No I don't, but when we started dating two years ago I negotiated good terms for my hobbies
.
They are actually basic "artificial white" (whitest of the whites), the lightning just makes them appear creamy. I think they will look quite nice when ready, considering the unescapable size.
Yes one of the goals was to make them something that one man can handle, and it was accomplished. I also like the removable side panels.
The scariest part was to lift them standing and, when needed, to make them lay again on the ground from the standing position. It was not so bad after all to make them "fall", they are somewhat down-heavy.
Thanks. They were basically designed for 50Hz and down, but I did not find their SQ too bad at higher freqs either. I will post more measurements after the finishing touches in a next couple of days. The strategic damping along the pathway and at the mouth might smoothen the response a bit. The duct could also be fully stuffed if one wishes, making it an "aperiodic tapped pipe". Then the response would be very smooth all the way to 200Hz or even beyond and the rolloff would be between 6-12dB/oct. All the resonant output would also be lost, thus the sensitivity would be quite similar to a sealed box.
The notch at 70Hz is actually smaller than simmed. I tried to get as much acoutical lowpass filtering as possible, but with two drivers in nonidentical acoutical points along the duct smoothens things out a bit making the null smaller. I simmed these only with HR, I did not try to create a more accurate Akabak script.
They are "tapped pipes", not tapped horns per se. Everything "tapped", whether the flare increases, stays the same or decreases towards the mouth, is quite the same imo. Although the decresing pathway is a mix of reflex and horn loading.
Tapped pipes (at least with these drivers I have simmed all sorts of different aligments) usually sim quite well without the typical "saddle" in the response before the tuning that many tapped horns have. Meaning propably that the tapped pipes are more damped resonators. Sensitivity is completely dominated by the actual size/litres of the flare, it does not really matter whether the flare is a "horn" or a "pipe" as long as the litres are the same.
Last edited:
Made some quick&dirty 15ply plywood back plates for the aluminum phase plugs like I did for also the BMS 18N860 in smaller THs. They hold the phase plugs in place and seal the voice coil vents that circle around the perimeter of the VC and would act as a direct leaks with "phase plugged" woofers if not sealed. Phase plug woofers always leak a little, it's the price to pay for otherwise better sound quality.
I also put on the binding posts and have been painting all the small details that needs to be painted before they can be assebled.
I also put on the binding posts and have been painting all the small details that needs to be painted before they can be assebled.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Nice progress today but full assembly has to wait until tomorrow.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Both giant's bass flutes are up and playing again. I like how they turned out visually in the end.
I also cut a good looking mouth damping pieces from 10cm/4" thick whatever that "mattress foam" stuff is called, plastic foam.
The impedance plots are very close to the simmed impedance. Now the impedance minimum lowered to 18,8Hz and 19,1Hz. Simmed is at 18,2Hz, but I think it depends just on the small variables in the drivers, which had been unused for almost 2 years before I played them again. Their suspension will soften out a bit durning the next weeks.
Impedance of the pipe with 2x18N850:
Impedance of the pipe with 2 x 18S430v2:
It seems that the room "loop back" actually changes the radiator itself, messing with it's response and impedance curve, if the radiator is a sensitive microphone like horns are. This is the reason why previously shown mouth responses all had a zigzag near the 33Hz and why also the previous impedance plots had the "twin peaks" near 30Hz. I could eliminate the twin peaks today by opening the balcony door.
Horn world is interesting.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I also cut a good looking mouth damping pieces from 10cm/4" thick whatever that "mattress foam" stuff is called, plastic foam.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The impedance plots are very close to the simmed impedance. Now the impedance minimum lowered to 18,8Hz and 19,1Hz. Simmed is at 18,2Hz, but I think it depends just on the small variables in the drivers, which had been unused for almost 2 years before I played them again. Their suspension will soften out a bit durning the next weeks.
An externally hosted image should be here but it was not working when we last tested it.
Impedance of the pipe with 2x18N850:
An externally hosted image should be here but it was not working when we last tested it.
Impedance of the pipe with 2 x 18S430v2:
An externally hosted image should be here but it was not working when we last tested it.
It seems that the room "loop back" actually changes the radiator itself, messing with it's response and impedance curve, if the radiator is a sensitive microphone like horns are. This is the reason why previously shown mouth responses all had a zigzag near the 33Hz and why also the previous impedance plots had the "twin peaks" near 30Hz. I could eliminate the twin peaks today by opening the balcony door.
Horn world is interesting.
Here are some more measurements, with the balcony door open.
Mic 10cm away from the mouth, without XO and with 40Hz But 18dB/oct filter:
Group delay (GD) and excess group delay (EGD) of the above non-filtered response:
To factor the room out even more, here's the response when the mic is 10cm inside the horn:
These measurements are also very close to what I simmed (like the impedance curves). They are quite nicely "damped" resonators with any obvious saddles in the response or high-ish Q "bump" right before or at the cutoff, it's just a smooth rolloff. Just like everytning seen before from me, also these are measured with completely uncalibrated "measurement loop". I have owned a calibrated ECM8000 once (it broke after ~3 years) and the calibration made a huge difference in infra bass levels. This also changes the GD and EGD.
Mic 10cm away from the mouth, without XO and with 40Hz But 18dB/oct filter:
An externally hosted image should be here but it was not working when we last tested it.
Group delay (GD) and excess group delay (EGD) of the above non-filtered response:
An externally hosted image should be here but it was not working when we last tested it.
To factor the room out even more, here's the response when the mic is 10cm inside the horn:
An externally hosted image should be here but it was not working when we last tested it.
These measurements are also very close to what I simmed (like the impedance curves). They are quite nicely "damped" resonators with any obvious saddles in the response or high-ish Q "bump" right before or at the cutoff, it's just a smooth rolloff. Just like everytning seen before from me, also these are measured with completely uncalibrated "measurement loop". I have owned a calibrated ECM8000 once (it broke after ~3 years) and the calibration made a huge difference in infra bass levels. This also changes the GD and EGD.
Last edited:
Boundaty loading a TH
Tried today some extra mouth damping for stronger acoustic low pass effect and also measured what a "wall loading" actually does to to a tapped horn in real life.
Mouth facing the room vs. mouth facing the wall (10cm away):
Mouth facing the wall (10cm) vs. mouth facing the wall (5cm):
The minimum impedance, which appears quite close to the tuning freq of the tapped horn where the cone movement is minimized, drops from 18Hz to 17,5Hz when the TH is wall loaded at 10cm away from the wall, and further to 16,5Hz when the distance to wall is reduced to 5cm. In addition the middle impedance peak drops from 27Hz to 25,5Hz (10cm) and again to 23,7Hz (5cm). The lowest impedance peak dropped only slightly from 12,7Hz (unloaded) to 12,4Hz (5cm).
Corner loading would drop the tuning of the TH even more.
These differences are much greater than the smulation predicts, as in real life one has to face the horns's mouth to the wall whereas simulation thinks the mouth is flush mounted to the wall/corner.
To be honest I never thought the tuning would reduce this much, but I quess it's good news for TH owners?
An externally hosted image should be here but it was not working when we last tested it.
Tried today some extra mouth damping for stronger acoustic low pass effect and also measured what a "wall loading" actually does to to a tapped horn in real life.
Mouth facing the room vs. mouth facing the wall (10cm away):
An externally hosted image should be here but it was not working when we last tested it.
Mouth facing the wall (10cm) vs. mouth facing the wall (5cm):
An externally hosted image should be here but it was not working when we last tested it.
The minimum impedance, which appears quite close to the tuning freq of the tapped horn where the cone movement is minimized, drops from 18Hz to 17,5Hz when the TH is wall loaded at 10cm away from the wall, and further to 16,5Hz when the distance to wall is reduced to 5cm. In addition the middle impedance peak drops from 27Hz to 25,5Hz (10cm) and again to 23,7Hz (5cm). The lowest impedance peak dropped only slightly from 12,7Hz (unloaded) to 12,4Hz (5cm).
Corner loading would drop the tuning of the TH even more.
These differences are much greater than the smulation predicts, as in real life one has to face the horns's mouth to the wall whereas simulation thinks the mouth is flush mounted to the wall/corner.
To be honest I never thought the tuning would reduce this much, but I quess it's good news for TH owners?
Last edited:
Corner or wall loading might be difficult to add in simulation, sort of extension of the horn pipe length and flare? There are spesific corner horns that make elementary use of walls and the floor, for bass mainly.
Corner Horn Imaging FAQ - diyAudio
Corner Horn Imaging FAQ - diyAudio
Yes one could try to add the external flare extension in the sim, it's quite easily calculated. I haven't cheked if this kind of mouth extension piece would give similar results as measured, quite likely it would after some fiddling.
-------------------------------------
The bass towers went on the front wall where they work better. No wall loading for now since I want easy access to cabling.
Low corner knee is approx. 16Hz. The crossover is 57Hz But2, which emulates the passive xo I'm going to make. No delays or EQ is used.
Unsmoothed L and R from listening spot:
Last edited:
I measured some infra bass distortion with spectrum view, which allows more accurately to see the distortion components than the distrtion plots "filtered" from a sine sweep.
