In my living room, which is located about here:
https://goo.gl/maps/91YyTBdN2rArwa6U6
You are cordially invited to take a listen!
https://goo.gl/maps/91YyTBdN2rArwa6U6
You are cordially invited to take a listen!
That's an insert of birch plywood. There was another Tweeter used with 110mm OD, so i decided to rebuild this with router + circle. It also looks nice to me, but I had not spent that efford just for optics.
I compared that 1:1 non-blinded during measurements and noticed an obvious difference with fast sine sweeps up to~48kHz. Without the notch there was a sharp note in the reverberation after the measurement that disappears with the filter applied. I did not compare 1:1 with stereo music, and I wouldn't bet on being able to tell the difference in a blind test.
The Tweeter crossover is now rebuilt:
Current tuning compared to the original:
1) Tweeter's level has been decreased a bit around 3-4kHz intentionally, there was slightly noticeable overemphasis in the power response before.
2) The phase response of the drivers around the crossover frequency has been significantly improved and is now nearly flawless, for the architecture costs of an additional 15uF Capacitor.
3) Around ~1kHz there is minimally more level by the mirdrange notch. 5.6 ohms is too "fat" for me, 6.8 ohms is too "lean", 6.2 ohms fits well.
Crossover:
6.2 and 4.3 ohms are somewhat unusal values, but can be found in the E24 series, for example, as Jantzen Superes without issues.
Power Response and DI only hor +/-60°, only ver +/-40° (both dashed) and hor+ver together:
Power Response and Predicted In Room Response based on hor +/-60° only:
Inverse-Zero, phase response:
Effect of notches applied to diaphragm resonances:
Directivity response based on hor +/-60° and ver +/-40°:
Vertical response +/-40°:
Best regards
Peter
Last edited:
Just a very minor update: By the time I pushed the tweeter upper range ~ +half a dB by adding a 0,1uF capacitor in parallel to most of the filter network. I think it sounds a bit more balanced now, but maybe it is a matter of taste...
You can use both depicted 100nF = 0,1uF capacitors to adjust the tweeter level in a limited range, as there is no voltage divider in the tweeter network for tuning. Increasing the 0,1uF || 2,2 uF pushes the lower treble, increasing the 0,1uF || "everything" pushes the upper treble. Leaving both 0,1uF out gives the lowest level ~ -1dB to my actual voicing, which may sound a bit too dark:
The speaker is driven now by a relatively cheap chain of SMSL C100 AKM4493 DAC and a Topping PA5 II Amp. Pretty good match for a ~300€ in electronics, sounds very flawless and balanced to me. The power of the PA5 II non-"Plus" is generously sufficient to drive the speakers to max volume.
Best regards
Peter
You can use both depicted 100nF = 0,1uF capacitors to adjust the tweeter level in a limited range, as there is no voltage divider in the tweeter network for tuning. Increasing the 0,1uF || 2,2 uF pushes the lower treble, increasing the 0,1uF || "everything" pushes the upper treble. Leaving both 0,1uF out gives the lowest level ~ -1dB to my actual voicing, which may sound a bit too dark:
The speaker is driven now by a relatively cheap chain of SMSL C100 AKM4493 DAC and a Topping PA5 II Amp. Pretty good match for a ~300€ in electronics, sounds very flawless and balanced to me. The power of the PA5 II non-"Plus" is generously sufficient to drive the speakers to max volume.
Best regards
Peter
Last edited:
You're right, definitely not the easiest load to drive!
DC resistance of the woofer + coils is already < 3,5 Ohm, but the large parallel cap + the parallel notch are dropping the impedance < 3 Ohms. Had tried other possibilites in simulation, but I haven't found any other sensible solution without using a lot more components, or changing the SB woofer to 8-Ohm type.
In practice, till now I had the speakers connected to a Sony TA-F530ES, a Rotel RA-970BX, a Sony TA-F210 and now the Topping PA5 II Amp. All of them worked fine in general and were able to drive the LS to max volume stable and without complains, none of them went to protection or went unusually hot or so.
The pretty old, worn out and simple Sony TA-F210 played really nice with a balanced, long-term enjoyable sound.
Rotel RA-970BX gave more control, but sounded a bit more washed-out in the mid-highs. Well, it's a Rotel....
Sony TA-F530ES was a dissapointment, the fundamentals got a bit uncontrolled and the mids got a bit rough. Maybe it really didn't like the impedance profile.
