New speaker design by NOOB

[Adam Frandsen]

Hi,

I have done the following calculations and selections, and now I need help correcting any errors and would also appreciate any other inputs. If someone could also help make a pcb/schematic of the crossover, pads and BSC it would be amazing. I tried using easyEDA, but I cannot get the nets to connect properly and am really not sure I am doing it right. Thanks in advance:

Speaker build:



4-way speaker build:



**Bass Driver: Scanspeak 32w/4878t00**

- Frequency Range: 20 - 225Hz

- Power: 200W RMS, 550W long-term

- Sensitivity: 90dB

- Nominal Impedance: 4 Ohm

- Impedance at 225Hz: 4.5 Ohm

- Diameter: 320mm

- Depth: 151mm

- Resonance Frequency: 18Hz

- Vas: 203.9 l

- Qms: 7

- Qes: 0.33

- Qts: 0.32

- Price: 630 euro



**Mid Woofer: Scanspeak 18w/4531g00**

- Frequency Range: 225 - 1100Hz

- Power: 70W RMS, 110W long-term

- Sensitivity: 90dB

- Nominal Impedance: 4 Ohm

- Impedance at 225Hz: 4.5 Ohm

- Impedance at 1100Hz: 6 Ohm

- Diameter: 182.2mm

- Depth: 78mm

- Resonance Frequency: 33Hz

- Vas: 41.9 l

- Qms: 5.2

- Qes: 0.38

- Qts: 0.35

- Price: 200 euro



**Mid Tweeter: Scanspeak d3004/662000**

- Frequency Range: 1100 - 7500Hz

- Power: 90W RMS, 150W long-term

- Sensitivity: 91.5dB

- Nominal Impedance: 4 Ohm

- Impedance at 1100Hz: 4.5 Ohm

- Impedance at 7500Hz: 4.5 Ohm

- Diameter: 104.25mm

- Depth: 29mm

- Resonance Frequency: 500Hz

- Vas: 0.02 l

- Qms: 3.79

- Qes: 0.62

- Qts: 0.54

- Price: 240 euro



**Tweeter: Scanspeak d2104/712000**

- Frequency Range: 7500 - 40000Hz

- Power: 50W RMS, 120W long-term

- Sensitivity: 92dB

- Nominal Impedance: 4 Ohm

- Impedance at 7500Hz: 3 Ohm

- Impedance at 40000Hz: 5 Ohm

- Diameter: 92mm

- Depth: 33mm

- Resonance Frequency: 500Hz

- Vas: 0.01 l

- Qms: 2.3

- Qes: 0.48

- Qts: 0.40

- Price: 220 euro



These are the driver specifications with the correct crossover points. All drivers are of the Scanspeak brand and all have a nominal impedance of 4 Ohm, with specific values at crossover points as listed. Their respective power ratings, sensitivities, diameters, depths, resonance frequencies, Vas, Qms, Qes, and Qts values are also provided.



Dimensions of the speaker cabinet:



**Main Cabinet Size:**

- Total Height: 187 cm

- Width: 35 cm

- Depth: 38 cm

- Thickness of MDF: 24 mm



**Driver Placement Spacing:**

- Bass Driver: 31.25 cm from bottom of the cabinet (center of driver)

- Mid-Woofer: 87.25 cm from bottom of the cabinet (center of driver)

- Mid-Tweeter: 142.5 cm from bottom of the cabinet (center of driver)

- Tweeter: 167 cm from bottom of the cabinet (center of driver)



**Inner Compartments:**

- Bass and Mid-Woofer Compartment: starts from the bottom to 125 cm height. The volume of this compartment is approximately 137 liters.

- Mid-Tweeter and Tweeter Compartment: starts from 125 cm to the top (187 cm). The volume of this compartment is approximately 27 liters.



**Port Specifications:**

- The port is placed in the Bass and Mid-Woofer compartment.

- Diameter: 7.62 cm

- Length: 38.9 cm



The port should be placed on the rear panel of the Bass and Mid-Woofer compartment, ensuring that there's sufficient clearance from the edges of the cabinet and the internal components.



