Two-way using Esoteric 7" Woofer & New Esoteric AMT, with First-Order XO?

[stellarelephant]

7" 2-way is only a stretch with a "non-robust" tweeter (like a ribbon, planar, or a too small/high Fs dome).

Those would be some expensive notches.
You know, you can target the same notch freqs. by halving your coil values, and doubling the caps (and probably adj. the resistors).

Great point. Spend $10 less per inductor, and $5 more per cap...saves a few bucks in the end. I tried again and did manage to shrink two of the inductors. And I think I got a smoother result overall. Of course this is all just a game if we don't trust PE's driver response files for the AMT...

[stellarelephant]

It looks like a possibly very nice combo, but you may end up settling for a bit of a dip in the midrange. That's not necessarily a bad thing in my experience.

I would be looking at something closer to a 1 mH inductor for the woofer. This could result in a 3500 Hz crossover with at least a 2 or 3 decibel shallow dip in the midrange. In effect, your bass and highs will be boosted relative to the mids.

Also, I like your second sim without the notch filters much better. The impedance is so much nicer. They look like nice drivers, keep it simple!

Billet, thanks for your input. Can you explain why you'd prefer a smoother impedance curve versus a smoother frequency response?
I'm a sucker for vocal clarity, so I think I'd be annoyed by a midrange dip, BUT I can see how nudging the XO point somewhat lower could be helpful.

[Billet]

I see the uneven impedance curve as a sign of over filtering the signal. I generally don't try to filter every little bump. The smoother impedance curve is also a much easier load for the amp to follow.

[Adam_M]

I haven't played with the sim for that woofer, but it looks like a relatively common profile. Why not try a 2nd order or maaaaybe 3rd electrical on the woofer with a tank cap (and resistor, if the parallel leg of the 2nd order doesn't have enough resistance to stop a dead short to ground at high frequencies) on the first inductor? That should get a lot of the junk you are trying to notch out suppressed. No matter how hard I try for something else, that's almost always the topology I end up with.

[LOUT]

It looks like you don't have any BaffleStepCorrection with that small/0.5mH series inductor. If these aren't going to be in-wall (or huge baffles), you might want some +3-+6db rising bass below 400-800hz depending on the box's front baffle size/shape. I'm guessing you might need something like a 1.5mH-2mH inductor instead of that 0.5mH.

On your cap+coil shunts to ground, a larger cap and smaller inductor (to get the cut centered at the same frequency) will also widen the cut a bit...probably a helpful thing if you're trying to knock down a slightly wider area, but might be a problem if you're trying to make more narrow/surgical cuts. Just something to keep an eye on if you shrink the shunt coils a LOT.

I know it's in the title and everything, but is 1st order XO's one of your top priorities?

[Adam_M]

It looks like you don't have any BaffleStepCorrection with that small/0.5mH series inductor. If these aren't going to be in-wall (or huge baffles), you might want some +3-+6db rising bass below 400-800hz depending on the box's front baffle size/shape. I'm guessing you might need something like a 1.5mH-2mH inductor instead of that 0.5mH.

On your cap+coil shunts to ground, a larger cap and smaller inductor (to get the cut centered at the same frequency) will also widen the cut a bit...probably a helpful thing if you're trying to knock down a slightly wider area, but might be a problem if you're trying to make more narrow/surgical cuts. Just something to keep an eye on if you shrink the shunt coils a LOT.

I know it's in the title and everything, but is 1st order XO's one of your top priorities?

And to the OP, please understand that 1st order electrical is very different than 1st order acoustic. 1st order electrical, at least in my experience, is used in budget builds or the occasional unique situation where it applies - although it never has for one of my builds. 1st order acoustic is a very complex beast that I can't even pretend to understand but it requires drivers with exceedingly wide and smooth bandwidth and often resorts to tricks beyond my knowledge like ladder delay networks.

[stellarelephant]

My gravitation towards first-order (electrical) stems from:
1) Reading mythical reports of uber-realistic spatiality and timbre coming from this topology
2) Seeing in XSim that with first-order, the overall phase shift is minimal AND the drivers track well together
3) Hearing, during my last project, that even the simplest crossover I could get away with compromised some realism versus running full-range.

[fpitas]

Well, most people here will sympathize with your trouble with crossovers. I use LR4, and it works very well, but it took quite a long while to really get it dialed in.

[Adam_M]

My gravitation towards first-order (electrical) stems from:
1) Reading mythical reports of uber-realistic spatiality and timbre coming from this topology
2) Seeing in XSim that with first-order, the overall phase shift is minimal AND the drivers track well together
3) Hearing, during my last project, that even the simplest crossover I could get away with compromised some realism versus running full-range.

At the risk of sounding too much like an objectivist, I can't think of a reason that a first order electrical would sound "uber-realistic", and if there are any benefits, surely all of those notches are defeating the "purpose".

I really can't imagine you can hear the 360 degree phase shift of lr4. I know with certainty I can't. The phase shift is easy enough to deal with by sightly altering the acoustic slopes from textbook lr4, which will be necessary in any xo unless the driver acoustic centers are on the same plane.

I think that nailing the frequency response and dispersion, and limiting distortion would have far more impact than having a single cap or coil on the driver.

It's not that the first order electrical isn't a fun experiment - it is. I'm not sure I'd use the expensive drivers with fairly large breakups for that experiment.

Regarding the crossover altering realism, are you sure the xo was optimal? More of my crossovers than not alter the realism for the worse. Sometimes with some driver combinations, I never got there. I'm quite sure that some of my non-optimal xo's alter realism for the worse.

[Billet]

I like to use first order crossovers also. They just always seem to sound better to me as long as I choose reasonably well behaved drivers. And these two drivers do look pretty well behaved.

[stellarelephant]

Regarding the crossover altering realism, are you sure the xo was optimal? More of my crossovers than not alter the realism for the worse. Sometimes with some driver combinations, I never got there. I'm quite sure that some of my non-optimal xo's alter realism for the worse.

To call myself a rookie would be overstating my credentials. So NO, I am not sure that my last speaker's XO was optimal. I did tweak on it until I got a compromise that I can live with.

I don't see many DIY speaker projects using first order XOs, probably due to some of the drawbacks that have been shared. A few folks swear that when feasible they are the best option, so I felt somewhat excited that I was able to make it work, more or less, with these cool Esoteric drivers, on XSim. Other PE driver combos I've sim'd with have not fared this well.

I'd like to move away from using subs, so for this speaker, I had it my head that I might try a floorstanding transmission line cabinet for the 7" woofer, and MAYBE try to integrate the AMT as a dipole on top, to get a hint of that open baffle magic into the room. That moves the drivers further apart on the baffle, though--another tradeoff, most likely.

Anyway, to tackle a "from scratch" build, using fairly expensive drivers in a somewhat unconventional way, without measurement gear, and relying on dubious tweeter response files, PLUS having very little experience, I feel I would face pretty significant risk of failure. I'm really grateful for all the feedback, which so far I've interpreted as ranging from "Try it--It might sound good," to "Too many caveats--Not worth it."

Also, in between those extremes, I've gotten advice to:
-Use steeper XO slopes
-Keep first-order, but lower the XO point
-Simplify by ditching the notches
-Add a midrange
-Use baffle step correction

Hmm...

[billfitzmaurice]

At the risk of sounding too much like an objectivist, I can't think of a reason that a first order electrical would sound "uber-realistic"

I can't either. The usual reason I see stated for using them is that the fewer components and the less phase shift the better, but when you consider the number of components in the signal chain, starting with the source medium, and then all the electronics between it and the drivers, every one of them contributing to phase shift, the number of components in and phase shift contributed by the crossover is rather piddling. By the same token I'm not a fan of BSC. If the speaker is being used outdoors it would serve a definite purpose, but indoors the wave that wraps around the speaker below the baffle step isn't lost, it's reflected off the wall behind back into the room. That reflected wave will be attenuated compared to the direct wave by dint of the distance traveled, but it's still contributing. That being the case that distance affects the result, so it would have to be taken into consideration. It's a lot easier to just tune the in room response via DSP.

[Billet]

I think a significant key to success in a first order two way project like this is sound of woofer by itself with no crossover. A first order system combines some of the qualities of full range and multi driver systems.

It is a great help if the mid bass driver sounds nice as a full range. A spec sheet with graphs unfortunately does not supply enough information. If the Esoteric woofer has the quality of sounding good on its own, then this project is far more likely to be successful.

Your planned design looks great in the modeling phase. Many of the suggestions probably hinge on the idea of the first order crossover not working as planned for various reasons.

[stellarelephant]

Good point about full-range sound of the woofer being important factor for first order success. To me, the 7" Esoteric woofer response looks sort of like a full ranger with insane bass and a tightly damped high end.

I tried inching up towards a 1mH inductor as a couple folks suggested. 0.7mH is the highest I can go without sucking out the mids or sacrificing the awesome phase agreement. Now the tweeter contributes a bit more of the output through the mids--which definitely helps off axis.

One cool thing in defense of the notches is that all three of them actually improve phase alignment as well as frequency response. See attached for another comparison with and without the notches, as well as predicted response at 45 degrees off axis. I think I will leave it here for now and see if anybody with first-hand experience of these drivers turns up to weigh in on the feasibility of this combo.

[ernperkins]

Noticed that you're not using any z-axis offset (mod delay in XSim). I don't know what it will be for these two drivers and maybe some one else has an estimate. The offset will change the phase response, especially at the crossover frequency. I wouldn't order any crossover components until you can measure and sim with the actual value.

[dcibel]

Looks like the datasheet data is being used for that AMT, it will be miles off. Real measurements always recommended for successful design, especially with decent drivers like these. Odd that you continued designing using manufacturer data after finding some guy on DIYA that measured the AMT and found it to have very sloped response, I'd trust his measurements . The Esoteric AMT response is similar to AMT3-4 but with a bit more extended response on both ends.