It's definitely worth a try. Here's a real picture.
An externally hosted image should be here but it was not working when we last tested it.
Some updates:
T-braces to the horn section that also have M6 thread inserts for attaching various things like mass dampers and/or hanging the super tweeters to the horns mouth.
Minimal PVA-doping on both sides of the Deltalites, adding only 0,5-1g of mass per cone. Slightly tighter sound from locking the paper fibers. Cones got faint satin sheen to them:
M6 thread inserts also to the front chamber for easy attaching/detaching of front chamber filler insert pieces:
I'm currently testing with 6pcs of hockey pucks and an 25mm thick oak disk of diameter of ~16cm. Total volume of the pieces is around 1,2 litres, resulting to ~2,4-2,5 litre front chambers when everything is accounted for. Works quite ok. I get (hopefully working) measurement equipment later this week and can see how things measure.
T-braces to the horn section that also have M6 thread inserts for attaching various things like mass dampers and/or hanging the super tweeters to the horns mouth.
An externally hosted image should be here but it was not working when we last tested it.
Minimal PVA-doping on both sides of the Deltalites, adding only 0,5-1g of mass per cone. Slightly tighter sound from locking the paper fibers. Cones got faint satin sheen to them:
An externally hosted image should be here but it was not working when we last tested it.
M6 thread inserts also to the front chamber for easy attaching/detaching of front chamber filler insert pieces:
An externally hosted image should be here but it was not working when we last tested it.
I'm currently testing with 6pcs of hockey pucks and an 25mm thick oak disk of diameter of ~16cm. Total volume of the pieces is around 1,2 litres, resulting to ~2,4-2,5 litre front chambers when everything is accounted for. Works quite ok. I get (hopefully working) measurement equipment later this week and can see how things measure.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Nice work Legis. I like the use of available materials - I assume you or your kid plays hockey? You should check out some new work by cookiemonster on the effect of offset vs centered ports on the bass/mid driver chamber. Offset seems to cause a full octave drop in bandwidth.
http://www.diyaudio.com/forums/multi-way/267083-synergy-horn-3d-printing-entry-6.html#post4183445
Based on this, I would try a simple phase plug using stacked sheets of XPS foam glued together that matches the cone profile with clearance for driver movement. I would make the a hole starting in the middle where the dustcap is and angle it to the side to line up with the actual port location. This should help and the foam acts as a bit of damping material.
http://www.diyaudio.com/forums/multi-way/267083-synergy-horn-3d-printing-entry-6.html#post4183445
Based on this, I would try a simple phase plug using stacked sheets of XPS foam glued together that matches the cone profile with clearance for driver movement. I would make the a hole starting in the middle where the dustcap is and angle it to the side to line up with the actual port location. This should help and the foam acts as a bit of damping material.
Last edited:
I actually bought the pucks for this
. I wanted something that also damps the horn. They are quite good also under speaker's feet/spikes, reducing some of the vibration conduction to the floor. I tried also "casting" a filler piece using Deltalites own cone by using sealing urethane foam and "Saran wrap", but is was not really a success.Yes center port would be ideal and possible for small driver & low xo point -combination (but I doubt smaller cones would be as dynamic as bigger ones). The whole system with a bending cone (at least paper is not very pistonic at HF like the sim might predict) and the possible front chamber filling piece blocking the radiation from back of the chamber to front etc. makes things quite hard to predict with an offset port.
I found out that these filling pieces block some of the HF radiation and resonances compared to empty chamber and I propably don't need any damping material inside the chamber. The oak disk also blocks the direct view from the dust cap to injection port.
Last edited:
I could try that. But I wonder what would it do to the response if the injection port would be "carved" to the filling piece like that. Would it lenghten the effective injection tube lenght which might cause trouble? HR sims show best results with 2-3cm thick tube.
Got the new measurement gear today. Here is the measured response from the horn's mouth with the mentioned filler piece in the front chamber (the black trace). It reduces the 250-450hz "hill", makes the acoustic low pass somewhat lower order (unlike simulated, simulation predicted shaper cutoff) and also the nulls are not as pronounced which is good. Bass goes deeped but woofer has gotten many hundreds of hours of burn in between the measurements because the other trace is an old measurement.
