Help with 2-way crossover for TPL-150H and Faital M5N12

It turns out my original plan for a 4-way active, time-aligned, project needs to be scaled down to 3-way active (still time-aligned) plus a passive crossover between midrange and tweeter at 2500Hz. But alas, I have never built a passive crossover! I would very much appreciate your input!!

The active crossovers will be very steep and easy to experiment with as they are software based.

I guess the first decision needs to be 1st or 2nd order passive xo. I understand 1st order xo are in-phase, so would be a nice match with the active xo. And they seem easier to build. Seems the big downside is the strain they place on the drivers.

Can my drivers handle 1st order at 2500Hz xo point?
  • Beyma recommends the TPL-150H to be crossed over at 1000Hz or higher and at minimum 12 dB/octave.
    • Using 12 dB/oct and crossing at 1000Hz, one would get -12dB at 500Hz.
    • If using a 6 dB/oct and crossing at 2500Hz, I would get -6dB at 1250Hz, -12dB at 625Hz. Is this safe? The speakers are to be used in a 25m2 living room, so I will never ever deliver the rated 80W (AES) into these.
  • Faital recommends the M5N12-80 up to 6300Hz. The frequency response and impedance show what looks like cone breakup starting at 5kHz. FWIW, these drivers will be horn-loaded and used from 400Hz.
Should I go for 1st or 2nd order?

Your guidance will be greatly appreciated!
 
The TDP-150H is almost a perfect 4.9-ohm resistor, so textbook passive crossovers will generate accurate shapes, BUT time and phase alignment between the TPL-150H and M5N12-80 midrange horn is a challenge, both in circuit design and blended sound quality. Naturally you can put the TPL-150H back in line with the M5N12-80 voice coil to get time alignment - and allow the horn to physically interfere. OR you can put the TPL-150H in line with the mid-horn mouth and design a crossover which also provides the correct time-alignment difference. VERY EASY with active crossovers. FINE ART with passive crossovers.
-------
Either way, you will need to make M5N12 horn SPL measurements in order to design a decent crossover, and at ~2Khz you will require at least 2nd order slopes to maintain the M5N12 horn polar control and mash down cone breakup.
--------
Passive crossovers which modify T - M frequencies and phase leave auible artifacts. You will need to research "quasi-optimal" crossovers, like the LeCleach work, which use 2-different slopes to improve phase and time alignment with long horns at crossover frequency Fx.

Marco built this list of “quasi-optimal” crossovers in order of increasing offset:

3rd order Butterworth Low Pass, -3dB @ Fx*0.87 (+)
3rd order Butterworth High Pass, -3dB @ Fx*1.15 (-)
Offset = 0.22*c/Fx

4th order L-R Low Pass, -6dB @ Fx (+)
3rd order Bessel High Pass, -3dB @ Fx*1.4 (-)
Offset = 0.29*c/Fx

4th ordrer Butt Low Pass, -3dB @ Fx* 0,93 (+)
4th order L-R High Pass, -6dB @ Fx (-)
Offset = 0.31*c/Fx

6th order Bessel Low Pass, -6dB @ Fx*1.25 (+)
2nd order Butterworth High Pass, -3dB @ Fx*1.3 (-)
Offset = 0.40*c/Fx

6th order L-R Low Pass, -6dB @ Fx * 1.06 (+)
3rd order Butterworth High Pass, -3dB @ Fx* 1,13 (-)
Offset = 0.445 c/Fx

6th order L-R Low Pass, -6dB @ Fx * 1.07 (+)
4th order L-R High Pass, -6dB @ Fx * 0.92 (-)
Offset = 0.465 c/Fx

------------------------
 

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
I understand 1st order xo are in-phase, so would be a nice match with the active xo.
Ideal (textbook) first order filters put the responses in quadrature (90 degrees apart) at the listening position. Reality departs from the ideal.

For example, impedance compensation ordinarily becames a part of attaining a first order network influence. Also response tailoring becomes less simple with fewer components available. Fortunately the TPL is considered easy to work with (so I've heard).

It is normal with any crossover for a blanket choice of filter order to go out the window when the drivers aren't time aligned. Not that time alignment is essential.. the HP and LP filters end up being of different orders.

Having the responses in phase is not essential either, nor necessarily optimum for a given design.

