Time Align ?

[jeff_free69]

I noticed it's rare to see any designs where the tweeter is backset in relation to the woofer, allowing them to perfectly time aligned.

Is this because ?:
1) It doesn't provide any significant audio benefit
2) Its too hard (the transition between the M & T needs to be a smooth contour to avoid nasty diffraction issues)

I am building an RS150 + RS28 in a PE .25 box. For aesthetic and space limitation reasons, I am mounting the tweeter on top, at least in this first build. The tweeter will be in its own baffle, about 7.5 X 6 with sloping sides which are highly countered with a large roundover. I should be able to locate the drivers pretty close and have thought about sliding the tweeter back an inch or so to align the front of the voice coils. This of course makes a smooth transition difficult, unless I raise the tweeter even higher.

Any thoughts on the tradeoffs of aligning vs not aligning, and if I do, on raising the tweeter to allow better contour of the baffle transition hwo far is too far?

since it's an HT application I'm hoping I can get away with it; If it completely sucks I can always go back to basics and make a new main baffle

[rc white]

The effect of lining up the acoustic centers of the drivers is to remove lobe tilt, the amount of tilt being..

angle = arctan(c/d), where d = center to center distance and c = acoustic center difference.

The actual value of c depends upon frequency but putting c=20mm., d=120mm.
gives a 9.5 degree downward tilting lobe, and is equivalent to 57degrees of phase lag in the high pass, 9.5 degrees being the total tilt assuming that everything else is exactly in phase.
With the usual symmetrical crossover orders the high pass phase lags the low pass and this causes extra tilting, in the case of the third order Butterworth this tilting can often be removed by anti phase connection of the tweeter, many people are reluctant to do this because it has much greater group delay than in phase connection.

If you for instance use a second order low pass and a third order high pass, depending upon the driver roll off characteristics, you can get the high pass to lead, in which case the lobe tilts the other way, and can be enough to completely compensate for acoustic offset. With a low fr tweeter you sometimes need an extra 2nd. order section on the high pass otherwise the phase curves cross over at a frequency low enough to again give the tweeter a net phase lag.
Summing up you can find crossover solutions that make physical offset compensation unnecessary.
rcw

[Deward Hastings]

aligning acoustic centers

rcw sums it up nicely, but there are some other considerations (there is a discussion of this very subject going on now over at the PE board).

The relative acoustic offset of the RS150-RS28 combo is probably closer to 25mm than to 20mm, and your c-to-c will probably be a bit more than 120mm, but the differences are so slight as to be inconsequential.

If these speakers are for center channel or surrounds a slight downward tilting lobe may be desirable, as such speakers are commonly mounted above the listener (or below in the case of the center).

With a LR4 crossover phase tracks at and around the crossover point . . . as a result the crossover is quite forgiving of slight deviations in offset and even in component values. When phase shift is introduced in one signal path to compensate for offset (asymmetric crossover) phase no longer tracks between the two halves . . . to function properly everything must be adjusted so the phase curves cross at the crossover frequency, and everything in the crossover becomes more critical.

Where tweeter offset is accomplished by a setback diffraction can be avoided with copious application of felt or other absorbing material (see Vandersteen speakers for an example). Setback can also be accomplished by sloping the baffle (see Thiel or the Spica T-50).

[jeff_free69]

I'm starting with the cjd's BSC version of the crossover and already have the parts.

https://www.htguide.com/forum/attachment.php?attachmentid=9526

It sounds like I'd need to do a major overhaul for an aligned version, which may not even be worthwhile, if i understand the gist of your responses.

I'm going to build it in a way for minimal distance between drivers and normal alignment. But I should be able then experiment with offsets later. A nice wedge of felt seems like a viable solution...

[rc white]

This is a simulation of the rs150-28 with the crossover you posted..



The tweeter has around 25degrees phase lag at the crossover, the combined output is..



This is not too bad.
From the above applying acoustic center adjustment is necessary with this crossover to reduce lobe tilting, otherwise a fourth order filter on the low pass is indicated.
rcw

[Dennis H]

RC White, what numbers are you using for those models, minimum phase as published by PE or actual measured phase from a common point with the drivers mounted on the baffle? 25 degrees of phase lag sounds about right to delay the tweeter enough to match the woofer with its recessed acoustic center so I question the lobe tilting as well as your summed dB.

Bottom line, I'm skeptical about your conclusions regarding Chris's crossover. He knows what he's doing.

Jeff, build it the way Chris designed it. It's already electrically 'time aligned' with the crossover. You'll just muck it up if you change it.