Mic (uncalibrated...) is at the listening position. Whole system is reproducing the signal, both channels.
Here is the raw freq response at the mic's position:
8Hz @ 95dB @ 5,87% THD:
10Hz @ 100dB @ 3,7% THD:
14Hz @ 101dB @ 1,1% THD:
16Hz @ 102dB @ 0,51% THD:
These are by far the lowest distortion bass reproducers that I have ever owned or listened in home enviroments.
By the way, I got the "acoutical filtering sceme" to work also with these new tapped horns/pipes. They have no electrical filtering whatsoever, like the upper TH's and the 15" woofers in synergy horns. Sound like the maddest thing ever, but works like a charm.
Mic (uncalibrated...) is at the listening position. Whole system is reproducing the signal, both channels.
Here is the raw freq response at the mic's position:
An externally hosted image should be here but it was not working when we last tested it.
8Hz @ 95dB @ 5,87% THD:
An externally hosted image should be here but it was not working when we last tested it.
10Hz @ 100dB @ 3,7% THD:
An externally hosted image should be here but it was not working when we last tested it.
14Hz @ 101dB @ 1,1% THD:
An externally hosted image should be here but it was not working when we last tested it.
16Hz @ 102dB @ 0,51% THD:
An externally hosted image should be here but it was not working when we last tested it.
These are by far the lowest distortion bass reproducers that I have ever owned or listened in home enviroments.
By the way, I got the "acoutical filtering sceme" to work also with these new tapped horns/pipes. They have no electrical filtering whatsoever, like the upper TH's and the 15" woofers in synergy horns. Sound like the maddest thing ever, but works like a charm.
Here's the latest iteration of the acoutical low pass filter for the twin tower THs. Added some limp membrane in front of the mouths covering them partially.
Measured from listening spot (somewhat behind it to eliminate certain room modes).
Also made a tweak for the synergy horns. Haven't decided which configuration sounds better. Tried also a felt on the top wall, but it went too dry sounding, leaving the top off saves the airyness mostly but keeps the majority of the added resolution, which the damping gives. Quite little change in listening spot freq response, the effect comes mainly from altered (narrower/narrowing) power response and damped horn inside reflections. Tweaking possibilities are perpetual.
One should look these pictures through 3D glasses.
Measured from listening spot (somewhat behind it to eliminate certain room modes).
An externally hosted image should be here but it was not working when we last tested it.
Also made a tweak for the synergy horns. Haven't decided which configuration sounds better. Tried also a felt on the top wall, but it went too dry sounding, leaving the top off saves the airyness mostly but keeps the majority of the added resolution, which the damping gives. Quite little change in listening spot freq response, the effect comes mainly from altered (narrower/narrowing) power response and damped horn inside reflections. Tweaking possibilities are perpetual.
One should look these pictures through 3D glasses
.An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Instead of taking pictures I shot a video with Galaxy S4, enjoy! Sound quality leaves much to be desired, but it's not the point to evaluate different systems through one's own system.
https://www.youtube.com/watch?v=YqiEbD6vhPE
edit. Authentic level is approx. 90dBC at the listening spot. Inside the horn it's much larger but the camera's mic has some kind of compressor that evens the sound levels.
https://www.youtube.com/watch?v=YqiEbD6vhPE
edit. Authentic level is approx. 90dBC at the listening spot. Inside the horn it's much larger but the camera's mic has some kind of compressor that evens the sound levels.
Last edited:
Legis,
Does that 34Hz peak audibly bother you? Is that the first peak from the tapped horn? Or is it some room mode? A 1/4 wave stub at last turn in TH mouth can suppress that quite well but it is all buttoned up now. Also 1/4 wave for 34 Hz is huge length. The felt on inside of mouths makes it more stealth. But ruins the my big fast Ferrari is better than your wimpy car effect.
Does that 34Hz peak audibly bother you? Is that the first peak from the tapped horn? Or is it some room mode? A 1/4 wave stub at last turn in TH mouth can suppress that quite well but it is all buttoned up now. Also 1/4 wave for 34 Hz is huge length. The felt on inside of mouths makes it more stealth. But ruins the my big fast Ferrari is better than your wimpy car effect.
Legis,
Does that 34Hz peak audibly bother you? Is that the first peak from the tapped horn? Or is it some room mode? A 1/4 wave stub at last turn in TH mouth can suppress that quite well but it is all buttoned up now. Also 1/4 wave for 34 Hz is huge length. The felt on inside of mouths makes it more stealth. But ruins the my big fast Ferrari is better than your wimpy car effect.