**Summary:**

The main cabinet's outer dimensions are 187 cm x 35 cm x 38 cm, built from 24 mm thick MDF. The total internal volume is approximately 164 liters, divided into two compartments: one for the bass and mid-woofer drivers with an approximate volume of 137 liters and another for the mid-tweeter and tweeter with an approximate volume of 27 liters. The port is in the bass and mid-woofer compartment with a length of 38.9 cm and a diameter of 7.62 cm.



**Crossover Network:**



*4-way 2nd Order Linkwitz-Riley design (with crossover frequencies at 225Hz, 1100Hz, and 7500Hz)*



For the Bass to Mid-Woofer Crossover (225Hz):

- Series Capacitor: 176.37uF (100V, 200W)

- Shunt Inductor: 7.06mH (200W)



For the Mid-Woofer to Mid-Tweeter Crossover (1100Hz):

- Series Inductor: 1.36mH (70W)

- Shunt Capacitor: 8.87uF (100V, 70W)



For the Mid-Tweeter to Tweeter Crossover (7500Hz):

- Series Capacitor: 1.06uF (100V, 90W)

- Shunt Inductor: 0.23mH (90W)



**L-Pad Attenuator:**



- For the Mid-Tweeter:

- Series resistor: 0.64 Ohm (90W)

- Parallel resistor: 19.5 Ohm (90W)



- For the Tweeter:

- Series resistor: 1.08 Ohm (50W)

- Parallel resistor: 13.7 Ohm (50W)



**Baffle Step Compensation (BSC) Circuit (6dB):** (Not sure if I will use)



- For the Bass Driver:

- Series Inductor: 1.39mH (200W)

- Parallel resistor: 6.73 Ohm (200W)

• For a 176.37uF capacitor, I might use a 100uF and 68uF capacitor in parallel, and add an 8.2uF and 470nF capacitors, to get approximately 176.67uF.
• For a 7.06mH inductor, I could use a 4.7mH inductor in series with a 2.2mH and 0.15mH inductors to get approximately 7.05mH.

Rough schematic:

AMP/IN --[C1=176.37μF]-- --[L1=7.06mH]-- || --[R1b=0.64Ω]-- --[R2b=19.5Ω]-- Bass Driver
|
|
--[L2=1.36mH]-- --[C2=8.87μF]-- || --[R1m=0.64Ω]-- --[R2m=19.5Ω]-- Mid-Woofer
|
|
--[C3=1.06μF]-- --[L3=0.23mH]-- || --[R1mt=1.08Ω]-- --[R2mt=13.7Ω]-- Mid-Tweeter
|
|
--[C4=1.06μF]-- --[L4=0.23mH]-- || --[R1t=1.08Ω]-- --[R2t=13.7Ω]-- Tweeter
|
|
Grnd-------------------------------------- Ground (Common Ground)

 

[Ricardus]

I don't know how many speaker designers there are on here, but there are a bunch of them on Audio Science Review if you aren't already a member there.

 

[Adam Frandsen]

I don't know how many speaker designers there are on here, but there are a bunch of them on Audio Science Review if you aren't already a member there.

Thanks, will try and post there as well!

 

[jensenmann]

Hi Adam

I´m not sure if I understood everything right. It seems you´re planning the woofer and low-mid driver to share the enclosure. That´s not a good idea. The woofer will modulate the low-mid driver. Make sure that both drivers have an enclosure of their own.

Did you simulate your enclosure and crossover?