And they seem easier to build.
I would think that second order offers simplicity based on the amount of control that two components can offer.
 

manninen

Member
2014-07-16 7:49 pm
i tried 1st and 2nd on 4way and it was way different than tight linear phase.
i could listen it like weather forecast but even little bit louder on music i felt lack of details and stream went to porridge.

when you´re active and it´s raining, it´s fun to try free different XO´s
 
Mmmm...I guess I'm facing a crossroad. Any Steve Vai fans? :)

Seems there will not be an easy path forward with a passive 2-way xo. Seems like wishful thinking on my part.

What if I give up the time-alignment constraint?
FWIW, the solution I seek here is not a final solution, but rather one that I'll use for about a year and then buy an 8-channel DAC and go active 4-way.


Taking a step back I can see two scenarios:

Scenario 1
  • Go for a time-aligned subwoofer-to-midabass-to-midrange and have the midrange to tweeter missaligned (due to passive xo there).
  • Settle for a suboptimal xo between midrange and tweeter and live with it for a while.
  • Pros: misalignment typically smaller at higher frequencies.
  • Cons: a) complexity of building an adequate passive xo; b)no complete time-alignment for a while.
Scenario 2
  • Use the passive xo inside the subwoofers and have time-aligned active xo between tweeter-to-midrange-to-midbass. So the lower channel from the DAC would output 20 to 400Hz and send it to the subwoofer plate amp, and from there output 80 to 400Hz to the midbass amp.
  • Pros: easy to implement, no fiddling with passive xo.
  • Cons: a) Time misalignment is usually larger at lower frequencies (e.g.: 80Hz vs 2500Hz in Scenario 1); b) still no complete time-alignment for a while.
Which path do you think I should follow?
 
Because the soundcard/DAC is a Lynx Hilo that can only do 6-channels. My initial goal was a Prism Titan (8 channels) but logistics didn't work out and for at least another year I will have the Hilo.

I would keep the time alignment on the tweeter and mid and remove it from the subwoofers.

Are you using multiple sub's?

What room treatment will you be implementing?

Remember that there are

1. Time / phase issues between drive units due to physical misalignment in space and crossover induced delay

2. Time / phase issues from the location of room boundaries and secondary reflections

No. 2 will have a FAR greater impact on the percieved accuracy of the recording and the illusion of " soundstage", imaging, correct timbre, "musicality" and all the other buzz words in the HiFi glossary :)

In other words, don't sweat the small stuff until you have fixed the important stuff
 
Last edited:

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
Some turn to active filters looking to make their task easier.. at the end of the day, none of the goals of the crossover change. Think of a crossover as a physical thing, and the filters simply trim the response to fit.

Time alignment is apparently overrated, and there are other priorities in positioning horns.
 
I would keep the time alignment on the tweeter and mid and remove it from the subwoofers.

Are you using multiple sub's?

Stereo subs for now. I've tried them in mono and mono per Geddes' approach trying multiple locations within the room, and settled on stereo placed at both front corners with digital room correction by Acourate. The difference in sound was small when a small sweet spot is considered, and a cleaner layout. Keep in mind this is in my living room, so aesthetics count.

What room treatment will you be implementing?
I've treated the room, overtreated, and now I'm looking to balance. But some treatments are directly related to the type of speakers I use, I think.
I had all 1st reflection points treated with absorption except the ceiling, big DIY tube bass traps, DIY Helmholtz resonators tuned to specific frequencies displaying peaks, etc. Some showed measured improvements larger than what could be heard, like the tube traps, and they looked big and nasty...out the door. Same for the Helmholtz resonators.

I now have a thick natural wool carpet at the floor 1st reflection point, reflective diffusion at the front wall 1st reflection, absorption at the sides 1st reflection, and a bookcase at the back 1st reflection (diffusion/absorption/diffraction/whatever). Ceiling still without treatment.

But these are for front-radiation speakers and might change if I end up with horns, I think. Diffusion at the front wall would seem useless, and probably the same for L/R 1st reflections. My plan is to use the whole front wall to build-in traetments, although the kind is still TBD. If horns, I'm thinking the front wall will likely just hide multiple bass traps.