[Kevin Haskins]

Without getting into specific designs, the amount of time difference you get from typical dome tweeter to midwoofer are pretty minimal. When you are designing with horns you get situations where the physical offset is large and it makes sense to start thinking about delay issues and DSPs.

My view on it from a design standpoint is that you deal with it if possible and don't sweat it if it isn't. Most of the mechanical methods for dealing with it present unsavory choices. I'd rather have the slight physical time alignment issue than either drivers on a slopped baffle or a irregular shaped baffle to try to mechanically align them. After all, they will only be aligned for one point in space anyway since the drivers are separated in space.

The one situation that doesn't present a tradeoff, and one I'm doing in some current designs, is using a waveguide/horn loaded tweeter with a physical offset that matches the midwoofer. The depth of the horn lines up the acoustical centers.

DIYers don't have that choice, except in exceptionally complex situations where they may make their own horn. I only do because I'm tooling both the parts to begin with and aligning them didn't cost me anything, either in terms of money nor design tradeoffs.

[rc white]

Those simulations use the Matlab LTI toolbox Denis and assume minimum phase behavior with well behaved Hurwitz polynomials.

These simulations are admittedly rather basic and since they have no time delay information assume time alignment of both transfer functions.
The rs150 is modeled with its high end cone break up, but the simulation shows the same thing withought it. With a fourth order L-R filter on the low pass instead of the 3rd. Butterworth the simulation shows..



In this the tweeter phase curve has moved around 30degrees to the other side and can compensate for around 5 degrees of lobe steering, the sum of the high and low pass outputs is..



As I said these are fairly basic simulations and I present them as such.
As these simulations indicate that the third order scheme results in around 20 degrees of lobe steering, and the fourth plus third order scheme this reduces to about 5 degrees.I would be interested to see how these simulations correspond with measured data people might have.

I agree Kevin that in general overall phase shifts in crossovers have much more effect upon lobing performance than driver acoustic offsets do, and especially in the commercial world offset compensation is as much to do with marketing hype, and appeal to aesthetics, than any thing else.
rcw

[Dennis H]

Those simulations use the Matlab LTI toolbox Denis and assume minimum phase behavior with well behaved Hurwitz polynomials.

These simulations are admittedly rather basic and since they have no time delay information assume time alignment of both transfer functions.

Thanks, that's what I thought. Chris's design uses real measurements so I'd be careful about criticizing it based on basic sims. As Anderson Cooper says on CNN, "just keeping 'em honest."

[---k---]

I was going to mention that also. CJD takes measurements of the drivers in the real box prior to doing the final crossover. I've seen the phase plots of his crossovers, they usually are really excellent. Maybe we can get him to post one. But, he is really busy with work these days.

[cjd]

Actually, I've not built the TM. So, it's always quite possible something is off.

It is, however, minimum phase modeled based on experience from reverse-engineering the MTM, which I designed using both a full sim process and with real measured data (separately - I wanted to compare and see where the pitfalls were if I could, in the "sim" process).

Besides, everyone around here knows I take so long that lots of delay is added to *all* my filters...

[Dennis H]

It is, however, minimum phase modeled based on experience from reverse-engineering the MTM

But you modeled the acoustic center of the woofer behind the tweeter, right? How much Z offset does it work out to be for the RS150?

[cjd]

Yeah, I don't remember if it's 18mm or 22mm. But unless I was waaaaay off my game somehow I would have used one of those two values z-axis offset.

[jeff_free69]

Actually, I've not built the TM. So, it's always quite possible something is off.

It is, however, minimum phase modeled based on experience from reverse-engineering the MTM, which I designed using both a full sim process and with real measured data (separately - I wanted to compare and see where the pitfalls were if I could, in the "sim" process).

Besides, everyone around here knows I take so long that lots of delay is added to *all* my filters...

Glad to see you're still around - I've got a question on this XO -

My location will be somewhat closer to the wall than your full BSC design might call for. (I chose BSC over nonBSC because my Zaph Sr71's (also BSC) sound fine there with just a slight tweak up in the tweeter level - changed an R from 4.1 to 3.7). I will probably want want to try the same with your design, what would you suggest?

the 4 ohm tweeter has the pad wired in, giving a net resistance of
4 + (30 in parallel with 4) = 4 +3.5... = 7.5...

can i use an 8 ohm LPAD instead ?
is there even such a thing as a 4 ohm LPAD?
Are the some other hard wired R values you might suggest as a starting point?