 

[abbey road d enfer]

**Crossover Network:**



*4-way 2nd Order Linkwitz-Riley design (with crossover frequencies at 225Hz, 1100Hz, and 7500Hz)*



For the Bass to Mid-Woofer Crossover (225Hz):

- Series Capacitor: 176.37uF (100V, 200W)

- Shunt Inductor: 7.06mH (200W)



For the Mid-Woofer to Mid-Tweeter Crossover (1100Hz):

- Series Inductor: 1.36mH (70W)

- Shunt Capacitor: 8.87uF (100V, 70W)



For the Mid-Tweeter to Tweeter Crossover (7500Hz):

- Series Capacitor: 1.06uF (100V, 90W)

- Shunt Inductor: 0.23mH (90W)



**L-Pad Attenuator:**



- For the Mid-Tweeter:

- Series resistor: 0.64 Ohm (90W)

- Parallel resistor: 19.5 Ohm (90W)



- For the Tweeter:

- Series resistor: 1.08 Ohm (50W)

- Parallel resistor: 13.7 Ohm (50W)



**Baffle Step Compensation (BSC) Circuit (6dB):** (Not sure if I will use)



- For the Bass Driver:

- Series Inductor: 1.39mH (200W)

- Parallel resistor: 6.73 Ohm (200W)

• For a 176.37uF capacitor, I might use a 100uF and 68uF capacitor in parallel, and add an 8.2uF and 470nF capacitors, to get approximately 176.67uF.
• For a 7.06mH inductor, I could use a 4.7mH inductor in series with a 2.2mH and 0.15mH inductors to get approximately 7.05mH.

Looks to me like you have all info needed. What do you miss?

Rough schematic:

AMP/IN --[C1=176.37μF]-- --[L1=7.06mH]-- || --[R1b=0.64Ω]-- --[R2b=19.5Ω]-- Bass Driver
|
|
--[L2=1.36mH]-- --[C2=8.87μF]-- || --[R1m=0.64Ω]-- --[R2m=19.5Ω]-- Mid-Woofer
|
|
--[C3=1.06μF]-- --[L3=0.23mH]-- || --[R1mt=1.08Ω]-- --[R2mt=13.7Ω]-- Mid-Tweeter
|
|
--[C4=1.06μF]-- --[L4=0.23mH]-- || --[R1t=1.08Ω]-- --[R2t=13.7Ω]-- Tweeter
|
|
Grnd-------------------------------------- Ground (Common Ground)

Not a very nice way of making schematics. You know you can attach pictures here...?

 

[Adam Frandsen]

Hi Adam

I´m not sure if I understood everything right. It seems you´re planning the woofer and low-mid driver to share the enclosure. That´s not a good idea. The woofer will modulate the low-mid driver. Make sure that both drivers have an enclosure of their own.

Did you simulate your enclosure and crossover?

I did not simulate the crossover and enclosure, as I do not have the program to do so. Making separate compartments for the mid woofer and bass should not be a problem, thanks for pointing that out. I will have tp recalculate the port and decide if there should be two…

 

[Adam Frandsen]

Looks to me like you have all info needed. What do you miss?

Not a very nice way of making schematics. You know you can attach pictures here...?

I was just hoping someone could assure me the calculations were correct… As mentioned I tried easyEDA, but must be doing something wrong, because the nets are not connecting properly, so I gave up.

 

[abbey road d enfer]

I was just hoping someone could assure me the calculations were correct…

I don't know easyEDA, I understand it's a PCB software.
How did you calculate the x-overs?
Calculating x-overs is a quite difficult task; the best packages in this respect use the speaker's true frequency and impedance response, as well as their efficiency. The cookbook calculators that assume a single resistive impedance result in poor frequency response.
Even with a good software package, the practical results often need to be tweaked.
I understand you're willing to spend serious money (€1000+ on speakers) on this project. I know for certain that you will not achieve a result equivalent to a €1k commercial speaker.
If you want to build your speakers, I would recommend buying an established kit, one that has been fully developed.
Unfortunately there are not many 3-way kits (and justifiably).

As mentioned I tried easyEDA, but must be doing something wrong, because the nets are not connecting properly, so I gave up.

Can you post pictures of the EasyEDA schemo and layout ?

 

[Adam Frandsen]

I don't know easyEDA, I understand it's a PCB software.
How did you calculate the x-overs?
Calculating x-overs is a quite difficult task; the best packages in this respect use the speaker's true frequency and impedance response, as well as their efficiency. The cookbook calculators that assume a single resistive impedance result in poor frequency response.
Even with a good software package, the practical results often need to be tweaked.
I understand you're willing to spend serious money (€1000+ on speakers) on this project. I know for certain that you will not achieve a result equivalent to a €1k commercial speaker.
If you want to build your speakers, I would recommend buying an established kit, one that has been fully developed.
Unfortunately there are not many 3-way kits (and justifiably).


