reverse null

[isaeagle4031]

What does a reverse null supposed to tell me in a crossover? How important is it?

I'm playing around with a 2way crossover 2nd order on the woofers, 3rd on the tweeter with a parrallel RC on the tweet. The RN is actually rather shallow at the crossover point

[mzisserson]

What does a reverse null supposed to tell me in a crossover? How important is it?

I'm playing around with a 2way crossover 2nd order on the woofers, 3rd on the tweeter with a parrallel RC on the tweet. The RN is actually rather shallow at the crossover point

RN matters more with symmetrical x-over slopes, espically 2nd to 2nd and 4th to 4th. Asymmetrical slopes may not always yield perfect phase, but mostly better power response and flexibility to use drivers in their better, respective bandpasses. As always its a compromise.

[isaeagle4031]

Aahh OK. Maybe I'm fretting too much then. It seems the more I work on the tweeter to tame its rising response and try to hit my target xover point the less the null became. It looks fairly smooth and is best with the tweet hooked up as reversed. The speaks in question are the Aura NS6's x2 and the Foster 025 horn.

BTW Miz any progress on the joint project?

[ReissM]

It's a quick way to tell if the woofer and tweeter are in phase at the XO freq. By intentionally reversing the phase, the two should then cancel and you should see a very deep "null". Approx -30 to -40 dB notch.

In the real world the phases don't always line up and you get a mish-mash result where there might be a 10 dB null when reversed. It usually indicates that the two drivers are not exactly in phase (or 180 degrees when reverse polarity) but instead, they are some other amount.

Often times when the phases don't line up, the two drivers will add a little bit below XO freq, and cancel above the XO freq. (Not a good thing!)

What to do?

Well, try altering your slopes (which affects the phase of the driver) to get a better match at the XO freq.

For example: you are using 2nd LP on your woofer and a 3rd order HP on the tweeter. Try 2nd order HP with 2nd order LP and see if your null gets deeper and more symmetric??? If not, try 3rd order LP with your 3rd order HP.

You can experiment with asymmetrical designs as well... no reason they have to have the same (electrical) slope. Remember, the filter will combine with the driver and the resulting acoustic slope may be quite different than your electrical XO slope.

At this point I'll get up on a soapbox and give my pitch... Passive Crossover Designer is perfect for these type of experiments. On one of my previous projects I simulated every possible combination of 1st, 2nd, 3rd, 4th order LP with every combo of 1st, 2nd, 3rd, 4th HP and found that for my particular drivers, 2nd order LP with 2nd order HP was just the ticket. Every driver combo will be different.

The project before that one worked out well when I used 3rd order HP and a 3rd order LP (with a zobel.) Once you get comfortable with PCD it will save you weeks of trial and error the old fashioned way. You can see the phase change in your simulations and determine very quickly that the phase of the two drivers are not lining up very well and will therefore not sum together very clean. (The reverse phase null will be wishy-washy and not a beautiful -35dB null.)

Keep working at it, there is usually a solution that allows the phase to line up... and thus, the reverse null will give a nice, deep notch at XO freq.

[Deward Hastings]

The "reverse null" is an aggregate of a number of things . . . how good your phase tracking is around crossover, for one, and lobe aiming (the reverse null is sort of the inverse of the primary lobe, unless the crossover is *really* messed up) for another. Looking at just one (listening) axis doesn't tell the whole story, since most of what we hear is reflected. Lack of attention to the vertical axis often leads to bad behavior of the first ceiling reflection, for example, and significant variation of the speaker's "sound" depending on listening distance and/or ceiling height.

[duanebro]

PCD is very fun to play with isn't it? The r. n. also shows where the xo point is. Phase will change as the ratio of the distance from the drivers to the Mic changes. If you change the listening height in pcd you will see the null change also.

I have a thread that I started last night you might like to keep an eye on. Simple 2-way design with phase and xo issues. I will be posting more later on design issues with lots of pcd screen shots.

[mzisserson]

Aahh OK. Maybe I'm fretting too much then. It seems the more I work on the tweeter to tame its rising response and try to hit my target xover point the less the null became. It looks fairly smooth and is best with the tweet hooked up as reversed. The speaks in question are the Aura NS6's x2 and the Foster 025 horn.

BTW Miz any progress on the joint project?

It's my kids'4 th birthday. Car filled with my little poly ballons as we speak I will likely get to it this week.

Though this is what you are working on. It's good so you can compare.

[Wolf]

Just as an arbitrary reference point, I try to get -15dB from reference minimum. Most of the lower order designs also allow this much depth.

As ReissM was saying, the null can be 'asymmetrical' and shift to left or right and not be symmetrical. The first thing I try is to alter the Q of the filter, by adjusting the cap one direction and the coil the other, and vice versa if it gets worse. Sometimes this will make a more symmetrical null, but cause a hole in the response, which means you have to bring the xover knees closer together to get good summation.

Most of my designs on flat baffles use a one-up higher order on tweeters than woofers, or maybe the opposite, depending on what I have to do to the woofer to make it behave.

If the tweeter has a recessed mount or the midbass is a shallower profile equal orders become more typical.

All this said- I watch the phase more than the reverse null.

Later,
Wolf