Why is an aluminum cone driver bad for BW1? Breakup? I am not sure a paper driver is any better band width wise from the plots I have seen.
I think I do enjoy the sound, as imperfect as the phase may be. As SATX says, the BW1 is challenging and perhaps stepping back to do the design with a mid alignment that can support the BW1. Or heck, maybe do the fancy B&O 3-way "LR2 hole-filler" with BW1 on bass and tweet for a phase-perfect box by slapping a tweeter on the top and getting kicked out of the Full Range forum
I am learning a lot though and seeing that XO's are tricky business even with miniDSP.
Appreciate all the help I have been getting.
I am only using these silly BW1's as they seem to be required to get perfect SR? Recall my initial design on post 1 in Akabak assumed LR4 at 500Hz! Is there a way to get perfect SR with higher order filters?
I think I do enjoy the sound, as imperfect as the phase may be. As SATX says, the BW1 is challenging and perhaps stepping back to do the design with a mid alignment that can support the BW1. Or heck, maybe do the fancy B&O 3-way "LR2 hole-filler" with BW1 on bass and tweet for a phase-perfect box by slapping a tweeter on the top and getting kicked out of the Full Range forum
I am learning a lot though and seeing that XO's are tricky business even with miniDSP.
Appreciate all the help I have been getting.
I am only using these silly BW1's as they seem to be required to get perfect SR? Recall my initial design on post 1 in Akabak assumed LR4 at 500Hz! Is there a way to get perfect SR with higher order filters?
Last edited:
Metal cones have a higher break up peak and also high ringing (CSD) compared to paper. Otherwise you'd be correct that the bandwidths are similar. Typcially, an aluminum cone driver needs to be crushed about 12db at breakup more than a similar paper cone driver. At 6db/oct that means 2 oct lower than you are now
Very likely. So do enjoy. Only suggestion I would have is try an LR2 and let go of the transient perfect fantasy. If the LR2 sounds better and you "enjoy the sound" more, then use that XO configuration. Even LR4. Dead simple really. Hardly have to do anything except dial in the settings.
I'd answer that question with a rhetorical question: Why get perfect step response?
I did see your OP with LR4 and didn't bother asking why go to BW1. Seems to me to be a step backwards. But everyone has their own goals so I didn't mention it.
To answer your question directly, with active you may be able to do LR2 and then delay the woofer until you have perfect step response and perfect phase summation. But I don't know, never been bothered to consider it. I'd rather have the woofer breakup buried in a grave and have a 1/4 millisecond of step response offset than have perfect step response and hear that awful sound of woofer break up.
No, I don't think you are. It's just that the woofer is already physically behind the mid. So delaying it further would seem to cause the time arrival to be even further off. But, it brings the phase in line and I'm assuming that it brings it around a full 360 so it's right again, but my mind is struggling to visualize this.
Delaying the woofer was the only way to get phase correct, thus proper summation in my sim with these BW1 filters
Posted while I posted
Yep, we do. I've been trying to help meet his goals anyways, but it's damn hard. LR2 would have been just about finished 20 minutes after reading the first post.
What he said
That's why I have been trying to get you to take a good look at that breakup for a while now. you're only down about 10-15db based on your measurements. Should be -40db at least with that driver. That's why I recommended to Byrtt earlier that the brother of the 10f, Scanspeak 22W/8534G00 would be a better choice. The dayton's great, especially for the money, but just like even the most expensive drivers it has to be used within it's limits to perform
I think I do enjoy the sound, as imperfect as the phase may be. As SATX says, the BW1 is challenging and perhaps stepping back to do the design with a mid alignment that can support the BW1. Or heck, maybe do the fancy B&O 3-way "LR2 hole-filler" with BW1 on bass and tweet for a phase-perfect box by slapping a tweeter on the top and getting kicked out of the Full Range forum
If you enjoy it you enjoy it! If you hated it then there'd be a problem. I personally think you could enjoy it as much or more with a bit more perfection regarding distortion, breakup suppression, ringing, phase and frequency response.
I don't think you should slap a tweeter on it per say. I prefer the dispersion and shimmer of a real tweeter personally, but if that's not your goal then you shouldn't do it. Nothing wrong with running a driver like the 10f full range on the top end.
You are, so it's worth it. even if you eventually have to compromise a bit to really be happpy with it. you're learning a lot now by making some mistakes.
If you got it right in the first 10 minutes then you wouldn't have learned much.
I made a lot of the mistakes that you have. Mostly picking a cool driver and trying to force it to do what I want. And trying to hit slopes and xo points that just didn't want to happen.
Anyways, no worries, just keep experimenting....but get some good measurements first; should be easier now that the amp and xo are in a box.
Ok I see what the issue is. What is being missed is what the XO filter does to phase, thus acoustic time. In PCD, change the low pass filter from say 400 to 200hz. See what the phase does. Now add delay. See what the phase does. The phase keeps moving left, or from positive to negative.