New Topping PA5 II is driving the speakers very nicely, transparent and balanced like the TA-F210 but with more control like the Rotel. A very good match, and although it is a switching amplifier, there is no slightest hint of sharpness in the treble.
So I think a "standard" Japanese amp or smaller, modern switching amplifier in the moderate 50-100W power range is pretty sufficient, but it should have a stable power cube and be capable to deliver some amperes. Could imagine that an NAD partners also nicely. The speakers a definitlely no match for a tube amp, Pass "First Watt" or such exotics.
Best regards
Peter
DC resistance of the woofer + coils is already < 3,5 Ohm, but the large parallel cap + the parallel notch are dropping the impedance < 3 Ohms. Had tried other possibilites in simulation, but I haven't found any other sensible solution without using a lot more components, or changing the SB woofer to 8-Ohm type.
In practice, till now I had the speakers connected to a Sony TA-F530ES, a Rotel RA-970BX, a Sony TA-F210 and now the Topping PA5 II Amp. All of them worked fine in general and were able to drive the LS to max volume stable and without complains, none of them went to protection or went unusually hot or so.
The pretty old, worn out and simple Sony TA-F210 played really nice with a balanced, long-term enjoyable sound.
Rotel RA-970BX gave more control, but sounded a bit more washed-out in the mid-highs. Well, it's a Rotel....
Sony TA-F530ES was a dissapointment, the fundamentals got a bit uncontrolled and the mids got a bit rough. Maybe it really didn't like the impedance profile.
New Topping PA5 II is driving the speakers very nicely, transparent and balanced like the TA-F210 but with more control like the Rotel. A very good match, and although it is a switching amplifier, there is no slightest hint of sharpness in the treble.
So I think a "standard" Japanese amp or smaller, modern switching amplifier in the moderate 50-100W power range is pretty sufficient, but it should have a stable power cube and be capable to deliver some amperes. Could imagine that an NAD partners also nicely. The speakers a definitlely no match for a tube amp, Pass "First Watt" or such exotics.
Best regards
Peter
Last edited:
I just wonder if an MTM version might be worth doing too. There is always a Mk.2 version...
So, here we go
:Bucking magnets ready:
I build the shown enclosures above around ~14 years ago for the mother of my girlfriend (now my wife), drivers were from an Audax remaining stock sold by the german distributor for ~20-25 € per piece:
Somehow the speakers have then found their way to my wife's younger cousin (he's a really nice and handy guy) who is a live sound engineer and musician by profession. He contacted me some month ago, because one driver was broken:
He told me he really likes the speakers, as they sound good and have accompanied him through some shared communities and a lot of parties, and he wants to keep them alive. A few weeks ago he visited me and brought the speakers, we had some discussions about the options, he had a listen to the Edelstoffs and some budget to spend and we decided to rework and upgrade the speakers then.
Tomorrow sunny weather is announced, so I plan to go out and grip my router and rework the enclosure to match the new drivers (fortunately the new drivers all are few millimeters larger in all dimensions), and refinish them.
Best regards
Peter
Last edited:
This one: https://www.magnetladen.de/ringmagnet-90x36x20-mm-y35/
Gives ~+5% BL instead of ~+7% with the neodymium magnet, but is much cheaper...
Gives ~+5% BL instead of ~+7% with the neodymium magnet, but is much cheaper...
Well, I would not say "significantly", but I percieved some improved defintion and precision of the sound subjectively.
The enclosure is small and according to TSP simulations FR and GD might improve with the magnet:
SB15/SB17 have exceptional clean midrange, especially CAC and NBAC, but for woofer duty with high excursion requirements there are out better ones.
Measurement morning...
DUT: Edelstoff XT
Some temporary guests from the neighbour farm:
Got pretty happy results:
Measurement setup:
Measurement data:
Observations:
Lower woofer is integrated textbook baffle-step-filler concept style; you can see that also on the electrical slopes, midwoofer stays flat generally, lower woofer fills up then +6dB below baffle step.
Xover is modified 3rd order electrically, aiming quasi LR4 acoustical slopes. Phase matching and tracking is slightly worse than with the original Edelstoff, but nice for the minimum efford tweeter filter, the additional 22 Ohm modification gives a degree of freedom helping to shape the response a bit. See inverse null response as dotted line in SPL window.
The baffle edges without rounding or chafering result in an on-axis FR drop around 3,5kHz due to edge diffraction. Never fill that up, it will sound sharp!