Can you post pictures of the EasyEDA schemo and layout ?

I calculated the crossover by typing in all the measurements of the individual drivers, nominal impedance as well as impedance at the crossover points, also developed the l-pads based on the speaker sensitivity and frequency response, same with the BSC. Physical driver placement on the baffle is calculated for a room at approximately 5x5m with the off angle frequency responses included as well as the crossover frequencies - I did calculate a horizontal offset for them too yesterday, but it doesn't quite fit with the baffle size, so have to do some more twerking with that.

 

[Adam Frandsen]

In terms of the schematic and PCB, I will make another effort either today or later this week

 

[Adam Frandsen]

In terms of a kit it doesn't make sense to me, I have Quad speakers already, so I am not interested in someone else's design, the fun part of this would be having done it all myself. That would lead to a perception of better sound for sure - the mind is a beautiful thing that way. Aesthetics. I hope I can achieve a good result, we shall see. Plan is to sell my current speakers and replace them with these.

 

[Disco Volante]

Sounds like you've got it pretty much figured out. I can only second Abbey's advice: some real-life tweaking is usually involved getting the crossovers to work well once you're done building boxes.
A 4-way system is ambitious I dare say, that's a lot of phase-relations and radiation patterns to get just right. LR2 filters might require at least the upper drivers to be "time-aligned" to work properly.

Did you use EDA for all the calculations? I shall have to look into that;-)

I might be able to give you a hand should you want fancier woodwork, veneer, etc.

Happy tinkering
Viggo

 

[abbey road d enfer]

I calculated the crossover by typing in all the measurements of the individual drivers, nominal impedance as well as impedance at the crossover points,

Did you measure them or are you using the manufacturer's data?

also developed the l-pads based on the speaker sensitivity and frequency response

From your list, I see the low-mid has only a low-pass filter, and the high-mid only a high-pass. Is it correct?
Looks to me like there would be a lot of interaction when speakers overlap.
Have you taken into account that some speakers must be reverse-connected in a 2nd-order x-over?
I strongly suggest you read this VituixCAD help 2.0
You don't have to use this software, but there are all the definitions and methods you need there.

 

[Adam Frandsen]

There is no 90 degree out of phase with the crossover type used.

 

[abbey road d enfer]

There is no 90 degree out of phase with the crossover type used.

Really? Then it's not a Linkwitz-Riley X-over.
A 2nd-order LR x-over requires inversion.
Anyway, LR x-overs are good only for 4th-order, whatever some may say. It's actually the only implementation that obeys the rules that Siegfried Linkwits himself has fixed (pressure summing, energy summing, transient response).
There are not many examples of 2nd-order LR x-overs, or calcs thereof; it stands to reason.

I don't understand why you don't have real passband filters for the low-mid and high-mid. This creates interference problems in the x-over area.

 

[Adam Frandsen]

Really? Then it's not a Linkwitz-Riley X-over.
A 2nd-order LR x-over requires inversion.
Anyway, LR x-overs are good only for 4th-order, whatever some may say. It's actually the only implementation that obeys the rules that Siegfried Linkwits himself has fixed (pressure summing, energy summing, transient response).
There are not many examples of 2nd-order LR x-overs, or calcs thereof; it stands to reason.

I don't understand why you don't have real passband filters for the low-mid and high-mid. This creates interference problems in the x-over area.

