|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
64.105.104.82
In Reply to: RE: Getting EV1823's and ST350's in phase..Ping Paul E. posted by Hiendmuse on May 30, 2014 at 14:33:32
H
The first step would be to get the tweeters to stand up by themselves. The horn of ST350 has a flat bottom, so it just needs a bit of shimming under the magnet to get it level. The T350 and T35 would need a wooden block screwed to the bottom of the horn mouth, as the long dimension being vertical is the preferred way to mount them. Cylindrical horn tweeters like the Fostex's work with dowel rod bases. You don't mention the tweeters you intend to use, but some improvisation is likely in order. Don't forget cardboard and duct tape, the horn man's secret weapons.
Paul
Follow Ups:
Done! Well at least the temp sled. There were three null points, which would make sense, given the distance front-to-back of the ST120 horn and the wavelength at 3.5KHz. The null was down about 7-8dB, from the measured peak. When the polarity was returned to normal, I got 9-10dB "reinforcement" at the XO point of 3500Hz. Does that seem "typical"? (base level was about 70dB)
I am planning to use the front-most point to avoid spurious reflections. This puts the ST350 about 2.6" back from the front lip of the horn. The ST350 would only be elevated maybe .5-.75" above the mid horn and, perhaps, pretty much level with the front lip. So, this point seems very manageable! (This will only yield phase alignment, not time alignment, if I understand correctly. I like Bill's idea of the tweeter output being a couple of wavelengths earlier than the mid for slight brightness lift.)
The change was not dramatic, but brought an overall improvement in clarity to the sound in timbre, imaging and sound stage. The treble actually seems slightly less, but that might be just better integration in the overall sonic signature. So, although not dramatic, I won't consider going back to the stock configuration!!! Well worth the effort! Hmmm...maybe that is dramatic!? ;-)
I am considering making 90 degree metal brackets to use the threaded inserts on the top front of the horns. Then cherry wood "risers" from the metal brackets to which I would mount the ST350. The rear of the ST350 would be supported with a cherry bracket. My mounts for the mid horns allow some tilt from the rear, so both horns could be brought closer to the horizontal axes at the listening position. Sound reasonable? Still thinking it through, so any thoughts would be greatly appreciated.
Again, thanks to Paul & Bill. You guidance got me up off my @$$!!! I am amazed at how much improved the overall sound is from my initial inquiry regarding alternatives to the ST350. Now they aren't going anywhere!!
Cheers,
Geary
Geary
A 3500 Hz sound wave is about 3.8 inches long.* The highest pressure (using a sine wave example) would be at the 1/2 wave point, or at about 1.9 in. So with the drivers reverse wired, you would see a null at about 2 inch intervals. You want to pick the null which is closest to the area where the drivers diaphragms are aligned in the vertical plane as seen from the side. This seems unlikely to be at a spot that near to the point where the mid and tweeter horns mouths are that close to each other, though it may be an improvement over just lining up the mid and tweeter mouths as previously.
A word about time alignment^tm, which is a copyrighted term, so I tend to use the phrase "phase alignment". Ideally you would check the alignment by feeding the drivers with an impulse signal and observing that their outputs are arriving at the listening position at the same time, with a test microphone into an oscilloscope. Using the reverse polarity null test is an easier way to approximate this. Either way the goal is to get the tweeter and mid output transients to arrive at the listening position with minimal delay between them. The rules are different between a speaker for home listening and a PA rig, though the equipment is often the same as is this case. In a PA rig you don't want a narrow sweet spot, and it's acceptable to have a rather vague stereo image as long as the coverage is acceptable. In the home you want the speakers to have a good stereo image with discreet image localization, and with broad coverage being secondary to this. In the cases of some current high end recommended components in well treated rooms, the off center coverage is non existent! Most Stereophile/TAS reading audiophiles seem oblivious to this, though they will patiently wait for the sweet spot at a demo, where half the orchestra does'nt disappear anymore.
Anyway the point is, try it both ways with music you like. With the drivers aligned in phase, the imaging will be better, with a stronger center phantom image between the speakers, and better instrument localization. The system will also sound more coherent with the drivers blending better. The whole "soundstage" thing has gotten blown out of proportion in recent years in my opinion, and most audiophiles seem oblivious to the fact that a live symphony orchestra dose'nt image all that well compared to some of the "one seat wonders" championed by the glossy magazines. Don't forget that moving the tweeter back and forth will change it's level, and you need to compensate for this with a L-pad. A nice female vocal is a good test, as the female voice bridges the area between the mid and tweeter. The Wail'n Jennys "One Voice" is my current fave test track, as the 3 part harmonies come in one at a time. Let us know how it goes.
Paul
* 13,500/Freq. in Hz=wavelength in inches.
Cr@p!!! I must have done something wrong. I was not getting nulls at 2" intervals, it was closer to 4" from my notes. 2" nulls makes sense. I was using a 3.5KHz sine wave. I will try again over the weekend. This time I will start measuring moving forward from the rearmost alignment first, after double checking the polarity of the driver leads from the XO. Hopefully I can recruit an assistant. That would greatly speed up the process.
Maybe wrong, but still sounds better!!! If "right" is even better I will be most pleased!
Cheers,
Geary
Geary
You should be measuring some differences at multiples of about 2 inches, so you're on the right track. Using a sine wave test signal can be problematic due to the fact that a steady state signal can excite room resonances which can confuse the results. The Stereophile Test Disc #2 CD has warbling test tones which continuously sweep about 1/3 of an octave which minimizes the effects of room resonances. Back in the day I bought the kit to build the oscillator used for the recording of the test signals, but I never got around to building it due to the handiness of the CD. In the end, if it sounds right it IS right!.
Paul
Thanks Paul. Yes there were some differences at intervals between the 3.9" points, but not the deepest nulls. I was focused on finding the deepest, so may have blown by the others.
I quickly realized the issues with the 3.5KHz sine wave. I had to reduce the base level to about 70dB to not excite the room resonances.
I do have the Stereophile Test CD 2, but the warble tones are centered at 3.15KHz and 4KHz. The XO point is 3.5KHz. So I didn't use, because I thought they weren't close enough. If either will work, that would be great!
I found a tone generator online:
http://www.wavtones.com/functiongenerator.php
Pretty cheap to buy a full function license. Could use it to generate a 3.5KHz warble tone and download.
Please let me know your thoughts on which may work the best.
Cheers,
Geary
Geary
The 3.15K Hz tone should work as the tone sweeps up and down over a 1/3 octave range, with 3.15K being the center frequency. You can check your results between this and the 4K test tone.
The NCH software program looks good, though I have'nt tried it, and the price is certainly reasonable compared to my unbuilt kit ; ) You would need to buy the upgraded version to generate tones longer than 5 seconds. If you select the lower end of the sweep at 2,333 Hz and the higher end at 4,083 Hz, this should put the center close to 3.5K Hz.
Another approach would be to use REW (Room EQ Wizard, a free download for both Mac & PC). With this you can generate a rising sweep tone between 2 selected frequencies, and a graph is made which is created from samples at fractions of a second intervals to minimize room resonances. Just remember to use the "smoothing" feature so you can see the trees instead of the forest. Initial un-smoothed results will look like "grass", while too much smoothing will look unrealistically flat (like some of the freq. response graphs in TAS).
Paul
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: