Yes, I know that's a good way. It doesn't always tell you tracking though, it can also point to the phase crossing at the crossover, but being wildly different each side of the crossover.
This is what I get with your settings, xo on woofer @250, xo on 10f @350. -.22 delay on 10f. polarity reversed (case D) Gray line null is pos. pol. My peq settings are different so the response is probably different.
Anyways, see how the phase crosses but does not track even though the null is fairly deep and symetrical? And the point of the null is centered on where the phase crosses and not necessarily where the acoustical crossover is.
Your null is certainly at 350 and mines not, but that could be due to peq settings and overall level of the 10f in mine
The placement of the null is based on the FR too. Phase will change with FR and with reflections it's altering the phase. I'm betting that if you took it outside, your null would change too
That would lower it a little, but not much. I have a bit of peq on the 10F and the roll off of the rs225 is smoother in mine. Since yours dips past the crossover that could have some effect on the observed xo point and overall response.
My point was just that with a deep reverse null, it doesn't always point to real great phase outside of the crossover as shown in the graph
My point was just that with a deep reverse null, it doesn't always point to real great phase outside of the crossover as shown in the graph
satx, the reverse null can tell you everything you need to know about phase if you know what to look for. In your simulation you can see the response "flips" at 2khz. This tells us the response is out of phase there.
The only thing the rev null can't tell you is the phase angle which frankly doesn't matter. All we care about is whether summation is maximal for the entire overlap.
Having said the above, getting proper measurements and viewing it in a program like PCD certainly is the easiest way to do it. Looking at the rev null isn't easy sometimes. Especially indoors with such a low XO. Basically impossible in that scenario. So I do agree with you there.
The only thing the rev null can't tell you is the phase angle which frankly doesn't matter. All we care about is whether summation is maximal for the entire overlap.
Having said the above, getting proper measurements and viewing it in a program like PCD certainly is the easiest way to do it. Looking at the rev null isn't easy sometimes. Especially indoors with such a low XO. Basically impossible in that scenario. So I do agree with you there.
Hmm, I should have collected some .FRD files with no XO when I was outdoors. At least would have had some data good out to 10ms (100Hz). I am just starting to use Bwaslo's Xsim software (with Byrtt's help) for passive XO's. I can model some of it in Akabak as an ab initio sim, but can't import an frd and propagate through a theoretical XO model.
Xsim doesn't do active though. PCD will do it.
Yes, always get raw frd data when you're outside. Even active XO'ers should be run through a modeller. You can see things more quickly and easily in software than you can in real time. Then once you have something you can sit down and listen for the problems and correct.
Yes, always get raw frd data when you're outside. Even active XO'ers should be run through a modeller. You can see things more quickly and easily in software than you can in real time. Then once you have something you can sit down and listen for the problems and correct.
See how both drivers cross at 500hz? That means the null should hit the floor right there, but it doesn't. This tells us the phase is not very good around there.
Yes, you're right. I didn't want to get into the width of the null vs xo order and the symmetry.
I've been talking about the cancellation shown with proper polarity and the flip elsewhere when reversed for days, but X's assertion is that if the step is good then the phase must be good regardless of what the FR tells us.
And the graph I posted was just to show that the null may be wide and fairly deep, but you can still have poor phase tracking either side of the crossover, pretty much what you said above if I'm not mistaken.
He is going to really struggle interpreting these indoor measurements with the xo that low. I'm glad someone else said it because X believes that he has the resolution to do this work down there and I've been a bit disagreeable on the point
Quality measurements with high resolution is your best bet, along with modeling to help you along.
What I'm working with is not going to be perfect because I had to make a couple of leaps, trust manufacture's data and mix in a dash of guess work. But, I think it can give you a little insight into the what and why.
Tuxedocivic, what are your thoughts on delaying the woofer to bring the phase in line. Does this work in reality or is this a simulation anomaly? I've never done it and really have no idea. It doesn't seem to make sense to me.
X, you'll also need Jeff Bagby's response modeler with PCD. If you decide to try them, I can help you get started
X, you'll also need Jeff Bagby's response modeler with PCD. If you decide to try them, I can help you get started
I can provide you indoor FRD and ZMA files measured at 0.5m - not sure if that helps any. Not looking forward to setting up outside again.