Seems you are right and I have been misguided - this should be more correct, but will have to do more research now, because my source was wrong:

Bass to Mid-Woofer Crossover (200Hz):
- Impedance at 200Hz: 4.5 Ohm
- First Stage:
- Series Capacitor: 176 uF (100V)
- Shunt Inductor: 8.8 mH (250W)
- Second Stage:
- Series Capacitor: 176 uF (100V)
- Shunt Inductor: 8.8 mH (250W)

Mid-Woofer to Mid-Tweeter Crossover (1200Hz):
- Impedance at 1200Hz: 6 Ohm
- First Stage:
- Series Inductor: 0.22 mH (250W)
- Shunt Capacitor: 22 uF (100V)
- Second Stage:
- Series Inductor: 0.22 mH (250W)
- Shunt Capacitor: 22 uF (100V)

Mid-Tweeter to Tweeter Crossover (8000Hz):
- Impedance at 8000Hz: 3 Ohm
- First Stage:
- Series Capacitor: 6.6 uF (100V)
- Shunt Inductor: 0.079 mH (250W)
- Second Stage:
- Series Capacitor: 6.6 uF (100V)
- Shunt Inductor: 0.079 mH (250W)

​

 

[ccaudle]

The cookbook calculators that assume a single resistive impedance result in poor frequency response.

The combined impedance of all the interacting drivers can sometimes be challenging for amplifiers as well. Active crossovers solve those problems, and allow for simpler amplifier designs since the individual amp channels do not have to deal with low impedance and challenging phase angle where the driver responses overlap.
Even if you eventually want a passive crossover using a digital active crossover for prototyping can allow you to quickly measure different options to guide you in the passive crossover implementation.

 

[abbey road d enfer]

Seems you are right and I have been misguided - this should be more correct, but will have to do more research now, because my source was wrong:

Bass to Mid-Woofer Crossover (200Hz):
- Impedance at 200Hz: 4.5 Ohm
- First Stage:
- Series Capacitor: 176 uF (100V)
- Shunt Inductor: 8.8 mH (250W)
- Second Stage:
- Series Capacitor: 176 uF (100V)
- Shunt Inductor: 8.8 mH (250W)

Mid-Woofer to Mid-Tweeter Crossover (1200Hz):
- Impedance at 1200Hz: 6 Ohm
- First Stage:
- Series Inductor: 0.22 mH (250W)
- Shunt Capacitor: 22 uF (100V)
- Second Stage:
- Series Inductor: 0.22 mH (250W)
- Shunt Capacitor: 22 uF (100V)

Mid-Tweeter to Tweeter Crossover (8000Hz):
- Impedance at 8000Hz: 3 Ohm
- First Stage:
- Series Capacitor: 6.6 uF (100V)
- Shunt Inductor: 0.079 mH (250W)
- Second Stage:
- Series Capacitor: 6.6 uF (100V)
- Shunt Inductor: 0.079 mH (250W)

​

I see no low-pass for bass, no low-pass for low-mid, no high-pass for high-mid.
What you describe are 4th-order filters with a weird response. Is that what you want?
In order to make a bandpass, it would be something like

Bass to Mid-Woofer Crossover (200Hz):
- Impedance at 200Hz: 4.5 Ohm
- First Stage:
- Series Capacitor: 176 uF (100V)
- Shunt Inductor: 8.8 mH (250W)
- Second Stage:
- Series Inductor: 0.22mH (?)
- Shunt Capacitor: 22 uF (?)
Although I seriously doubt these values, seems to me like a factor 10 error in calculations

 

[JohnRoberts]

Really? Then it's not a Linkwitz-Riley X-over.
A 2nd-order LR x-over requires inversion.
Anyway, LR x-overs are good only for 4th-order, whatever some may say. It's actually the only implementation that obeys the rules that Siegfried Linkwits himself has fixed (pressure summing, energy summing, transient response).
There are not many examples of 2nd-order LR x-overs, or calcs thereof; it stands to reason.

I don't understand why you don't have real passband filters for the low-mid and high-mid. This creates interference problems in the x-over area.

I am not enthusiastic about going down this rabbit hole. Designing loudspeaker crossovers is not trivial, further in the real world how drivers work together or not is a 3D problem trying to be solved with 2D solutions.
==
On the subject of LR crossovers the iconic 4th order LR is two 2nd order Butterworth filters cascaded in series.

Books could be written on this subject and several have been...

Good luck...

JR

 

[Adam Frandsen]

Won't be easy I am sure, gonna have to do some more research