That is the lowest longitudal room mode. The tapped horns have smooth response with acoustical filtering, the shown response was RAW, no electrical filtering was used.
That specific room mode is not that bad in precisely at the listening position (although other modes are worse), which is little further from the back wall than were the mic was. But yes, it bothers me with the right material that has very strong bottom end around ~35Hz. I have made resonators to suppress it but this low frequency in concrete room is hard to kill. The room is 5m x 5m, so the modes are the worst, very high Q modes. Oh well...
Some super tweeter modding.
I potted the side cavities of the JBL 2404H, which I'm using as ambience super tweeters, with 2-component potting epoxy. Made it "dead" and non-resonant. I also experimented enlarging the throats with a step drill bit, which increased the sensitivity 10-20Khz around 4-6dB. Massive high freq sensitivity for a "voice coil tweeter" now (which all have quite high Mms). The new throat is almost 1". I also made a smooth transition to the wavequide, sanded off the black paint in various places in the metal adapter and sanded the wave quide's diffraction slot's outer edges a little.
The reason why I sanded off the paint from the bottom of the adaptor is because it now makes all metallic parts in the tweeter electrically as one, as I tested it with multimeter. I can now ground the chassis (like Tannoy does) with a single wire. The inductance of the motor will also drop a tiny bit since the aluminum adaptor ("shorting ring") has now lower Re because it's connected to other metals.
They also got some metal flake candly-like red paint on them. I had polished the phasing plugs previously.
I think they are now one very fine examples of the 2404H super tweeters. Quite nice sounding at 10khz 3rd order filtered, very wide radiating top octave. If only Truextent or some other company would make a Beryllium replacement diaphragms.
I potted the side cavities of the JBL 2404H, which I'm using as ambience super tweeters, with 2-component potting epoxy. Made it "dead" and non-resonant. I also experimented enlarging the throats with a step drill bit, which increased the sensitivity 10-20Khz around 4-6dB. Massive high freq sensitivity for a "voice coil tweeter" now (which all have quite high Mms). The new throat is almost 1". I also made a smooth transition to the wavequide, sanded off the black paint in various places in the metal adapter and sanded the wave quide's diffraction slot's outer edges a little.
The reason why I sanded off the paint from the bottom of the adaptor is because it now makes all metallic parts in the tweeter electrically as one, as I tested it with multimeter. I can now ground the chassis (like Tannoy does) with a single wire. The inductance of the motor will also drop a tiny bit since the aluminum adaptor ("shorting ring") has now lower Re because it's connected to other metals.
They also got some metal flake candly-like red paint on them. I had polished the phasing plugs previously.
I think they are now one very fine examples of the 2404H super tweeters.
Quite nice sounding at 10khz 3rd order filtered, very wide radiating top octave. If only Truextent or some other company would make a Beryllium replacement diaphragms. An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Legis,Some super tweeter modding.
I also experimented enlarging the throats with a step drill bit, which increased the sensitivity 10-20Khz around 4-6dB. Massive high freq sensitivity for a "voice coil tweeter" now (which all have quite high Mms). The new throat is almost 1".
They also got some metal flake candly-like red paint on them. I had polished the phasing plugs previously.
I think they are now one very fine examples of the 2404H super tweeters.
They certainly look pretty now!
An increase of 4-6 dB on axis sensitivity would imply a 30 to 60 degree reduction of coverage.
My guess is your 2404H "Baby Butt Cheeks" may now be closer to the 2402 "Bullets" polar response. Have you measured the polar response after the mod?
Art
Hi Art! Yes they certainly are cuties
. I also believed at first that was partly the reason. However A/B listening the same driver with modded and unmodded wavequide (changin on the fly without attaching screws) showed that the modded one did output much more energy in to the room, it was just plain louder allround.I presume the diffraction slot in the wavequide is the one responsible 90% of the widening of the radiation, not the small port/throat.
What I suspect to be the main reason of the increased SPL, is that the (quite large) front chamber and the very small hole/port (which enters the wavequide) might have created an acoutical low pass filter and also have limited the SPL output by being so small. They gained most SPL at 20kHz (+6dB) and somewhat less (+4dB) at 10kHz and around 3dB below that.
I did not measure the polars response prior to modding but we can trust JBL's PDF. In after-measurements they are still very "CD". I'd say pretty close to the 2404H polars: http://www.jblpro.com/pub/obsolete/2404.pdf
I used a chair and duct tape as a mic stand, so measurement is not totally "reflection free". Measuring distance is 50cm, time window is 3ms.
Horizontal polars 0-90deg of the modded throat 2404H:
90 degree horizontal vs. 90 degree vertical:
Last edited: