Physical alinement of drivers

[maxwedge]

This question has been on my mind for a while and maybe I could get some of your opinion's. I know we try and align these horizontaly, in general, but what is the best or proper location? Magnet to magnet? Top plate on a speaker to surface (can't remember proper name) that the diaphram is attached, on a compression driver. What about ring radiators?

Any opinions are apprecated.

Scott

[Guenter]

I've thought about this and my take, without proof, is that for tweeters alignment is not critical. The frequency range over which there might be phase anomalies is relatively narrow and inconsequential (I'm thinking, for example the 075 crossed over at 7kHz). In contrast, for the crossover at - say - 500 Hz from woofer to midrange the consequent phase anomaly because of the diaphragm displacement could be very serious and noticeable. Particularly I'm thinking of the folks who use the big horns on compression drivers that often leave the driver way behind the woofer with lots of delay. Possibly this is one reason JBL used to use the small horn and slant plates on their monitors. re finding out a correct placement I believe you basically just have to measure the impulse response to determine the relative position (ie arrival times) of the source since horns have an effect on the speed of sound etc. Nevertheless, I've got a feeling that placing the magnets at similar positions is probably 95% correct.

- interesting topic

[Ian Mackenzie]

Time displacement in distance is a variable issue...the ear is more susceptible to at absolute delay at 2kherts than other frequency...the Blauert curves show about 0.8 milliseconds at 2 kherts, nearly 2 milliseconds at 8 khertz & 1.0 khertz and 3 milliseconds at 500 khertz.

Dramatic displacement more pronounced with gentle crossover slopes like 6 db per actave than say 12 or 18 db, the reason is in the area of overlap with longer delays you get a comb filter effect or ripples in the response.

If you align the voice coils that will be near enough..the acoustic centre of a given driver varies with frequency as the phase response of a driver varies through its bandwidth.

If the voice coils closely aligned (with 1-2 inches) the only variation with the the angular geomtry of the vertical axis lobe and this is seen as a tilt in the vertical polar response in the region of crossover overlap.

Regards the distance of displacement sound travels at 1130 feet per second (344metres a second ) (roughly) so I millsecond is .34 metres or 1.13 feet for you Damn Yankees

I have had time displacement issue my self earlier with big horns and I would agree with the above Blauert studies..under 1 Ms except for 2 kherts is a bit of a wank, particularly with steeper crossover slopes....the 4343 falls within the curves, the 4333 does not.

Ian

[maxwedge]

So, I guess we should try and align the voice coils for best alignment and not nessesarly the magnets, because the voice coil in a compression driver is at the rear edge of the magnet. I've got 2ms of delay, for each driver, built into my crossovers but I'd rather get them as close as possible before using the delay.

Now what about crossing over between a horn and a UF like a 2404 or 2405, where the ring radiator is mounted on the baffle. The UF is going to be approx 1 ft (30.5cm??) or more in front of the horn. Is this a big problem or because of the hi crossover frequency's it may not be a big deal?

Oh, I don't like the Yankee's either.....I'm more into cycling (pedal power!)

[Tom Loizeaux]

If you go to the Rane web site and look up the tech papers on their active crossovers, you'll find a chart that gives the time delay relative to displacement distances.
I had a good time clicking in different frequency crossover points and then dialing in the delay until I got a time-corrected response. One of the benifits of using a sophisticated active crossover, especially when crossing between a baffle mounted woofer and a compression driver. I used a sweep generator as it was easier to hear the out-of-phase signals come together when I was sending just that frequency.

Tom

[Robh3606]

Now what about crossing over between a horn and a UF like a 2404 or 2405, where the ring radiator is mounted on the baffle. The UF is going to be approx 1 ft (30.5cm??) or more in front of the horn. Is this a big problem or because of the hi crossover frequency's it may not be a big deal?

Sure is look at the wavelength at say 8K which is 1.7 inches. You may be in phase but you will not be "Time aligned" without delay on the UHF.

I just put them in a vertical array use 24db L/R and not worry. I don't use any of the larger horns where the delay would be significant.

Rob

[Ian Mackenzie]

Tom,

I would use a steep slope on the horn and UHF radiatior, if the distance is only 1 ft dont stress, more than 2 feet it will become audible in time program material..but not movies.

Ian

[Mr. Widget]

My experience with driver spacing (front to back relative to the baffle) is that to my ears it is not that critical. It is quite possible that I am more bothered by other anomalies while others are disturbed by time/phase issues, but in my experiments, I have not been able to conclusively say that alignment is worth the bother. Certainly you can hear the effects of early reflections much more clearly than time/phase problems, so if you mount the UHF driver 12" back into a tunnel or behind a large mid horn this causes greater problems than it fixes. Using digital or analog delay solves this problem, but I really can't say that I "hear" an improvement either.

Two issues to think about if you want to try time/phase alignment. Simply aligning voicecoils isn't enough as some drivers' acoustic center is in front or behind it's voicecoil. The second issue is they need to be aligned relative to the listener's ears. This means they should be aligned in an arc with the radius of the distance between the listener's ears and the acoustic centers of the drivers as shown in the drawing. This speaker is obviously not time/phase aligned.

Widget

[maxwedge]

Ok, so I guess there's not a simple rule of thumb and it all boils down to a measurement with a RTA and where the mic is placed. I guess I need to play around with my RTA some more and try out the sweep function as Tom suggested.

I do use Rane crossovers but I was having a hard time seeing the difference when testing with pink noise, so I pretty much just used Rane's charts they have. One thing Rane didn't do, is to put any delay for the UHF driver....maybe because it would need too much delay at 3k-7k??

And I won't be placing my UHF driver's back near the compression driver's (I was thinking of doing this, lol) because of early reflection issues.....thanks Mr. Widget.

Thanks for the great info guys, you are way above me on this stuff, but I'm trying to understand.

[John Y.]

Now we see the synergism of the Altec Voice of the Theater approach (later incorporated into the JBL look-a-likes). Mr. Widget had a discussion about time alignment earlier when I questioned the obvious implementation of the Hartsfield. I own a Hartsfield and, to my ears, sounds cohesive.

My current home theater set up has 4648's as the mains with the 2440's on 537/509 HF horns crossing at 500Hz, 12 db passive. For movies it is great. I do intend to physically time-align the mains when I later recess them into the rear wall. For now, though, such alignment out in the room would not be aesthetically pleasing.

I don't really listen too much to music on this system preferring, instead, to use my Theils and CJ tube electronics for two channel music. For the SACD and DVDA, I do like the home theater setup, even though the forum will groan over my use of L100's for center and surround. Sounds pretty good to me.

John Y.

[Guenter]

Well, the consensus appears to be pretty much as I assumed.
- don't worry about the tweeter's allignment
- measure the impulse response of the system to allign mid/bass
- although not mentioned before; there is also no issue with the
subwoofer alignment which is very likely to be off considerably.
However, with it we are simply after a uniform Power response

Now, what I would like to ask Widget, he has in his post above not adressed this (nor in the past posts). How does one deal with the large horns - like the diffraction horn - in practice. that is, stick the horn wayyyy out in front or have the driver stick wayyy out in the back?

[maxwedge]

I say stick the horn as far forward as reasonable and time align the rest of the distance.

[Mr. Widget]

Originally posted by Guenter

Now, what I would like to ask Widget, he has in his post above not adressed this (nor in the past posts). How does one deal with the large horns - like the diffraction horn - in practice. that is, stick the horn wayyyy out in front or have the driver stick wayyy out in the back?

Yes, I have pushed the horn way forward to try and align it, but it didn't make the system sound more coherent or improve the imaging. It did get in the way (reflection wise) of the other drivers. While I am sure the system's impulse response looks frightening on a scope... come on over, this system reproduces a snare or tom like few others.

My current set up is still in the experimenter box state, but I am about to commit to real cabinets. Here is the system as I currently listen to it. The side view shows the actual alignment I am using. The 2405 VC is 3.5" in front of the 2202 VC and the 2441 VC is 8.75" behind the 2202 VC. The offset between the 2441 and the 2405 is a full 12.25"! This is certainly not a phase coherent design!!! It sounds quite good and laid out in the room as depicted in the next drawing, the imaging is extremely good.

The crossover frequencies are 1200Hz and 10KHz. (There is also an active crossover at 290Hz to the twin 2235s.)

Widget

[Mr. Widget]

This is the actual layout of the room. The arcs showing the driver offsets in the previous drawing were derived based on the mid horn being at ear level and set up as shown here.

Since these horns are a wide dispersion design, having the walls lined with moderately high sound absorbing drapes helps enormously in the imaging department. If I open the drapes revealing the reflective walls, the imaging goes to hell.

[Ian Mackenzie]

Nice scene Widget....must hear it go.

Had a similar project here, but the mid cone was still displaced some distance ...13 inches ...as was using the 1 ~ 2 inch adapter for the 2397 and the slot more like 22 inches.

It had great dynamics but the depth and imaging shot to hell...now I know why.

I Actually time aligned it once with a digital delay effects box in quad amp mode...wow that made a difference......but was not practical at the time.

Ian