Here's the response from the listening spot with the new equipment. The new mic is uncalibrated Superluc EMC-999.
Here's the filtering setup. I have always aimed for low order filters and minimal EQ'ing.
2 x BMS 18N860 - Tapped horns:
2 x Deltalite 2515 - Synergy midwoofers (no signal correction at all)
JBL 2446J + Truextent - Synergy comp driver
2 x JBL 2404H - supertweeters (mounted currently inside the synergy horns, one facing front and other facing backwards in same phase, works nicely)
An externally hosted image should be here but it was not working when we last tested it.
Here's the response from the listening spot with the new equipment. The new mic is uncalibrated Superluc EMC-999.
An externally hosted image should be here but it was not working when we last tested it.
Here's the filtering setup. I have always aimed for low order filters and minimal EQ'ing.
2 x BMS 18N860 - Tapped horns:
An externally hosted image should be here but it was not working when we last tested it.
2 x Deltalite 2515 - Synergy midwoofers
(no signal correction at all)An externally hosted image should be here but it was not working when we last tested it.
JBL 2446J + Truextent - Synergy comp driver
An externally hosted image should be here but it was not working when we last tested it.
2 x JBL 2404H - supertweeters (mounted currently inside the synergy horns, one facing front and other facing backwards in same phase, works nicely)
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 work Legis! Thanks for the detailed DSP settings.
I find it remarkable that you got an extra 20Hz of bass extension, now down to 50Hz with the volume phase pucks. Or is that just plain old drivers breaking in?
You are basically able to listen to most music without the TH sub and it will sound quite full. What is your feeling of how it sounds without the TH going?
For what it's worth, I would not be so shy of going in and applying a negative low Q (circa Q=0.9 to 1.3) peak cut (circa -6dB) at 150Hz to flatten the response. Also apply a -3dB (circa Q=0.7 to 1.2) at 325Hz peak cut to get rid of the "hill". This will level set the overall response at 107dB which isn't bad at all and will flatten out your minimum phase, and reduce distortion. Broad (low Q) peak cuts are not bad. It is DSP so hurts nothing to try and you can always go back easily. You may find that doing this transforms your speaker to have a much nicer balance and cleaner transient response.
I find it remarkable that you got an extra 20Hz of bass extension, now down to 50Hz with the volume phase pucks. Or is that just plain old drivers breaking in?
An externally hosted image should be here but it was not working when we last tested it.
You are basically able to listen to most music without the TH sub and it will sound quite full. What is your feeling of how it sounds without the TH going?
For what it's worth, I would not be so shy of going in and applying a negative low Q (circa Q=0.9 to 1.3) peak cut (circa -6dB) at 150Hz to flatten the response. Also apply a -3dB (circa Q=0.7 to 1.2) at 325Hz peak cut to get rid of the "hill". This will level set the overall response at 107dB which isn't bad at all and will flatten out your minimum phase, and reduce distortion. Broad (low Q) peak cuts are not bad. It is DSP so hurts nothing to try and you can always go back easily. You may find that doing this transforms your speaker to have a much nicer balance and cleaner transient response.
Last edited:
Nice work Legis! Thanks for the detailed DSP settings.
I find it remarkable that you got an extra 20Hz of bass extension, now down to 50Hz with the volume phase pucks. Or is that just plain old drivers breaking in?
You are basically able to listen to most music without the TH sub and it will sound quite full. What is your feeling of how it sounds without the TH going?
For what it's worth, I would not be so shy of going in and applying a negative low Q (circa Q=0.9 to 1.3) peak cut (circa -6dB) at 150Hz to flatten the response. Also apply a -3dB (circa Q=0.7 to 1.2) at 325Hz peak cut to get rid of the "hill". This will level set the overall response at 107dB which isn't bad at all and will flatten out your minimum phase, and reduce distortion. Broad (low Q) peak cuts are not bad. It is DSP so hurts nothing to try and you can always go back easily. You may find that doing this transforms your speaker to have a much nicer balance and cleaner transient response.
Yes the synergies are quite listenable without the tapped horns, especially if the room and placement supports them. Extension of a shelf speaker with outrageous dynamics
. Tapped horns integrate very seamlessly to them (even without any electrical crossover) and their sound signature. The new 2404H super tweeters are also good match in the sound signature area so the system is going forward.Quite many things change when measured from further away, I try to add some more measurements today.
I think the effect in bass extension is at least partly caused by break in but the smaller front chamber - at least via compression and air load - might add to it somewhat even though simulation did not predict it. Also, maybe the front chamber is starting to act more efficiently as an extension of the horn/throat and thus the loading and the lenght of the horn effectively changes? It becomes more like a longer, small opening angle (like expo) throat when the throat chamber is understood as a whole with the injection port. I'd see this as a very good possibility.
I also like the sound of the smaller front chambers, I'd say they sound slightly more dynamic in their whole pass band propably doe to higher compression (which introduces more air nonlinearity effects making the reproduction to sound & feel more explosive - a phenomenom I'm trying to understand).
Last edited:
I drew a quick picture with coaxial and offset injection port variations how the effective horn path lenght might change with a shaped front chamber.
In both cases the furthest corner of the cone (= cone's surround) could be understood to become more the "start of the horn" when shaped like that (or somewhere in between the surround and the dust cap). Meaning the offset port makes the horn efectively slightly longer than the coaxial port. Since this added path lenght is "high compression path" along the cone and filling insert piece, the effect in tuning is greater.
If the front chamber is not filled, it propably does not act as efficiently as a part of the horn, but acts mainly just as a lowpass chamber without adding many inches to the path lenght. By reducing the size and shaping the front chamber it becomes more efectively part of the horn at least at bass freqs where even the crudest approximations of different flare rates are just fine for the long bass frequency waves. In short synergy horns every centimeter counts
. Also added compression path in (quickly opening) conical flare can never hurt the loading! Also the bigger the cone, the more loading and path lenght can be achieved with shaping the front chamber.Some analogy would be the Paraline concept.
Maybe this can explain some of the more extended bass response - or not.
I believe it can be highly propable. It might be most obvious if the offset ported cone would be 1m wide and the horn it's attached to would be 1m long, just enought for the cone to be attached to the horn (imagine that!) - this would propably result in effectively ~2m path lenght with shaped front chamber.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:
For a quarter wave horn to go from 70Hz to 50Hz requires extra 0.5m of horn length! Probably the extra distance in chamber is not responsible for this. It is probably more along the lines of increased efficiency at the lower wavelength which raises the floor of the SPL up, thereby extending the apparent bass extension. Looks like a 7dB boost at 50Hz occurred.
It's quite interesting what really happens inside the front chamber, how long the pressure/waves travel there and what it the effect on total horn lenght. 15 inch woofer is ~34cm wide, the "perimeter" (like in the previous pictures) of the cone is somewhat more. The increased compression will also add to the equation by raising the sensitivity at the bottom end. I think the horns actually play little louder the whole pass band but the sensitivity increased the most at the bass.
Finally finished the front chamber of other horn also.
Response of the old unfilled, but damped front chamber vs. the new filled (but undamped) front chamber. The first mentioned was taken 2 day ago, the latter today so the possible burn in effect does not affect to this anymore.
Green is the new filled front chamber. Both are measured at the mouth (0cm).
Here's both channels with the new front chambers, quite small tolerance between the channels.
Here's the response from 170cm away fron the mouth (lively concrete room). Those near field effects/bumps are different when measured from afar, I presume they pretty much disappear but the room modes' tearing effect replaces them.
Response of the old unfilled, but damped front chamber vs. the new filled (but undamped) front chamber. The first mentioned was taken 2 day ago, the latter today so the possible burn in effect does not affect to this anymore.
Green is the new filled front chamber. Both are measured at the mouth (0cm).
