Black Friday has spurred a new build. High efficiency coaxial design using with Dayton 12" PCX12-8 and SB Audience ROSSO-34CDN-PK. I will be crossing with a DSP-408.
Now my conundrum is ported or sealed? Either way I will be crossing over above the tuning frequency (have a UM-18 to handle the low end). With being above the tuning frequency does the box have less impact on the frequencies above? Is its not resonant (and out of the Q band) my thoughts are that it might not make a big difference?
Now my conundrum is ported or sealed? Either way I will be crossing over above the tuning frequency (have a UM-18 to handle the low end). With being above the tuning frequency does the box have less impact on the frequencies above? Is its not resonant (and out of the Q band) my thoughts are that it might not make a big difference?
Maybe closed would be easier to deal with. The impedance will be simpler, the rolloff won't be as steep. This is not to say that you couldn't infer some benefits from going vented... at the end of the day it may not be overly critical.
Thats what I'm leaning towards.
Recommended vented is ~59L 70Hz and Qtc just over 1 (this makes me nervous). Ported is ~65L. Seems like I could be safe just building to the ported volume and adding some ports if I really change my mind.
Recommended vented is ~59L 70Hz and Qtc just over 1 (this makes me nervous). Ported is ~65L. Seems like I could be safe just building to the ported volume and adding some ports if I really change my mind.
It's not (necessarily) going to be this after you cross it. Starting with a higher Q doesn't do irreparable damage.
If you were crossing passively you'd have to worry about the impedance of what you were doing, but with DSP that isn't an issue. A ported enclosure might see some group delay that would be easier to not have, depending on how far apart your tuning freq and XO freq are. A ported box is probably larger, I'd probably go sealed for system size and based on the price of wood these days.
If you want to do it, you could go smaller box and engineer in a higher resonance into the 12. Then compensate with a cut in the DSP. It would save you some power output from your amp, and some heat in your 12", and give you the smallest box. But the cross would be more complex, you'd probably need to do a little measurement to set it up, if you've got the capacity.
But if you wanted it to be able to do stand-alone duty away from the 18, if you wanted to give it away or sell it later, then you might want it to be ported.
If you want to do it, you could go smaller box and engineer in a higher resonance into the 12. Then compensate with a cut in the DSP. It would save you some power output from your amp, and some heat in your 12", and give you the smallest box. But the cross would be more complex, you'd probably need to do a little measurement to set it up, if you've got the capacity.
But if you wanted it to be able to do stand-alone duty away from the 18, if you wanted to give it away or sell it later, then you might want it to be ported.
After absorbing the comments and usual deliberation I've settled on a 60L sealed box. I am planning on making it well stuffed which I estimate will get the Q just under 1 and a Fsc ~75Hz. I figure this gives me sufficient room to go lower in xo for the woofer. With a 4th order filter at 80Hz I should have more than enough SPL's (>102dB) for regular listening. I have no qualms about bumping up that cutoff 30Hz for party mode.
I would say I am 99% out of the water with impedance. With the CD's being 109dB it may behoove the gain schedule from dsp to amp output to have 10dB of padding on it. If not for dynamic range in signals then possibly for noise floor of the amps.
The next big question, paint or veneer?
I would say I am 99% out of the water with impedance. With the CD's being 109dB it may behoove the gain schedule from dsp to amp output to have 10dB of padding on it. If not for dynamic range in signals then possibly for noise floor of the amps.
The next big question, paint or veneer?
The holidays were productive and was able to get these put together. First impressions are well... ...they bring the beans!!!
The box is 60L and loosely fully stuffed with polyfilll. The roll off starts well above 100Hz. Not sure if this is related to break in or not, but also not the end of the world as there's plenty of sub and a little EQing fixes it. Their performance so far is crisp and detailed. Punchy with percussion and impactful with movies.
The CD OTOH is being a bit of a challenge. This is my forst go with CD's and i presume not being a horn Im getting significant rolloff >8kHz more than the bare driver itself. To get these to sound acceptable I have a +/-12db HF/LF shelf filter at 5kHz. This is pushing the dynamic range of the DSP and has me considering giving it a first order low pass at 10-12k to help things EQ in better.
These have more directivity that I anticipated as well, at 12ft away the difference between standing and sitting is large.
Later this week I should have more time to play around. Luckily NYE celebrations have provided 12hrs of break in at normal and club level listening levels.
I would like to hear peoples opinions on how to get the 10kHz and up balanced in more. Worst case I can sit a horn on top (presuming they will gain up the HF), or maybe experimenting with an acoustic lens in front of the coax woofer (help dispersion and response?)?
The box is 60L and loosely fully stuffed with polyfilll. The roll off starts well above 100Hz. Not sure if this is related to break in or not, but also not the end of the world as there's plenty of sub and a little EQing fixes it. Their performance so far is crisp and detailed. Punchy with percussion and impactful with movies.
The CD OTOH is being a bit of a challenge. This is my forst go with CD's and i presume not being a horn Im getting significant rolloff >8kHz more than the bare driver itself. To get these to sound acceptable I have a +/-12db HF/LF shelf filter at 5kHz. This is pushing the dynamic range of the DSP and has me considering giving it a first order low pass at 10-12k to help things EQ in better.
These have more directivity that I anticipated as well, at 12ft away the difference between standing and sitting is large.
Later this week I should have more time to play around. Luckily NYE celebrations have provided 12hrs of break in at normal and club level listening levels.
I would like to hear peoples opinions on how to get the 10kHz and up balanced in more. Worst case I can sit a horn on top (presuming they will gain up the HF), or maybe experimenting with an acoustic lens in front of the coax woofer (help dispersion and response?)?
This JBL article seems to confirm that "this is normal".
https://www.google.com/url?sa=i&rct...aw1N-099MyNRO-gIBcAj-aPT&ust=1672857059405462
https://www.google.com/url?sa=i&rct...aw1N-099MyNRO-gIBcAj-aPT&ust=1672857059405462
Braces run the wrong way, there is a much better way to do it.
Rule 1: The aspect ratio of the subpanels created by a brace should be larger than the panel being braced.
dave
Braces run the wrong way, there is a much better way to do it.
Rule 1: The aspect ratio of the subpanels created by a brace should be larger than the panel being braced.
dave
Well its built now so I'd look to solutions that I can change now. Thanks for the advice tho.
The PCX12 coax has built in 80 degree conical dispersion horn. If it is actually 80 degree at high frequency, the response of the HF driver may drop anywhere from 6dB-12dB from 2kHz to 15kHz. Corrective EQ to compensate for this HF rolloff is normal for most HF drivers.
An acoustic lens could widen dispersion, but will reduce the high frequency level compared to the 80 degree horn, as well as being difficult to implement and sounding "fuzzy" .
Assuming the HF driver response is within spec, the combination of the woofer's huge on-axis peak with the compression driver's rising response in that range may make the HF response above seem weak by comparison, especially at lower listening levels, where hearing is more sensitive in the 2-4k region.
The Rosso-34CDN recommended crossover point is 2500 Hz (12dB per octave is typical on spec sheets, though not stated on theirs), but the woofer has a +12dB on axis peak around 2500 Hz, and becomes very beamy, only about 40 degree dispersion that high.
It's dispersion widens to near 90 degrees at 1600Hz, using a 24dB LR crossover at 1600 Hz and reducing the woofer peak with EQ may bring the response into balance through the crossover region, while still offering adequate LF protection for the HF driver.
Art
@weltersys thanks for pointing out that 2.5k peak adding from both sides. was a bit of a blind spot that was not entirely obvious. I have been tuning around a 1500-1700Hz xover freq. I wanted to stay away from the big on-axis hump. This is lower than the recommended by SB but as a Livingroom setup there is excessive headroom. I will experiment with a steeper slope on the woofer (currently 18), this makes intuitive sense now with keeping the 'bump' out of things.
Noted on the lens. The more I have been reading into it the more it became apparent its a not a great solution.
I was pushing the DSP a bit more in the upper ranges last night and have some better results with balanced highs, however I think I am reaching limits on how far i can stack eq gain in the dsp.