Optimized the system for linear power response of +90° ... -90° hor data, and linearity of listening window based on +30° ... -30° hor data.
Above 10kHz, the DXT tweeter unfortunately has low sensitivity, so the response drops a few dB till 20kHz. This is typically for the DXT. The applied overshoot of the tweeter filtern nework >10kHz helps to gain 1-2dB back.
In dotted lines you see the effect of an additional parallel notch filter in series of the tweeter to surpress the 27kHz dome breakup resonance.
Both 3rd order electric filters of the woofers have each a small capacitor in parallel to the 2nd coil as modification; 2nd coil sized to serve the filter slopes, the capacitors tuned then so that the >10kHz cone breakup resonances are surpressed. The lower woofer network has small 2nd coil as not needed to contribute to the filter slope slope, so the notch is for fighting the cone ringing only:
Resulting system impedance, minimum is 3 Ohm @ 250Hz.
Still have to optimize the internal damping - losses of the vent in the usable range are a bit high, therefore it shows unwanted lenght resonance...
On-axis responses vs. unfiltered driver responses:
These are responses with very long gate containing the 1st floot bounce, 1/6 octave smoothing:
Should work, or? Any ideas before I will order missing xover components?
Best regards
Peter
EDIT:
may start with additional 150nF || 3,9uF main tweeter capacitor for bit fresher highs voicing:
DUT: Edelstoff XT
Some temporary guests from the neighbour farm:
Got pretty happy results:
Measurement setup:
- Monacor ECM-40 mic, Monacor MPA-102 preamp, FR calibrated
- T.amp PM40C measurement amplifier, ARTA measurement switchbox.
- Computer with ASUS Xonar DX soundcard for I/O, ASIO drivers, ARTA/LIMP V1.9.4.1
- Mic height ~2,2m, mic distance ~2m
Measurement data:
- drivers individually measured +90° ... -90° horizontal, gated with ~140Hz resolution. Both woofers directivity data was measured for each driver, with the other then short-circuited.
- On-axis response is merged with nearfield responses. Both woofers nearfield data was measured individually, with the other driver then short-circuited. Vent response with both woofers conected in parallel, reduced by -6dB for merging with each woofer response (measured/merged other electrical combinations and compared results in simulation also - many look similar in the result).
- Impedance response for each woofer is the measured impedance of both in parallel, then doubled for simulation. (measured other electrical combinations, e.g. one woofer connected, the other one short-circuited and compared results in simulation also - most looks similar in the result).
Observations:
Lower woofer is integrated textbook baffle-step-filler concept style; you can see that also on the electrical slopes, midwoofer stays flat generally, lower woofer fills up then +6dB below baffle step.
Xover is modified 3rd order electrically, aiming quasi LR4 acoustical slopes. Phase matching and tracking is slightly worse than with the original Edelstoff, but nice for the minimum efford tweeter filter, the additional 22 Ohm modification gives a degree of freedom helping to shape the response a bit. See inverse null response as dotted line in SPL window.
The baffle edges without rounding or chafering result in an on-axis FR drop around 3,5kHz due to edge diffraction. Never fill that up, it will sound sharp!
Optimized the system for linear power response of +90° ... -90° hor data, and linearity of listening window based on +30° ... -30° hor data.
Above 10kHz, the DXT tweeter unfortunately has low sensitivity, so the response drops a few dB till 20kHz. This is typically for the DXT. The applied overshoot of the tweeter filtern nework >10kHz helps to gain 1-2dB back.
In dotted lines you see the effect of an additional parallel notch filter in series of the tweeter to surpress the 27kHz dome breakup resonance.
Both 3rd order electric filters of the woofers have each a small capacitor in parallel to the 2nd coil as modification; 2nd coil sized to serve the filter slopes, the capacitors tuned then so that the >10kHz cone breakup resonances are surpressed. The lower woofer network has small 2nd coil as not needed to contribute to the filter slope slope, so the notch is for fighting the cone ringing only:
Resulting system impedance, minimum is 3 Ohm @ 250Hz.
Still have to optimize the internal damping - losses of the vent in the usable range are a bit high, therefore it shows unwanted lenght resonance...
On-axis responses vs. unfiltered driver responses:
These are responses with very long gate containing the 1st floot bounce, 1/6 octave smoothing:
Should work, or? Any ideas before I will order missing xover components?
Best regards
Peter
EDIT:
may start with additional 150nF || 3,9uF main tweeter capacitor for bit fresher highs voicing:
Last edited:
