They both look amazing. Looks and measurements. Do you have a favorite sound wise?
I finally gave up on the inconvenience of my previous tube rolling ways, and got a second turntable and extra headshells for rolling cartridges. And now that I have one, and two favourites for it as extras, I don't really change cartridges much anymore.
For speakers, I use a sealed bipole subwoofer and two old M&K Satellite 1B, mixing active and passive crossovers to knock down room modes with wide crossover points, and a boxed mono midbass dipole coupler made from parts of two previous open baffles.
For speakers, I use a sealed bipole subwoofer and two old M&K Satellite 1B, mixing active and passive crossovers to knock down room modes with wide crossover points, and a boxed mono midbass dipole coupler made from parts of two previous open baffles.
Attachments
Hey peeps! 
I just built a new crossover for a coaxial speaker based on Beyma 15XA38ND and it turned out very eye-catching so I thought why not to share it here. Maybe it will give dudes idea to build networks with similar idea. The owner of the speaker decided to place the crossovers on top of the speaker instead of shoving it inside the cabinet.
I just built a new crossover for a coaxial speaker based on Beyma 15XA38ND and it turned out very eye-catching so I thought why not to share it here. Maybe it will give dudes idea to build networks with similar idea. The owner of the speaker decided to place the crossovers on top of the speaker instead of shoving it inside the cabinet.
Attachments
I particularly like your DIY potting of the coils! Did you purchase magnet wire and wind them yourself?
On thing I prefer is NOT having it on a solid base - I actually prefer perforated *glass board that's had damping applied to it. Still, you do have those cool risers for the Cap's and Resistors.
*..that could be veneered with a similar wood to get that sweet finish you have.
Did you purchase magnet wire and wind them yourself?On thing I prefer is NOT having it on a solid base - I actually prefer perforated *glass board that's had damping applied to it. Still, you do have those cool risers for the Cap's and Resistors.
*..that could be veneered with a similar wood to get that sweet finish you have.
High efficiency 95dB/2.83V, 3-way, 1st order series crossover.
I began this project around 2012 armed with a mini dsp, 12" woofers in a 125L cabs using the pr170mo along the PR17 supertweeter xo'ed at 5khz. Passive xo between the mids-tweet and active woofers via mini dsp. It sounded great but i could never make the bass sing along the mid and treble.
2012:
2020:
About a year ago, i had way too much time on my hands and so this project came back to life!
A friend that used to be active here and is an "Audio Industry" personality who prefers to named as "Magnum Innominandum" helped me greatly to design this speaker. He helped to design the cross over and the entire system really. I'll quote some of his comments about the design.
Drivers:
Planar tweeter: HiVi RT2HA: https://www.parts-express.com/HiVi-RT2H-A-Planar-Isodynamic-Tweeter-297-407
Midrange: Audax 170MO: https://www.madisoundspeakerstore.com/pro-sound/audax-pr170m0-6.5-midrange-100db/
Woofers: Eminence Omega Pro 15's https://usspeaker.com/OmegaPro-15A-1.htm
Measurements:
FR measurements: Yellow line is 1.5M away from the speakers, mic at the mid level. the green line is the same speaker but now measured from the listening position.
Series xo:
The whole crossover contain no resistors to lower sensitvity, only two Inductors and two Capacitors.
The reason I always prefer series networks for 1st order crossovers, is that due the use of "shunt elements" at frequencies where the driver impedance is approximately resistive means the drivers inductance and LF resonances do not require compensation networks.
This is totally independent from all the articles all over the net where debunkers try to prove that there is no difference between series and parallel network. This case is only true for resistors, not for real drivers. A parallel type crossover would require both voice coil inductance and driver LF resonance to be compensated, making the parts count very large, especially for the midrange.
In effect, the planar Tweeter is a resistor, the midrange will need RC "Zobel" to compensate the voice coil inductance and at least one RLC "Trap" to compensate the impedance "humps" at low frequencies to make it behave like a resistor. The Bass Driver would need a RC "Zobel". Then an equivalent parallel crossover will work broadly equal. The series network does not need these compensation elements and is super simple.
Midband sensitivity is around 95dB/2.83V/1m with around 6 Ohm impedance.
final crossover pics. MIT multi cap PPFX caps used in the critical midrange. MKP type caps for the bass.
test xo:
Mid-tweet offset:
normally, a offset of the mid-tweet in relation to the woofer result in a smoother transition between bass to mid. I would have prefered to have the woofer, the mid and the planar all centered but here's the simulations:
no offset:
offset:
Sub enclosure for the mid:
closed enclosure to isolate the mid-tweet.
