L-R 12db/oct tweeters reversed or not?

[kscharf]

I've read in various books and on line that the 2nd order L-R crossover requires the tweeter be wired out of phase with the woofer. Yet, I've seen some speaker projects on the web where this was NOT done (mostly M-T-M designs, but not always). Which way is correct?
I also thought that in M-T-M designs either the low pass, high pass, or both were usually 3rd order networks. What's correct here?

[Jonathan Bright]

I think one of the reasons that you haven't got many replies is that the topic is complex and your question is a bit too big or general. In one sense you're right about phase reversal in second order networks. (Although I'm not up to speed on L-R networks as they are a separate family). But on the other hand people use everyything from 1st to 4th order (And more!) with each side proposing good arguments in favour of their position.
Can I suggest you put "crossover networks" in Google and chase up the various tutorials that appear? I've just done that for my own information and about 60% of the sites are helpful.
This area is just complicated and people who have some experience will be reluctant to offer an opinion until your question is a bit more specific. Even then, while it is generally correct to reverse polarity in the second order case, you may decide not to for a speciifc pair of drivers. And that is the sort of thing running through peoples minds as they address your post. And again, I think some of the MTM designs are very specific. e.g. the D'Apolito designs are made to have a defined response and I recall the c/over in that case in not conventional.Sorry I can't be more informative......but that's the situation....

[kscharf]

I agree that the topic of crossovers takes in quite a bit. I have looked at the tutorials on line, and most agree that in the case of 2nd order networks it is necessary to invert the phase of the tweeters to avoid a null condtion. Of course, the mounting position of the speakers might change this so the physical layout of the drivers is part of the equation.

I guess I should re-phrase the question to "what worked for you in your project", and "is there anything special about your design?" I'm sure many speaker builders on this forum have found cases where they had to throw the textbook away to come up with something that sounded good for them. Maybe they had some special drivers, crazy room layout, or unique source material.

I've built three 2 way boxes so far (a pair for left/right) and just added a center channel. The L/R boxes use different drivers than the center because they were built some 13 years ago and the same drivers are no longer made. The old boxes have 6.5" eclipse woofers and the new one a 6.5" vifa. The tweeters in all three are vifa's but the one I used back then also is no longer in production so I used a more current, but very similar model. I don't remember the exact model numbers off the top of my head, but can post them later this evening. Cross over frequency is about 3khz, and I have a zobel network on the woofers. The crossovers are 2nd order L-R and I do have the tweeters wired with the phase reversed. Tweeters are flush mounted (so are woofers) so the top of the baskets are level with the baffle board. They seem to sound ok, but the new speaker is still having it's drivers broken in. (It sounds a little 'fuller' each time I listen)

[Jonathan Bright]

Are some of the confusion is cleared up! I wasn't sure whether you meant Linkwitz-Riley or left -right but I assume that you used different notation. (ie L-R ....L/R) Look it sounds like you know what your doing. I'm not familar with those drivers unfortunately..

[MJK]

Which way is correct?

Both. The textbook solution only assumes the driver behaves as a resistor. In real life the drivers have magnitude and phase as well as being in different positions on the baffle. You have to design with more data then a textbook solution. I have used a second order crossover several times on woofer-tweeter two ways and have never wired the drivers out of phase. Others have done the opposite for thier combination of drivers nd geometry. Unfortunately, it is just not that simple for one rule to be correct all of the time.

[Jay_WJ]

I think one of the reasons that you haven't got many replies is that the topic is complex and your question is a bit too big or general.

Another (and more appropriate) reason why expert people do not reply is that this question is too often asked to be answered each time.


In a L-R 2nd order (acoustic 2nd order, not electrical 2nd order) XO design, the tweeter polarity should be reversed for correct phase alignment. But a real acoustic 2nd order is rare. An electrical 2nd order network, in most cases, results in approximately 3rd or 4th order slope due to natural roll-off of driver FR. This is why you see tweeter polarity not reversed in most designs with an electrical 2nd order network. I'm not 100% sure, but your XO will also most likely be "in phase" when the tweeter polarity is not reversed.

[Svante]

The branches of a 2nd order Linkwitz-Riley crossover sum to magnitude 1 only if one of the branches is phase-inverted. This means, if the drivers are ideal a bit beyond the crossover frequency, that one of them should be reversed. If it is not reversed, there will be a big dip at the crossover frequency.

In real life, the drivers add a phase shift. This makes the answer less clear.

This is why there is a lot to gain from fiddling a bit with the crossover. The crossover is a great place to put compensations for non-ideal behaviours of the drivers. Typically, this results in crossovers that are not textbook Linkwitz-Riley, electrically, but something else that makes the whole design behave well.

[spizznel]

Originally Posted by kscharf

I've read in various books and on line that the 2nd order L-R crossover requires the tweeter be wired out of phase with the woofer. Yet, I've seen some speaker projects on the web where this was NOT done (mostly M-T-M designs, but not always). Which way is correct?
I also thought that in M-T-M designs either the low pass, high pass, or both were usually 3rd order networks. What's correct here?

2nd order crossovers reverse phase 180 degrees and I have built several for PA systems. I wire it (high pass reverse) which sounds better to me, but there are circumstances where it differs. What I do is put a DPDT phase reverse switch on the high frequency driver so you can toggle it on the back of the cabinet. I have used it in different positions at different venues.

[sreten]

Originally Posted by kscharf

I've read in various books and on line that the 2nd order L-R crossover requires the tweeter be wired out of phase with the woofer. Yet, I've seen some speaker projects on the web where this was NOT done (mostly M-T-M designs, but not always). Which way is correct?
I also thought that in M-T-M designs either the low pass, high pass, or both were usually 3rd order networks. What's correct here?

Hi,

Many crossovers use 2nd order electrical slopes with the drivers responses
to end up with 4th order acoustic crossover functions, for these you do not
reverse the drivers polarity, e.g. : Zaph|Audio - ZA-SR71

rgds, sreten.

[oszkar]

Hello!
My question is, do i need phase inversion for 2 tweeters mounted on each crossover, 2 basses and 2 midrages 120w each minimal, in a 3 way system on the crossovers BRAUN L630!