Reversing polarity on 1st order Butterworth

[rmjeeves]

Hello Guys,

I just put together a pair of bookshelf. I am using a 1st order butterworth. When I first put it together, the speaker sounded good but I notice the sound was way to chirpy. I guess the tweeter was way to active. The crossover incude Lpad and Zobel. Just to experiment, I reversed the polarity on the tweeter and suddenly the speakers sounded awesome. Here is where I am confused. Generally, it is not required to reverse polarity on 1st order butterworth. Now I tried and it sound good. Why? Is this going to damage my drivers in anyways. They say only 2nd order LR needs reversing and not butterworth. Trying understand why speakers are sound so much better and will this damange my driver at also. Also will it be ok if I keep them reversed or it is not advisable?

Thanks for the insights.

[wintermute]

Hi I had the exact same thing with my tweeters. When I did a measurement I had a deep null when the tweeters were wired the way I thought that they should have been (ie normal polarity).

What I have discovered is that my tweeters when measured raw are 90 deg out with my mid woofers (also measured raw). The addition of the single cap brings this to 180 degrees difference hence the need to reverse (at least in my case).

Changing the polarity will not damage the drivers, the reason it sounds better (assuming what I said above is the case) is because both speakers are in phase at the crossover frequency, if they are out of phase by 180 degrees sound around the crossover frequency is canceled out.

Tony.

[T101]

With any crossover, you should be after 90 degree phase difference or in some cases 270 degree phase difference.

At 0-degree phase difference, you get the signals summed and thus peaking at the crossover frequency and at 180 degree, you get annihilation, thus a dip.

It is tricky, because you can have both simultaneously, but at different angles vertically in front of the speakers.

Generally, first order crossovers have driver phases always parallel to each other at ~270 degrees out of phase relative to each other. With the reversed polarity, now you have added 180 degrees of phase lag to the tweeter and it is now 450 degrees.

What happened? Well, you changed the focus of your speakers. In other words, you acoustically rotated your speakers upside down.

Try: Xdir
Enter your driver center-to-center distance and then try 270 and 450 degrees of phase difference. - Nothing changes, only it's upside down

Now, read that post: Can you have sparkling treble but without sibilance and you'll get the whole picture

Have in mind that the phase induced anomalies are only around crossover frequency, there are other anomalies such as sibilance which occur due to other reasons - see the second link again.

In essence with a woofer, sometimes you can get 18db/oct with just a first order 6db/orc crossover, the same does not apply to highpass crossovers for tweeters. It's the opposite, first order can mean no crossover at all down to tweeter resonance frequency.

Edit: my post is about electrical signal phase and not accounting for the driver induced phase lag. If you sum my post and the above one, you'll get the wholle picture.

[AllenB]

Originally Posted by T101

With any crossover, you should be after 90 degree phase difference or in some cases 270 degree phase difference.

At 0-degree phase difference, you get the signals summed and thus peaking at the crossover frequency

Your suggestion will permit a flat on-axis response along with a flat power response but a 0 degree design, which can be adjusted for a flat on axis response, gives better vertical lobing. Which to use will depend on the design.

[rabbitz]

This is common and in the 2-way ref thread, 2 out of 3 of the suggested 1st order series crossovers required -ve polarity on the tweeter. You have to think in acoustic slopes and not electrical so you have to consider the driver roll off as well as the filter transfer function.

Adding a zobel will in effect increase the slope of the crossover so you can end up with a pseudo 2nd order electrical. In the graphs, the LH shows 1st order electrical on a woofer and the RH shows the increase in slope when a zobel is added. Playing with the zobel cap size is a good shaping tool as well as getting better phase at the crossover point. Rarely have I ever used a textbook zobel.