An externally hosted image should be here but it was not working when we last tested it.
Here's both channels with the new front chambers, quite small tolerance between the channels.
An externally hosted image should be here but it was not working when we last tested it.
Here's the response from 170cm away fron the mouth (lively concrete room). Those near field effects/bumps are different when measured from afar, I presume they pretty much disappear but the room modes' tearing effect replaces them.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I'm propably going back to damped front chambers. I have been focusing too much in the >500Hz band and got somewhat blind to the effects in the actual pass band . I noticed that something was different (in a not good way) when listening the synergies for the past days with familiar records I have played alot and I compared the responses in the old records again. Damped front chamber has had the best response in the pass band this far (when measured near field). The filled chamber actually has the biggest saddle in ~200hz. Undamped and unfilled (= original) chamber has the most pronounced resonanse 250-400Hz region but it lacks the big saddle in the 200Hz what the filled chamber has.
Damped front chamber is blue, filled undamped front chamber is red. Measured at the mouth (ie. 0cm).
I suspect the filler piece lenghtens the effective lenght of the front chamber injection port/pipe, which causes the response to be more uneven compared to the most shallow injection port possible (when the front chamber is unfilled). I don't know if this can be prevented, my filler pieces do not form any sort of "wall" extension with the ports, but the story in acoutical domain can be very different from visible domain.
The effect of deeper injection pipe is pretty much like measured difference also according to HR sim - more saddle in the 200hz area and more hill in the 250-400hz area compared shallower ports.
Based on this I also started making the injection ports as shallow as possible without being fragile. They were rounded before but not nearly as much as now. The "tip" is only couple of millimetres thick. This should give quite transpared injection ports with least amouth of the tube-effect, hopefully.
I never though the mids would be the most sensitive part to get right in a synergy. Comp drivers have worked like being in the best horns I have yet heard them playing in from the start. Mids are the cranky b*tches.
. I noticed that something was different (in a not good way) when listening the synergies for the past days with familiar records I have played alot and I compared the responses in the old records again. Damped front chamber has had the best response in the pass band this far (when measured near field). The filled chamber actually has the biggest saddle in ~200hz. Undamped and unfilled (= original) chamber has the most pronounced resonanse 250-400Hz region but it lacks the big saddle in the 200Hz what the filled chamber has.Damped front chamber is blue, filled undamped front chamber is red. Measured at the mouth (ie. 0cm).
An externally hosted image should be here but it was not working when we last tested it.
I suspect the filler piece lenghtens the effective lenght of the front chamber injection port/pipe, which causes the response to be more uneven compared to the most shallow injection port possible (when the front chamber is unfilled). I don't know if this can be prevented, my filler pieces do not form any sort of "wall" extension with the ports, but the story in acoutical domain can be very different from visible domain.
The effect of deeper injection pipe is pretty much like measured difference also according to HR sim - more saddle in the 200hz area and more hill in the 250-400hz area compared shallower ports.
Based on this I also started making the injection ports as shallow as possible without being fragile. They were rounded before but not nearly as much as now. The "tip" is only couple of millimetres thick. This should give quite transpared injection ports with least amouth of the tube-effect, hopefully.
An externally hosted image should be here but it was not working when we last tested it.
I never though the mids would be the most sensitive part to get right in a synergy. Comp drivers have worked like being in the best horns I have yet heard them playing in from the start. Mids are the cranky b*tches.
Last edited:
I will try that next with the same very dense (110kg/m3) foam plastic or whatever it's name what I used previously.
An externally hosted image should be here but it was not working when we last tested it.
The two pieces take around 1,5 litres of the internal air, over double the amount that I used on the first round of front chamber damping. This time around I also made the bottom piece bigger covering the whole area of the cone and it starts to act as a mechanical brake after around 6-7mm of excursion, quite ok for the 4,8mm xmax Deltalites. You hardly feel them moving though even at infra freqs.
Last edited:
I have been making as-simple-as-possible passive crossovers for the synergy + tapped system for the past week. I'm getting close but some tweaking still needs to be done.