Now do that to the tweeter except with the high pass filter. Maybe change from 200 to 400. Now try adding delay. The high pass is moving the phase from negative to positive yet the delay is changing phase from positive to negative. So really it's not all about where the drivers phyically are. It's also about their phase angle (natural) and the filter poles.
In PCD you can also play with the z offset to further see what happens to phase. This is essentially identical to delay, just in distance rather than time.
Ok I see what you mean. I could see why it worked in PCD, but I wasn't quite grasping the differences in phase and time once put into the real world.
I don't think he's tried it yet to know for sure if the sim's accurate as far that goes.
It seems that you guys are concerned that the woofer is not down low enough in the stop band. Well that won't get better unless I go to -12 or -24dB per octave filters. But looking at my response curve it is within +/-2dB (above XO) so reasonably flat and HD plot, albeit at low SPL doesn't t show any distortion that appreciable. I recall an earlier woofer only HD plot showing essentially no contribution from the woofer at the breakup frequencies. So it seems that the XO is doing its job from separating the drivers even though a lower order filter is used. Acoustically it is ending up around -12dB/octave.
I've been saying this since I first started contributing. And you've been saying this^ since before even then, so..... I guess we just have differing opinions.
I did post this yesterday or the day before. Not sure if it works, but maybe.
"Another thing you might try, is to target a BW1 on the woofer for 2-3 octaves past the crossover and then roll it off much steeper in the breakup region. Sims show it cleaning up the little phase problems in the treble and it starts getting the breakup down where it should be.
This may or may not effect the SR, but it should help if you find the breakup objectionable. Try listening to the RS225 on it's own with the xo attached. This'll help you pinpoint issues with the sound.
I don't see why you can't achieve the BW1 target slope on the dayton if that's what you want.
This is what I got with basically the same settings as you when you measured the woofer xo acoustically at 900ish and it matched your measurements perfectly so you should be able to match my sim without much problem.
Once again, and I'm not the first to mention it, you can't really get good distortion measurements at 82db @.5 meters. You're in the mid 70's at 1 meter. You should try 96db at .5 meters for the minimum. Or adjust for around 90db at 1 meter, but remember that if you measure distortion nearfield, the numbers probably won't be accurate higher up where we're interested in.
Also, try a CSD to look for ringing.
If you haven't seen this, it might be a good read for X. Goes over the pros and cons, difficulties of 1st order crossovers and also has a number of links and sound clips.
Siri's Killer Note
Siri's Killer Note
Thanks. Byrtt or Wesayso pointed out Siri's killer note on TG's site earlier in this thread.
Does anyone know what kind of xo the Dunlavy XC-IV uses because it seems to have perfect SR and time coherence.
Dunlavy Audio Laboratories SC-IV loudspeaker Measurements | Stereophile.com
Looks like what I am trying to make is like this but on a smaller scale.
Pretty sure Dunlavy used first order, time aligned. Always thought this was one of the smartest commercial designs.
They used to be commonly used for mastering monitors (different than mix monitors).
I vaguely recall Dunlavy hung out on forums (maybe the basslist?) occasionally back in the day.
They used to be commonly used for mastering monitors (different than mix monitors).
I vaguely recall Dunlavy hung out on forums (maybe the basslist?) occasionally back in the day.
"Generically, the rather complex crossover falls into the minimum-phase, first-order (6dB/octave) category, but Dunlavy says it cannot be accurately described in terms of conventional slopes and crossover points. There's also a network to notch out the tweeter's resonant frequency and conjugate load matching to produce an easy impedance load."
I guess it really doesn't matter what classification we put the xo into, the goal, for me, is time coherence and whether we achieve that with first order or fifteenth order doesn't matter. From my experience so far, I can only achieve it when I use low order slope filters. I might have to redo my Trynergy to be time coherent as well when I get a chance. It has very similar xO point at about 400Hz between the woofers and the full range driver.
I am back to LR4 at 400Hz and listening. It can handle more power, sounds cleaner, less distortion. But, the cue ball shot on Last Payday doesn't sound as convincing. Just that mainly. To a certain extent the phantom imaging on the BW1 at 350Hz is better. More full range-like.
Last edited:
The phase is a function of amplitude (FR), thus you can predict phase response visually from the shape of both drivers' responses during roll-off.
The good thing about passive crossover is that you can do an "acrobatic" improvisation with crossover components to shape the roll-off responses.
With textbook filter (such as DSP without programming feature) you will rely on the natural roll-off of the drivers. Here, RS225 is not suitable, not only because it has "wiggles" at roll-off response, but aluminum cone has harsh breakup sound (which you can not see from casual measurement).
RS225 is more suitable with complex and steep crossover. But they have more problems to solve. The problem that does not exist in full range approach is of course the phase related issue. This is confusing as we have too many phase-related "variables", and most people just cannot hear anything so they just "guess"...