Denim foam installed:
Here you can see the FR when the mid have been removed. interesting to note here that the sub enclosure was not closed, hence you see a peak at around 600-800hz. that peak is gone in the final speakers once the sub enclosure is closed and foam installed. mic at the LP
Bracing is based on this super simple design: http://www.realoldiesradio.com/docs/EV/ElectroVoice_TL606.pdf
ive installed on each inner panels 2x4 glued. i havent taken any pics of the bracing. ive really went the simpler and probably compromised bracing technique by gluing 2 x4 along the back and side walls, but I hear absolutely no cabinet colorations to the sound.
How does it sound? Like the real thing. The closest system to truly create the illusion of live music ive ever heard. Mindblowing sound. Very inspirational. I'll be building a couple of versions with different drivers and cabinet arrangment along different mid driver to find my favorite combination. very very exciting stuff.
I began this project around 2012 armed with a mini dsp, 12" woofers in a 125L cabs using the pr170mo along the PR17 supertweeter xo'ed at 5khz. Passive xo between the mids-tweet and active woofers via mini dsp. It sounded great but i could never make the bass sing along the mid and treble.
2012:
2020:
About a year ago, i had way too much time on my hands and so this project came back to life!
A friend that used to be active here and is an "Audio Industry" personality who prefers to named as "Magnum Innominandum" helped me greatly to design this speaker. He helped to design the cross over and the entire system really. I'll quote some of his comments about the design.
Drivers:
Planar tweeter: HiVi RT2HA: https://www.parts-express.com/HiVi-RT2H-A-Planar-Isodynamic-Tweeter-297-407
Midrange: Audax 170MO: https://www.madisoundspeakerstore.com/pro-sound/audax-pr170m0-6.5-midrange-100db/
Woofers: Eminence Omega Pro 15's https://usspeaker.com/OmegaPro-15A-1.htm
Measurements:
FR measurements: Yellow line is 1.5M away from the speakers, mic at the mid level. the green line is the same speaker but now measured from the listening position.
Series xo:
The whole crossover contain no resistors to lower sensitvity, only two Inductors and two Capacitors.
The reason I always prefer series networks for 1st order crossovers, is that due the use of "shunt elements" at frequencies where the driver impedance is approximately resistive means the drivers inductance and LF resonances do not require compensation networks.
This is totally independent from all the articles all over the net where debunkers try to prove that there is no difference between series and parallel network. This case is only true for resistors, not for real drivers. A parallel type crossover would require both voice coil inductance and driver LF resonance to be compensated, making the parts count very large, especially for the midrange.
In effect, the planar Tweeter is a resistor, the midrange will need RC "Zobel" to compensate the voice coil inductance and at least one RLC "Trap" to compensate the impedance "humps" at low frequencies to make it behave like a resistor. The Bass Driver would need a RC "Zobel". Then an equivalent parallel crossover will work broadly equal. The series network does not need these compensation elements and is super simple.
Midband sensitivity is around 95dB/2.83V/1m with around 6 Ohm impedance.
final crossover pics. MIT multi cap PPFX caps used in the critical midrange. MKP type caps for the bass.
test xo:
Mid-tweet offset:
normally, a offset of the mid-tweet in relation to the woofer result in a smoother transition between bass to mid. I would have prefered to have the woofer, the mid and the planar all centered but here's the simulations:
no offset:
offset:
Sub enclosure for the mid:
closed enclosure to isolate the mid-tweet.
Denim foam installed:
Here you can see the FR when the mid have been removed. interesting to note here that the sub enclosure was not closed, hence you see a peak at around 600-800hz. that peak is gone in the final speakers once the sub enclosure is closed and foam installed. mic at the LP
Bracing is based on this super simple design: http://www.realoldiesradio.com/docs/EV/ElectroVoice_TL606.pdf
ive installed on each inner panels 2x4 glued. i havent taken any pics of the bracing. ive really went the simpler and probably compromised bracing technique by gluing 2 x4 along the back and side walls, but I hear absolutely no cabinet colorations to the sound.
How does it sound? Like the real thing. The closest system to truly create the illusion of live music ive ever heard. Mindblowing sound. Very inspirational. I'll be building a couple of versions with different drivers and cabinet arrangment along different mid driver to find my favorite combination. very very exciting stuff.
Last edited:
Like the idea of full sealed and no resistors. Cool design.