This is very much a learn as I go experimental speaker build (taking advantage of holidays for build time rather than research). "Extensive EQing" looks to be a common theme with CD's the more I read. The next experiment in this series is this filter to go between the DSP and the amp for the CD to see if I can pull in a more reasonable EQ'ing above in the last couple octaves. Its not perfect, but its the parts I had on hand today. All in the name of science after all 😉 .
Noted on the lens. The more I have been reading into it the more it became apparent its a not a great solution.
I was pushing the DSP a bit more in the upper ranges last night and have some better results with balanced highs, however I think I am reaching limits on how far i can stack eq gain in the dsp.
This is very much a learn as I go experimental speaker build (taking advantage of holidays for build time rather than research). "Extensive EQing" looks to be a common theme with CD's the more I read. The next experiment in this series is this filter to go between the DSP and the amp for the CD to see if I can pull in a more reasonable EQ'ing above in the last couple octaves. Its not perfect, but its the parts I had on hand today. All in the name of science after all 😉 .
The woofer frequency on axis sensitivity in the 2500 Hz range is +10, the compression driver mid range may be similar.
Selective parametric cuts can be used to reduce those peaks and get the acoustic crossover range correct, then the VHF shouldn't need much gain at all- it should be more than 95dB one watt one meter even in the 10kHz and up range.
If you are adjusting "by ear", you may find after EQ, the cabinet measures flat, but still sounds dull, as the highs are first to go for most people.
By the time most reach age 40, the threshold of hearing at 12kHz is reduced by more than 10dB, so will sound less than half as loud.
Having now aged past the red curve, really miss the blue years of hearing😢
Art
@weltersys are you suggesting to gap the HF crossover point well above the 1700Hz of the woofer and let a less steep slope take care of that mid 2kHz range? On the highs, once attenuated down to match sensitivities, teh SB CD measures decently flat from 5k down
On a good day I am still in that blue curve, on other days I am easy into the green territory. Seems to relate to how loud of exposures I have had in the previous week or two. I am in the early green years and like to think I have been pretty good at protecting my hearing overall. Its the odd power tools w/o protection, concerts/live venues, givin'er the beans and driving with the windows down that are my occasional degrading exposures.
On a good day I am still in that blue curve, on other days I am easy into the green territory. Seems to relate to how loud of exposures I have had in the previous week or two. I am in the early green years and like to think I have been pretty good at protecting my hearing overall. Its the odd power tools w/o protection, concerts/live venues, givin'er the beans and driving with the windows down that are my occasional degrading exposures.
No, I'm suggesting using parametric equalization (PEQ) filters to individually flatten the raw response of both the drivers in the crossover region, so when combined, the acoustic response is flat.
The PEQ to individually flatten the raw response is the inverse of the deviation from flat, a +10dB peak requires a -10dB cut.
Your woofer could use as many as 6 narrow filters in the crossover region, or a single wider one to be "close enough".
With a 24dB/octave slope at 1700Hz, you'd still want to flatten the woofer response to around an octave higher (3400Hz).
Q controls the “sharpness” of a PEQ. For a "peak" filter type, lower Q gives a broader peak or dip, while higher Q gives a narrower peak or dip.
For high or low shelf filter types, Q controls how quickly the filter transitions from no gain to maximum gain.
Your HF horn could probably use both a high and low shelf filter and a few narrow PEQ.
Different DSP use different "Q" definitions, so you will have to play around a bit to see what yours does.
If you are lucky, it's "Q" bandwidth may conform to this:
http://www.sengpielaudio.com/calculator-bandwidth.htm
Art
Last edited:
Thanks for the explanation. I am an electronics designer by profession, so I am familiar with how filters work, etc. This sounds like the key tidbit that I was missing/deviating from before. I was being somewhat reliant on the xover to not worry about the out of band peaks and just smoothing the response as a system. I will start with this approach with my next go around at tuning. Thanks!
Some great success' were made last night. But first the failure. My filters did not do much and were basically an attenuator (seen in first plots). I suspected just from experience that the 100p would be too light and the sim was too ideal.
Here is the response of the CD and the Woofer. CD was measured with a lower high pass of 1200Hz and the woofer opened up to 10K low pass. High pass on the woofer was 85Hz. These measurements were taken at ~0.5m from the driver in its listening position. I believe this measurement setup is cause of the ~7k suck out. Closer NF measurements gave odd results with the CD.
Anecdotal observation. The 12" Dayton has better HF extension than the CD????? This is odd. Also the CD HF is highly attenuated over the published specs. Not sure why.
Both Drivers were then flattened in response
Finally crossed over and level adjusted. Only time alignment, and minor tweaking around the xover freq was done.
@weltersys Thanks for the tips! I presume that my fighting of the 2.4K hump by beating down the CD EQ was causing all sorts of havoc previously! Instant impressions was that the tonality finally sounds "real" and imaging is quite real now as well. Not having all the 2.4k in the woofer also smoothed out the dispersion around the room. There's still a road of experimentation and tweaking but am now much more pleased with the product of this project.
Is there any way to get the CD to go flatter in the HF outside of aggressive filtering in DSP? Put it in a horn? Is this just a limitation/ effect of the coaxial waveguide?
Here is the response of the CD and the Woofer. CD was measured with a lower high pass of 1200Hz and the woofer opened up to 10K low pass. High pass on the woofer was 85Hz. These measurements were taken at ~0.5m from the driver in its listening position. I believe this measurement setup is cause of the ~7k suck out. Closer NF measurements gave odd results with the CD.
Anecdotal observation. The 12" Dayton has better HF extension than the CD????? This is odd. Also the CD HF is highly attenuated over the published specs. Not sure why.
Both Drivers were then flattened in response
Finally crossed over and level adjusted. Only time alignment, and minor tweaking around the xover freq was done.
@weltersys Thanks for the tips! I presume that my fighting of the 2.4K hump by beating down the CD EQ was causing all sorts of havoc previously! Instant impressions was that the tonality finally sounds "real" and imaging is quite real now as well. Not having all the 2.4k in the woofer also smoothed out the dispersion around the room. There's still a road of experimentation and tweaking but am now much more pleased with the product of this project.
Is there any way to get the CD to go flatter in the HF outside of aggressive filtering in DSP? Put it in a horn? Is this just a limitation/ effect of the coaxial waveguide?
As a rule of thumb, HF measurements need to be around three times the diameter of the horn dimension (the 12" speaker cone is the outer dimension of the horn) to get out of near-field effects.
The 12" Dayton response being around +30dB at 10kHz compared to the spec sheet seems way off, as does the near inverse on the compression driver, like the result of filters inadvertently applied during testing.
If the screw on compression driver is not properly seated to the co-ax horn entrance it could cause frequency response problems, but not at all typical of what we see in the response, other than the 7kHz null.
Verify the output of your DSP (and amp) actually have flat amplitude response, and check both compression drivers have the same raw response.
The -10dB hole in response at the combined acoustic crossover point indicates either a polarity reverse is needed, or the time alignment is off, or a combination of the two.
The conical coaxial horn may have problems, and the driver may have peaks, but probably are not the source of the problems indicated in your measurements.
Without comparisons to another horn and another driver and verification of the test signals, can't tell what the limitations or the problems are.
Cheers,
Art
I wont lie, last nights tuning was done in duress of time crunch. I will sanity check the measurements at 4ft with only a high pass on the CD. Also for HF roll off in the system. Previously I have only verified LF flatness <=20Hz.
This null shows up in both woofer and tweeter measurement. Which is what lead my assumption of measurement geometry in room.
This null shows up in both woofer and tweeter measurement. Which is what lead my assumption of measurement geometry in room.
The only comparison I have for another set of speakers in the same position and mic in the same LP position is from the bookshelf speakers that were displaced by these. The below is Dayton classic 1" silk domes. No eq'ing in those upper ranges.
The 12" woofer spec sheet also shows a deep (but smoothed to be less deep..) null in response at the same 7.5kHz frequency, the HF driver's is slightly below.
7+ kHz wavelength is under 2" long, I'd suspect horn geometry rather than room.
The bookshelf speakers upper response of +/- 6dB response looks a lot more "sane" 😉 .