I will be looking into it. I think I will build an MDF prototype for one channel to see if I like it. Thanks
At the moment the drivers are in cabinets the previous owner built. They're 77 L (2.72 ft³) -ish BR tuned to 38 Hz, some stuffing, no crossover, connected directly to amp. EQ is completely bypassed most of the time, but it depends on the genre and mood (some music gains more from EQ in error correction than it loses in purity/coherence). I have a pair of JBL 2245H stored upstairs. I intend on making cabinets for them. These are going to be big so I hope to be able to shrink the cabinets for the Super 12s somewhat.
I want to keep ADC-DAC conversion out of the chain. For the subs, definitely, but for the full range drivers, it's a nope without room for negotation. I have tried a few electronic crossovers (reckhorn, behringer, marchand, jamo) and even though the details stayed intact, something in the soundstage and 'magic' is clearly lost
At the moment the drivers are in cabinets the previous owner built. They're 77 L (2.72 ft³) -ish BR tuned to 38 Hz, some stuffing, no crossover, connected directly to amp. EQ is completely bypassed most of the time, but it depends on the genre and mood (some music gains more from EQ in error correction than it loses in purity/coherence). I have a pair of JBL 2245H stored upstairs. I intend on making cabinets for them. These are going to be big so I hope to be able to shrink the cabinets for the Super 12s somewhat.
Thank you. So far I tried a single cap, LC, and RC. The RC had the best result crossover wise but degraded the midrange. The cap didn't sound good at all, frequency-wise. I tried to model it but I am struggling with the impedance spike at Fs
Thanks for the suggestion. I assumed that it is generally a bad thing to have two different woofers playing in the same frequency region. Especially for bass, since longer wavelengths have larger more pronounced phase differences between two woofers than shorter wavelengths do. Is this still the case here? Or can the woofers overlap in frequency coverage?
We use active crossovers as HENERYS of coils gets a bit hard to deal with. To sufficiently deal with sub Fb excursion ( and on a Maggie, critical) you need at lease second order. Do the math/simulation. I have.
Multiple woofers are fine. Even desirable. We know how to deal with phase. Your assumption is actually backwards.
Multiple woofers are fine. Even desirable. We know how to deal with phase. Your assumption is actually backwards.
I suggested to try a Series crossover exactly for that reason, to suppress the spike at Fs.
In a Series x-over an inductor is in parallel with the FR and a cap is in parallel with the woof and all of it is in series. Google "series 1st order crossover" and you'll see the configuration.
It's very simple, just a cap and a coil and the 2 drivers.
A few minor points on the above, FWIW:
- The AN specs. are sometimes a little wide of the mark. Not always, and different measurement techniques play a role here (probably) but production tolerances etc. may also come in
- In many cases, the CSA boxes aren't especially good alignments, which is probably one of the issues you're currently having with the LF response.
- The other issue you're probably having with the LF response (probably, depending on programme material, room size & SPL at the listening position) is they have a relatively short coil & coil gap. That's an observation, not a criticism BTW. They're a short throw unit that aims for a relatively (not massively) high sensitivity. So if you pump LF heavy music, especially material with significant energy in either the main power-band, or below Fb, distortion is likely to be rising quite quickly
- A passive crossover at 80Hz with them in the current box is not impossible, but will need to be relatively complex and with large components due to the LF impedance characteristic of the driver in the enclosure. Sticking to parallel for a minute, you would require a large primary series coil / coils + caps depending on preferred order and large caps & coils for LF impedance compensation. For the filter to work, especially at that sort of frequency, you'll need a series LCR Zobel shunted across the driver to flatten as a minimum the upper box impedance peak. Normally you wouldn't bother with the lower box impedance peak, but if you're targeting 80Hz, you'd probably have to do that too, with a modestly broad Q as the impedance characteristic may well change with drive level / dynamic loads (which is why T/S parameters change with drive level). Translation: you'll need at least three large components, possibly 6, just for impedance compensation, before you actually get your XO in there.
- Series 1st order in theory is simpler as it electrically speaking sums flat even with a varying impedance. This only goes so far though, and doesn't necessarily mean they will acoustically sum flat; in these conditions, they probably won't.
- Phase shift is linked to frequency response & visa versa, i.e. phase follows the FR, which should give you a better understanding of filters, however they are applied
- Realising you have woofers available, my suggestion would be that you seal up your current box as best you can, and add some more damping: not excessively, and keep from close proximity to the driver to prevent the air-load mass-loading the suspension, cone etc. Then use your woofers, assuming they are suitable, to build a pair of bass extenders / 'subwoofers', and power them with a dedicated LF amplifier with a built-in, adjustable, active / electronic crossover. Do not apply any filtering to the mains, for the time being at least. That way you may be able to preserve whatever it is you feel you need to preserve with the ANs (the issues you've had with the passive filters so far have been due to incorrect implementation -no offense, see above for the reasons / causes: without the impedance compensation, they won't be behaving as you think they are, and you'll have significant resonances & response deviations) and support them in a reasonably adjustable manner with a dedicated LF leg.
Last edited:
To be honest, Iam not sure about the first question. What I know is that distance between two drivers should be smaller than half wavelenght of cutoff frequency and they have to be aligned to reduce phase interactions. And even an active 2-way will have some issues around cutoff, beacuse there is some overlap there to. Anyway, they will be far away from vocal range in both cases.