What are the cut offs please ? Are they assymetric or symetric xo between drivers. Curve looks good to me whatever it says nothing from the sound I surmise to be good with that drivers... I know wolfteeth here who is an experienced diy designer likes this tweeter too if my reading memory serves me well.
Do you always have ECDESIGNS stuffs ?
What are the cut offs please ? Are they assymetric or symetric xo between drivers. Curve looks good to me whatever it says nothing from the sound I surmise to be good with that drivers... I know wolfteeth here who is an experienced diy designer likes this tweeter too if my reading memory serves me well.
Do you always have ECDESIGNS stuffs ?
if i look at the graph ive posted showing the mids removed, id say xo points are about: 300hz/ 1.8khz
i dont remember if i had to inverse the mid phase. ill confirm later once i swap the audax for the phl 1120 im waiting to try here
Having a sub enclosure for the mid proved to be CRITICAL. I should show more measurements, but my first set of measurements were done without the sub enclosure closed, and I had serious peak around 600hz. adding foam and closing the sub ennclosure removed entirely the peaks in the mids.
Actually, here you can see: this was measured without the sub enclosure. look between 600 and 1.2khz response:
then look between 600 to 2khz once its closed: dismiss under 100hz, there was no ports installed.
I think a sub enclosure was probably critical here.
Yes, im using the U192 + Mosaic T ec design dac for my speaker system.
the tweeter: im in love with it. complete love. if its good enough for the Ascendo Crescendo speaker, its good enough for my DIY speaker.
https://6moons.com/audioreviews/ascendo3/4.html
Looks to me more clever on the paper at least, than the big classic 3 ways from Gravsen... while I understand his limitations due to the tigth economical model.
I made an Harbeth like, while nothing exceptional I'd love to see more sealed systems..though a little attracted by open baffle or open no baffle w/o experienced them yet. What a hobby !
I made an Harbeth like, while nothing exceptional I'd love to see more sealed systems..though a little attracted by open baffle or open no baffle w/o experienced them yet. What a hobby !
..most of the higher spl non-linear distortion below 4kHz is 2nd order.
Note that it might be an ELECTRICAL first order - but it certainly isn't going to be a resulting 1st order crossover crossed that low (..if it is crossed that low). (..and that smaller Inductor forming the high-pass for the tweeter looks fairly small.)
Last edited:
Hello Scott,
Btw the other Scott (ScottL) if I remember used also the PR170Z0 (aerogel but with a very tiny usuable bandwidth due to more pronounced break-ups and faster low end roll off)
I think the loudspeaker you refer from Loesh indeed used the PR17M0 but was a two ways. I do not remember what is the loudspeaker that was his inspiration with the same mid (Advent ?)...
That PR17MO has a Xmax of a good tweeter...0,5 mm. Which is tricky for a first order. I think Youknowyou uses his speakers in a little rrom in a near field.
Btw the other Scott (ScottL) if I remember used also the PR170Z0 (aerogel but with a very tiny usuable bandwidth due to more pronounced break-ups and faster low end roll off)
I think the loudspeaker you refer from Loesh indeed used the PR17M0 but was a two ways. I do not remember what is the loudspeaker that was his inspiration with the same mid (Advent ?)...That PR17MO has a Xmax of a good tweeter...0,5 mm. Which is tricky for a first order. I think Youknowyou uses his speakers in a little rrom in a near field.
Yeah, I think he had several.
They (my old ZO's) really only need a 1mh Inductor in series (assuming a parallel network) to provide both their low-pass filter and correction (though that's assuming a complementary baffle/baffle-step result). They do require some serious suspension "break-in" to get really good depth from them (and a very low-loss interior cabinet). Target about 96 db average.
The new version of the ZO looks a bit different to me (surround, rear cone color, and rear magnet shape). I wouldn't trust the measurements shown for it (though that's true for most drivers).
https://en.toutlehautparleur.com/speaker-audax-pr170z0-8-ohm-7-48-inch.html
The MO version is easier to work with (overall), but doesn't have the same clarity.
They (my old ZO's) really only need a 1mh Inductor in series (assuming a parallel network) to provide both their low-pass filter and correction (though that's assuming a complementary baffle/baffle-step result). They do require some serious suspension "break-in" to get really good depth from them (and a very low-loss interior cabinet). Target about 96 db average.
The new version of the ZO looks a bit different to me (surround, rear cone color, and rear magnet shape). I wouldn't trust the measurements shown for it (though that's true for most drivers).
https://en.toutlehautparleur.com/speaker-audax-pr170z0-8-ohm-7-48-inch.html
The MO version is easier to work with (overall), but doesn't have the same clarity.