Advantages of Series Crossover vs Parallel for Subs & Open Baffle

[tubelectron]

Can you confirm this with real measurements of your designs?

Besides, the crossover of these small monitors has no baffle step compensation, it is too simple to talk about correct filtering of the speakers.
Those schematics you showed are book examples of theoretical filters that were not built based on real and correct acoustic measurements.

No, these are real examples, built and tested :

Thank you, I know how to make correct acoustic measurements :

I am also able to recognize by ear when a speaker is in phase or not, and if the transients match themselves between the speakers or not. No need to simulate, nor ot measure it...

I do not have measurements of my designs because for now, my measurement chain has become obsolete (thank you Windows).

T

 

[mV8]

Thank you, I know how to make correct acoustic measurements :

If you took measurements in the way you showed in the pictures, then unfortunately it is not about the correct way.

Don't get me wrong. I also like to use series crossovers, but what you have shown will not work correctly.

I encourage you to re-do the measurements one day according to generally known guidelines.

 

[wolf_teeth]

It really depends on the order of rolloff, not the electrical slope, and I've had them line up both ways on different designs. It is not just because they are SXO filters.

 

[tubelectron]

It really depends on the order of rolloff, not the electrical slope, and I've had them line up both ways on different designs. It is not just because they are SXO filters.

Yes. In essence, for the S- // Xover, the order of the rolloff starts at circa 3dB at the crossing frequency, increases to 6dB, then 9dB, and approaches slowly 12dB at the end. Here are two simulations made by an Audio friend and presented on his webpage (250Hz / 4kHz) :

* A 3-ways classic parallel 6dB/Oct. crossover shows the regular slopes at 6dB per octave :

* A 3-ways serial-parallel 1st order crossover shows the increasing slopes :

* Note that the speakers on the diagram above are all in same polarity, hence the synchronized transient response between speakers, shown below on square waves, by the serial-parallel configuration :

That said, the speakers must be linear enough around the crossing frequencies (2 octaves would be the ideal) : no great accident in the response nor in the impedance curves, like for a classic // 6dB/Oct. by the way, but fortunately in a lesser extent for that S-// configuration.

T

 

[mV8]

This is only a simulation on an ideal speaker with linear impedance, and such a situation does not occur in real life.

 

[Davey]

Unfortunately, in the real world, that doesn't work like this in a 1st order serie crossover, being in 2 or 3 ways, or over : any polarity inversion is heard as a scooped mid tone, and the transient response is erased. You guess that I made the test several times on several enclosure projects, and each time : same conclusion...

The very first post in this thread shows a real world example where a polarity inversion is required.

Dave.

 

[Mister Audio]

Very interesting discussion. If one designed a two way 1st order series crossover for woofer and wide band (FR) could one add a helper tweeter/super tweeter by simply paralleling a cap blocked tweeter or would a redesign as a three way be necessary?

Yes you certainly can.
I have just obtained a pair of Pioneer B350PRO bullet super-tweeters for a very similar reason.
Their very high efficiency and 'natural bandwidth' means you can bring them into SPL range
using a single low value capacitor, achieving attenuation & very high frequency output simply.
However, you must get your (+/- ) phase polarity correct for proper results.

 

[Mister Audio]

Hello all,

Some posts ago I read > "These are standard 'book shelf' crossover networks, without compensation for "baffle step".

I would like 1, 2 or more members to present 2 examples of: *1. basic XO and then *2. XO with baffle step compensation.

I would be very interested to see some examples of this

 

[Boden]

I do not have measurements of my designs because for now, my measurement chain has become obsolete (thank you Windows).

Your .frd or .txt and .zma measurement files are simple text files in columns and can survive fully independently from MS-DOS 3.1 or whatever until Win11.

 

[tubelectron]

Your .frd or .txt and .zma measurement files are simple text files in columns and can survive fully independently from MS-DOS 3.1 or whatever until Win11.

Yes. But I don't have them anymore - I even wonder if I had them. I just have the proprietary application files.

Now I have a ARTA license, and some other free software, but the perspective of reinstating a complete measurement chain makes me lazy.

T

 

[AllenB]

I would like 1, 2 or more members to present 2 examples of: *1. basic XO and then *2. XO with baffle step compensation.

I would be very interested to see some examples of this

https://www.diyaudio.com/community/...overs-without-measurement.189847/post-6910468

 

[mV8]

Hello all,

Some posts ago I read > "These are standard 'book shelf' crossover networks, without compensation for "baffle step".

I would like 1, 2 or more members to present 2 examples of: *1. basic XO and then *2. XO with baffle step compensation.

I would be very interested to see some examples of this

I will show such an example of filtration based on my design of a cheap stand monitor, in which I used a series crossover. In the next few posts I will describe its behavior. Measurements for the crossover were taken using Arta software, on a 0 to 90 degree rotary table.

The first screenshot is the measurements of the speakers connected in parallel. As you can see, this is a connection in the same polarity, and even without any crossover we see quite a bit of phase distortion.

 

[mV8]

Now let's see how the simplest 1st order serial crossover works. I added an Lpad for the tweeter to the schematic to lower its SPL slightly relative to the midwoofer.

Looking at the transmittance of the filter, you can see that such a crossover is unable to correct the Baffle step phenomenon that occurs with a typical loudspeaker enclosure. If we used an infinite baffle for these speakers, such a simple filter could work, but in the real world we still have a problem with the midwoofer, whose frequency response is not ideal and requires further intervention.

The two speakers are still connected in the same polarity, which manifests itself in a huge frequency response hole caused by a phase mismatch.

Reversing the polarity of the tweeter significantly improves the frequency response.

 

[mV8]

We need to add another coil to our schematic to correct the Baffle step effect. This is still far from the theoretical ideal, especially considering the frequency response of the midwoofer. We need to attenuate the break up of the cone, which can be achieved by using additional elements at our coil responsible for Baffle step correction. This is a typical LCR scheme, acting as a notch filter. I have added a series resistor to the schematic, which will make it easier to achieve optimal filtering characteristics and impedance without large drops in the higher bandwidth.

In the end, we have a series crossover with a small number of elements, but allowing for effective speaker filtering. This crossover scheme is quite versatile and even when using speakers with not very even frequency response, one can get quite good results.

Thanks to the use of a waveguide tweeter and an enclosure with chamfered sides, we get fairly even bandwidths off the listening axis. The waveguide also allows for better time alignment of the speakers.

 

[mV8]

A serial higher-order crossover can also work very well. The schematic creates a 3rd-order crossover, but acoustically you get an LR4 crossover. In this case, the polarity of the speakers will be the same.

 

[rolf101]

Maybe a little OT, but ...
I´ve been using the S- filter in the attachment in several 6,5 inch + tweeter constructions and been satisfied with the results.
But I don´t fully understand how the 6,8 uF capacitor and 2,2 ohm resistor attached between the two capacitors and the + terminal of the B/M works.
Has anyone an explanation?

 

[Davey]

@rolf101
Yeah, that's been discussed here on the forum previously.
It's sort of an "in-between" configuration for series-crossover advocates that can't fully commit.
The 2.2 ohm resistor would be a wire in a series crossover. But if you were to raise the value high enough you'd have a parallel crossover.

Dave.

 

[rolf101]

Many thanks Dave,
Of course I should have seen the shorted 2,2 ohm resistor as a vital part of the series X0.
But I still don´t understand the 6,8 capacitor and how to use it use it together with the resistor when I alter the XO.
Anyhow, I find it easy to use and I find it satisfactory.