The tapped horns and synergy midwoofers are completely unfiltered (just acoustic filtering). I get about the same acoutic lowpass with tapped horns than with the front chambered front horns.
2446J + Truextent compression drivers have 2nd order high pass at ~500Hz, supertweeter have 3rd order high pass.
Naturally no delays, no notch-filters.
I have adjusted the level of the compression driver "losslessly" with a good quality 4:1 step-down transformer (ie. -12dB) trying to avoid burning any power into resistors. Being lossless, it keep the efficiency of the driver and makes the effective impedance very high like 150-300ohms. Other bands have their dual drivers connected in series to reduce the voltage sensitivity just about right because all the bands have pretty close sensitivities (except the 2446J).
I cannot think of much simpler way to crossover a 4-way horn system.
Simplistic passive filtered setup is very nice change and simplifies the system and "cable-hell" quite much. I also like the sound very much. Now I'm able to use only one 2ch amplifier and one stereo DAC.
These are listening spot measurements. After taking the measurements, I have reduced the level of the midbasshorns by ~3dB with an L-pad.
Here's the impedance plots of the separate bands and the whole speaker.
Tapped horns:
Deltalites: (The spike at ~380hz is Deltalites modulating the 2446J membrane, I need to get rid of that!)
2446J + Truextent + passive filtering & 4:1 step-down transformer (the sensing resisitor could have been better suited for such high Z load to sense it more accurately without the haze but I did not bother):
2404H supertweeters:
Whole speaker:
Minimum impedance was around 6,2 ohms and phase shift is quite small, 35deg at maximum. Now the Deltalites have the -3db L-pads so the impedance of the whole speaker is even easier.
The tapped horns and synergy midwoofers are completely unfiltered (just acoustic filtering). I get about the same acoutic lowpass with tapped horns than with the front chambered front horns.
2446J + Truextent compression drivers have 2nd order high pass at ~500Hz, supertweeter have 3rd order high pass.
Naturally no delays, no notch-filters.
I have adjusted the level of the compression driver "losslessly" with a good quality 4:1 step-down transformer (ie. -12dB) trying to avoid burning any power into resistors. Being lossless, it keep the efficiency of the driver and makes the effective impedance very high like 150-300ohms. Other bands have their dual drivers connected in series to reduce the voltage sensitivity just about right because all the bands have pretty close sensitivities (except the 2446J).
I cannot think of much simpler way to crossover a 4-way horn system.
Simplistic passive filtered setup is very nice change and simplifies the system and "cable-hell" quite much. I also like the sound very much. Now I'm able to use only one 2ch amplifier and one stereo DAC.
These are listening spot measurements. After taking the measurements, I have reduced the level of the midbasshorns by ~3dB with an L-pad.
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.
Here's the impedance plots of the separate bands and the whole speaker.
Tapped horns:
An externally hosted image should be here but it was not working when we last tested it.
Deltalites: (The spike at ~380hz is Deltalites modulating the 2446J membrane, I need to get rid of that!)
An externally hosted image should be here but it was not working when we last tested it.
2446J + Truextent + passive filtering & 4:1 step-down transformer (the sensing resisitor could have been better suited for such high Z load to sense it more accurately without the haze but I did not bother):
An externally hosted image should be here but it was not working when we last tested it.
2404H supertweeters:
An externally hosted image should be here but it was not working when we last tested it.
Whole speaker:
An externally hosted image should be here but it was not working when we last tested it.
Minimum impedance was around 6,2 ohms and phase shift is quite small, 35deg at maximum. Now the Deltalites have the -3db L-pads so the impedance of the whole speaker is even easier.
Last edited:
Couple sound/video clips can be found here. Fully passively xo'ed like clarified in previous post: http://www.diyaudio.com/forums/multi-way/23208-system-pictures-description-269.html#post4221028
I have to say I propably like passive xo system better as-is. Utterly black background and more organic presentation.
I have to say I propably like passive xo system better as-is. Utterly black background and more organic presentation.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.