Bottom line, I' ve studied room acoustics quite a bit, and thought it through before looking into speaker building. I am far from being done, but have come to believe the "system" includes the type of speaker used and a matching room.


Remember that there are

1. Time / phase issues between drive units due to physical misalignment in space and crossover induced delay

2. Time / phase issues from the location of room boundaries and secondary reflections

No. 2 will have a FAR greater impact on the percieved accuracy of the recording and the illusion of " soundstage", imaging, correct timbre, "musicality" and all the other buzz words in the HiFi glossary :)

In other words, don't sweat the small stuff until you have fixed the important stuff
I completely agree.
Assume for a minute No.2 is taken care off - which is not, but will be handled separately. Aren't the subwoofer to midbass time/phase issues within No.1 far larger than those from midrange to tweeter?
 
If you'll make two-way passive crossover, I have ready PCBs for you ))
Tested, well done.

Please, check this thread.

Thanks for the offer. Having a PCB is not that important for me, though. I think, if I build something, it should be rather simple where component-to-component wiring shouldn't be an issue.

The issue for me is crossover design :(
 
Some turn to active filters looking to make their task easier.. at the end of the day, none of the goals of the crossover change. Think of a crossover as a physical thing, and the filters simply trim the response to fit.

Time alignment is apparently overrated, and there are other priorities in positioning horns.

Fair enough. Maybe it is overrated. I really don't know as I haven't experienced it.

Actually some people who do think time-alignment is important also point out the misalignment is less so in tweeter to midrange, such as Uli Brueggemann (the man behind Acourate).

What type of passive xo would you recommend for me, assuming I have complete freedom in terms of physical alignment between the TPL-150H and the M5N12 with a horn that's about 40cm long and a rectangular 30x60cm mouth?

No need for baffle step compensation, easy to work with tweeter, 2-way, no need for time alignment...aren't I relieving constraints to the point a relatively simple XO can be designed?
 
When i had fst and tpl same axle on front i needed 4cm delay for tpl

That is good to know. Was that with a passive xo? If so, which type? And the FST as direct radiator or in the horn?

The TPL and midrange horn will be independent "boxes", so moving one relative to the other would be a non-issue. Other related issues to be handled the best I can (e.g., TPL reflecting on the horn back, etc)
 

manninen

Member
2014-07-16 7:49 pm
sorry i lied.

0cm sub
0cm low mid
0cm fst horn
32cm tpl

[IMGDEAD]https://dl.dropboxusercontent.com/u/33080121/hifi/pikkujoulu/td.jpg[/IMGDEAD]
[IMGDEAD]https://dl.dropboxusercontent.com/u/33080121/hifi/pikkujoulu/tdoff.jpg[/IMGDEAD]

tweeter, low mid and sub is quite near on same axis, fst is around 28cm further back.

thats linear phase xo, sorry about that

baffle mounted ob i had, el cheapo had:
0cm bass
11cm low mid
15cm up mid
20cm tweeter

thats zig-zag without delay
 
Last edited:

AllenB

Moderator
Paid Member
2008-10-18 11:31 am
An extra order on the tweeter with a conventional dome/mid-woofer box speaker in some cases compensates the Z alignment where the tweeter is closer. It's not always mentioned within these discussions that the two drivers are not co-located in the Y dimension and so cannot be technically 'perfect'.

You are taking two devices with quite different needs, shapes and sizes but at the crossover frequency you want them to be doing the same thing in the same location. I see this as being at the middle of the crossover issue.

Normally you will produce a good speaker if you consider the limits of all the compromises and keep them within proven reason. It can be off-putting to think that this thing or that thing is only just within reason, especially where a particular design forces all of the compromises to allow you very little leeway to obsess over a favourite issue. Vertical separation limits with regards to lobing tend to be relaxed with a system that has controlled narrow vertical directivity.

A horn/waveguide system offers benefits that are achieveable. Horn lengths, raw mouth sizes and subsequent termination (position, size, quantity, radiation space conditions), loading and waveguiding qualities of the devices will need to be juggled, as will group delay and other separations, along with the audibility of a given compromise vs frequency.

A suggestion is to calculate the minimum quantities for each and make sure you can fit these together into a system, then branch outward from there as the compromises will be more apparent. Once these conditions are satisfied, the electrical networks should almost build themselves as the result will be more or less known.