Unwindowed MLS and Sinusoidal:
[Which speaketh the truth? ]
Unwindowed MLS and Sinusoidal:
[Which speaketh the truth? ]
Whichever one looks the best.
I'll do some more 2425/2307/2308 measurements in a week or so and see what happens.
The pic below has another opinion.
Here's Greg's -
Keep in mind that a ground plane measurement does not give the correct midrange curve. (100 Hz to 400 or 500 Hz) This is due to the enclosure dimension being doubled by sitting on the ground. The baffle effects are presented at the wrong frequency. I figure if you try different distances between the ground plane and near field, you should be able to zero in on common ground in a splice region. Another possibility is to do curves outside and just put the speaker on a stand. Even 3 feet will help substantially at the 1 m mic distance. Any different measurements you can make will help to show what is real and what isn't.
That last sentence basically reiterates Doug's view on the subject.
O.K., thank you, Giskard.
I updated the Sinusoidal to include gated and stepped.
Haven't used that enough to know what's optimum.
RTA says MLS and gated groundplane are correct.
[RTA's always my fallback frame of reference.... ]
They're remarkably uniform considering varying age and indefinite provenance, I must say.
If it's real, there's an easy fix (bottom two)....
Yup, Bo. 4-5 dB is a fairly major correction, actually. I'm reluctant to post the specifics until Giskard verifies this with his independent testing using 2425s, Todd's newly acquired driver type, as opposed to these LE85s. The "fix," if any, will likely be different. See the second graph here, 2426J on HL91:
http://audioheritage.org/vbulletin/s...5&postcount=15
Once running flat, the narrow (and collapsing) vertical beamwidth really becomes apparent. These need a tweeter, not so much for the VHF (13.5 kHz being plenty extended for some applications,) as for overcoming their beaminess.
David Smith documents the beamwidth as ~25° @ 10 kHz, and ~45° at 5 kHz. See Fig. 4 here:
http://audioheritage.org/vbulletin/s...ead.php?t=7852
Boosting the VHF would thus be a fruitless endeavor. Instead, I'm mating them with a 90° x 50° waveguide tweeter @ 6 kHz to see if that integrates well (bottom). Optimum might be somewhat lower, probably.
After that, I've gotta get back to doing my own stuff here. Todd always comes up with provocative subjects requiring empirical determinations....
I simply have got to get over there and see this place for myself!
bo
"Indeed, not!!"
Why not do 10k with a slot. 4344 measured 2 meters on axis. You know what the kicker is?? I am using Ti drivers and the filters help flatten it out a bit and you have the L-Pads. If you look real close there is a shallow rise from 3-8K about 4Db and the compression driver is down about 2Db relative to the average midrange level. Looks a lot like your Green and Blue curves imbedded in the overall response. You can use the L-Pad to center the response curve and still be in a +/-2 window as measured with an RTA we all know that a real CLIO measurement will be like the 4345 measurement.
Rob
Geez, Rob that looks GOOD!
I did try it with the slot at 8.5 kHz, and that worked well, but the rising midrange using LE85 was still apparent; I just didn't understand it:
http://audioheritage.org/vbulletin/s...ad.php?t=12522
The major problem I found with that approach was the narrow vertical beamwidth of the slot. Run it alone, and it's immediately apparent.
So, this time, I'm trying to maintain 50° all the way up. Down at 6 kHz, I'm finding the MF lowpass is well down at 7 kHz, and the HL91 anomaly is all but buried, as with your filter and driver combination. I'd expect that additional compensation is part of the filter design originally incorporated in the three-ways, and/or subsequent updates thereof.
In the two-ways like S7 and L200, if it's real, it matters, and seems to be a major portion of their voice I've never much cared for. Converting them to three-way with 3105 mitigates it considerably, as well, from what I'm measuring here.
I'm still hearing some horniness; now determining how much of that's coming from running HL91 at 800 Hz. No HL92 here, alas.
Thankfully, I'm not up to THAT task....
Ontogeny recapitulates phylogeny, I guess. It's part of figuring out the whys and wherefores of how this stuff works (and doesn't, some times.) I've linked to your N333 update several times, actually.
Too much like work. Two-ways are more fun....
Which version of my equivalent network are you running? First or second?
Are you running the aquaplased diaphragms here or are they in something else?
Which reminds me, why are you screwing around with these older networks Zilch when I already did the N333 equivalent network years ago?
Just for grins take a look at the response for the 4330 vs the 4333 in the Library. They tell a lot. Especially the change in the FR curve from 7K up
Rob
Unless Rob has changed something they are the very first equivalent 3145 croosovers we did that I left with Robert back in 2004. The voltage drives on that particular network were spot on (Hovland 5% tolerance).
Ian
NEW EQUIVALENT 3145
Here is a link to the new equivalent networks I built with on axis and 30 degrees off axis response curves of the horn mounted in the 4345 baffle. Note the drives are mounted off centre and on a large planar baffle. The resulting curves and tailoring of the crossover yield a remarkably smooth on and off axis response. The driver is a 2420 with titanium diaphragm.
http://www.audioheritage.org/vbullet...ght=equivalent
Hello Giskard
These are with the original version of the crossover you worked out with Greg. This is an old measurement using 2416's for the compression driver biamped with an M552 with 24dB L/R slopes all drivers in phase. The system now has the aguaplassed 2425's biamped with the DX-1 with 12dB slopes doing the crossover. Woofer out of phase like the original.
Rob
There are currently 1 users browsing this thread. (0 members and 1 guests)