I still wonder why it is that the good step response seems to be at odds with a good phase response ? Probably indoor reflection as the limitation? Given that, is the SR more indicative of a better XO or good phase?
I still wonder why it is that the good step response seems to be at odds with a good phase response ? Probably indoor reflection as the limitation? Given that, is the SR more indicative of a better XO or good phase?
I guess a lot of people do close mic with woofer to get the FRD data and moderate distance with gating for the tweeter FRD file.
I get a chuckle out of how complicated this is becoming of we are trying to eek perfection put of it. One of the reasons the Fullrange forum folks stay away from crossovers. Here, hoping that a low XO FAST keeps it simple, and it would be unless you start digging into phase and SR.
But if one were to stick with a single full range driver and not FAST, and sort of bass augmenting cabinet like BR or BLH will also not have perfect phase or ST either. Only way is a single driver in a sealed cabinet with DRC.
I get a chuckle out of how complicated this is becoming of we are trying to eek perfection put of it. One of the reasons the Fullrange forum folks stay away from crossovers. Here, hoping that a low XO FAST keeps it simple, and it would be unless you start digging into phase and SR.
But if one were to stick with a single full range driver and not FAST, and sort of bass augmenting cabinet like BR or BLH will also not have perfect phase or ST either. Only way is a single driver in a sealed cabinet with DRC.
Last edited:
Close mic limits high frequency resolution, so even if you're crossing low, the woofer's stop band won't be accurate and neither will the phase.
Most use gated indoor measurements. With a normal two way xo, this is no problem really. Three way and low xo fast is just harder and takes some extra work
This is a complicated hobby, especially at first. My brain hurt for a good year when I started. And that wasn't that long ago really so I still have a lot to learn.
As far as your project being complicated, you started with some tough limitations.
You want a low BW1 xo, but you're using indoor measurements without the necessary resolution to reliably trust what's going on down there. So, you're mostly guessing and then jumping to conclusions when the flawed measurements show something.
You picked an aluminum 8" woofer to use a BW1 crossover on.
Your chosen mid driver and the enclosure it's in don't allow a proper BW1 target slope below the crossover, or really even at the 350 xo.
Notice that many of these mention the first order xo. They're just tough to do right and you need to take special care when selecting drivers and xo points. But you kind of went about it from the oppisite direction- you picked drivers, baffle layout, enclosure alignment and xo point and then are trying to shoehorn a BW1 xo in there.
And you're trying to get a transient perfect, linear phase design here , when maybe you should get frequency response and phase right and then investigate how to refine it to perfection.
I'm not trying to be negatively critical here, I'm just saying that I think this is why this is proving a lot harder than you first expected
I actually think that I'll have better results using the manufacture's data here and modeling the baffle and enclosure effects than I would using the measurements that you have now. They're just not accurate that low.
If you had good outdoor measurements then that would be a different story.
BTW, you're likely to have some floor bounce effect near your xo that I don't know how to model and outdoor measurements won't show.
As far as the step response, I don't really know. I pretty much only know what it's supposed to look like when it's good
I would say that nothing can really be accurately deduced from the measurements you have except higher in frequency.
Gotcha. Sounds like we're on the same page. I probably should have read further back in the thread.
I don't see a problem with delaying a woofer if that's what is required to achieve summation. If the tweeter is lagging phase, then lag the woofer too. Perhaps I'm missing part of the conversation around this?
Step response only tells you time alignment and is quite coars. Phase tells you time as a function of frequency. Phase is a better indicator of summation. You can be 360° out of phase and have perfect summation and bad step response.
Do it right or do it again later. The only way to do it is outdoors or an anechoic chamber (or guess). Back up around 1m or 2m and get ground plane measurements and combine with a ladder measurement. Feel free to do it half way, but I've given up on that.
It's complicated, yes. Most people underestimate the complexity. Just think, an active FAST is about the easiest XO one can design. You're partially making it harder on yourself by using silly first order BW filters, but even so, there's a lot to be done. The FASTs that are usually described around here are fine, but they're not examined down to the nitty gritty. Of course, you can choose to ignore the nitty gritty and enjoy.