To second question, overlap at bass is the point od 1,5-way. They work together to certain frequency (usually to 100-400Hz depending on drivers size) and than one is "disconnected" by inductor.
From my point of view, simple solution without extra electronics its either 1,5-way or series 1st order x-over as suggested by Stanislav. If you go with second, I suggest higher x-over frequency, 200-300Hz.
Note that in both cases, if you have to 8ohm drivers resulting impedance will be closer to 4ohms.
If you wish to experiment with aperiodic box, you can use existing 77 liter bass-reflex boxes. Use some sandbags, pieces of wood, pet-bottles filled with water etc. to reduce the internal volume to around 50 liter and stuff the rest of the box and bass vent with poly-wadding or similar.
That should give you fairly good feeling what you may expect.
'No', for the first, 'yes' for the second. The wavelengths in question are long enough for the drivers to acoustically sum as one unit; it's when frequency rises & wavelengths decrease that you start running into issues due to the differing distances between the individual radiating sources to the listening position & the phase-angles no longer lining up. Good job too, or all those loudspeakers that have more than one LF unit (really quite a lot of them
) wouldn't work.With all due respect Davor, two 8 ohm drivers connected in a 1st order series xo will always produce an 8 ohm speaker.
My practice tells me that and also XSim can confirm it.
1.5 way is a completely different animal.
Yes, in a textbook example with resistors.
But I remember an old Visaton project with such x-over. Drivers were 8ohm, but system impedance was specified as 4ohm, and impedance plot really dropped at some places lower than usual for 8ohm system.
Unfortunately, I cant find it anymore on their site, not even in the archive. And my memory may not be correct.
Anyway, your suggestion to make some simulations is welcomed. Perhaps you can make some with mentioned drivers. I look forward to see some.
But I remember an old Visaton project with such x-over. Drivers were 8ohm, but system impedance was specified as 4ohm, and impedance plot really dropped at some places lower than usual for 8ohm system.
Unfortunately, I cant find it anymore on their site, not even in the archive. And my memory may not be correct.
Anyway, your suggestion to make some simulations is welcomed. Perhaps you can make some with mentioned drivers. I look forward to see some.
wel, if you got measurements and a plan how to align the drivers physically i can make simulations also. I've build serial crossovers before and one of my systems still uses one (fullrange with woofer 1st order serial at 250Hz). But i need the right data to start, otherwise i can't do nothing.
I manage to find that Visaton project with series 1st order xo, from which I made the conclusion about total impedance closer to 4ohms. Its Stella Light.
Two 8ohm drivers, but nominal impedance is given as 4ohm. If you check the impedance plot, slightly under 200Hz impedance drops to less than 5ohms. So, in real life series 1st order may give lower than usual system impedance.
Stella Light
Two 8ohm drivers, but nominal impedance is given as 4ohm. If you check the impedance plot, slightly under 200Hz impedance drops to less than 5ohms. So, in real life series 1st order may give lower than usual system impedance.
Stella Light
I needed that extra bit of information, thanks. Assumed a series crossover meant 'a cap in series with the woofer'. I will give this a try, I have some parts laying around.
There's a lot I hadn't considered and I learn a lot.
@waxx Kind offer, I would be curious as to how the crossover would turn out. The new speakerboxes are going to be omnidirectional radiating boxes (or at least ~160°) with the Super 12s firing upward into a cone. Would you know how to simulate that data? Unfortunately I don't have impedance graphs on the JBLs. I don't have any measuring equipment either.
That said, the bass will be turned up or down, depending on the music genre and time of day. Is there a way to do this (that does not include op-amps or DSP) when the drivers are crossed in series?
I am looking forward to experimenting with aperiodic vents and 1,5 way configuration
Agree crossover would be higher around 150 hz.
Well geez a 18" will definitely cover the bass department !!
I dont think the enclosure needs to be to exotic.
JBL would just go into a normal reflex cabinet. And the full range will now be on a simple high pass
So it could go into a smaller well lined sealed cabinet.
Assume reducing the bass content in the full range will also slightly improve mids.
should be a fun system.
Just mount 2 speaker connectors on the 18" enclosure, and put crossover inside the bigger box.
Amplifier goes to the big box, then full range goes to second set of connectors, which are wired to crossover high pass.
I found measured response of 2245H in a JBL 4518 enclosure
and attached PDF datasheeet
Well geez a 18" will definitely cover the bass department !!
I dont think the enclosure needs to be to exotic.
JBL would just go into a normal reflex cabinet. And the full range will now be on a simple high pass
So it could go into a smaller well lined sealed cabinet.
Assume reducing the bass content in the full range will also slightly improve mids.
should be a fun system.
Just mount 2 speaker connectors on the 18" enclosure, and put crossover inside the bigger box.
Amplifier goes to the big box, then full range goes to second set of connectors, which are wired to crossover high pass.
I found measured response of 2245H in a JBL 4518 enclosure
and attached PDF datasheeet
Bad example.
Stella Light has a pretty awful xo with a lot of overlap, the FR is connected with the wrong polarity and so on, see XSim simulation below, Untitled 1.
Just a minor change makes it look decent, Untitled 2.
