New speaker design by NOOB

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
IMD being dependant on the combination of several frequencies, the actual result depends very much on how the loudspeaker is used, particularly the range of frequencies.
Just a simple example: typical IMD measurement implies mixing a low frequency (50 or 60Hz) with a HF tone (typically 7kHz).
Applied to a woofer, the 7kHz signal is very low, so the modulated signal is bound to be difficult to measure, when applied to a tweeter, it will be (ab)used so much outide its nominal range that the results will be horrific. In these conditions, no manufacturer wants to publish these results.
DFD wouldbe a better assessment of performance though, but will also be of limited use, since measurements have to be adapted to the particular nominal range of the speaker, so no specs would be comparable.
Loudspeakers are actually kind of "3D" apparatus, so can't easily be described by single parameters.

IMD varies almost exponentially with level, so the designer has some leeway as to how he wants to use the loudspeaker. Using two speakers instead of one can result in seriously improved IMD.
There should be a standard made for each type of driver (woofer, sub, mid, tweeter etc.) that would help determine its real life precision in reproducing simultaneous frequencies within it's intended range. I don't see the use of a great frequency response, if it completely distorts once more frequencies are added simultaneously, which in praxis would be all the time when playing music. DFD is a type of IMD, no?
 
Last edited:
I tried to search "praxis" and apparently I can't help you...

Good luck.

JR
Thanks for your comment, JR. You're correct that 'praxis' might not be a term commonly used in everyday English. However, it's used in academic language, particularly in the fields of philosophy and education, and its root comes from Greek and Latin, where it refers to 'practice' or 'action'. In the context of our discussion here, I used 'praxis' to signify the practical application of theory, in particular referring to the real-world use of speakers. While English is not my first language, I'm glad we're having this exchange as it opens up an opportunity for us to learn from linguistic and cultural nuances that may not always be apparent. I look forward to more enriching discussions!
 
Wow - thanks 🙏 I am doing some more research at this stage and recalculating everything, I am even considering using cheaper drivers for the build, but nothing is final yet. I just realized I need to put more time into this, so I am sure to get a good result. Once I am further in the process, I may well draw upon your resources, thank you so much for being so generous!!
[thank you so much for being so generous!!] -- NO PROBLEM!!! I am "Standing By....." for a later time!!!

Besides.....what good does having all of this expensive CAD-design software do ya if you don't use it, huh???

/
 
[thank you so much for being so generous!!] -- NO PROBLEM!!! I am "Standing By....." for a later time!!!

Besides.....what good does having all of this expensive CAD-design software do ya if you don't use it, huh???

/
The program in question is actually and amazingly enough free, so I will try and see if I can gain some proficiency in it. But, like you say, I am not gonna be building speakers commercially or even repeatedly, so I don't necessarily need to become fluent in speaker design or other software for this one build. I will make an effort however :)
 
The program in question is actually and amazingly enough free, so I will try and see if I can gain some proficiency in it. But, like you say, I am not gonna be building speakers commercially or even repeatedly, so I don't necessarily need to become fluent in speaker design or other software for this one build. I will make an effort however :)
[The program in question is actually and amazingly enough free] -- It doesn't matter whether a software program is -- FREE -- or it costs "BIG BUCKS" like my software, they all take a fair amount of a "learning curve" in order to learn how to use the software, let alone become proficient with using it. In addition, if you are not familiar with knowing how to use CAD-software in general, then the process of learning -- HOW -- to effectively "use" the software can then become daunting in and of itself. In other words, using CAD-software is a whole different "world" than learning how to use a word-processor!!!

So.....if you want to go ahead and learn how to use your EasyEDA software for an occasional use every now and then, then.....forge ahead and feel good about learning something new. But, on the other hand.....if you find learning how to use the EasyEDA program is too challenging to bother with.....well.....here I am!!!

P.S. -- If I were going to use a -- FREE -- CAD-design program like how you are planning on doing, I would opt for a program that probably has more users on this forum than what EasyEDA does and that program is the popular -- KiCAD -- schematic and PCB program!!! Plus.....while the software itself is completely -- FREE -- to download and use, they also offer an extensive "Training Course" for around $55.00, I think. The training course comes with a manual and an extensive set of video tutorials that shows you exactly how to use the software in "easy to digest" bite-sized videos that range from just a minute or two, all of the way up to around 15-to-20 minutes. When I purchased my video course, the video files were downloadable and comprised of a total of 8.5GB of videos!!! This would be my suggestion to you as the training course will probably eliminate the -- frustration level -- you will experience with attempting to learn EasyEDA on your own with no assistance. But.....it's your call!!! Just my 2-cents worth!!! Here ya go!!!.....

https://www.kicad.org
/
 
[thank you so much for being so generous!!] -- NO PROBLEM!!! I am "Standing By....." for a later time!!!

Besides.....what good does having all of this expensive CAD-design software do ya if you don't use it, huh???

/
I don't know how you're doing it, but the text on your posts is quite a bit larger than anyone else's - comes across as a bit 'shouty'.
 
There are too many things to comment on your design. Anyway, if this is your first don't use those drivers, they are too expensive for the job regarding your (assumed) skills. Also start with something simple like a 2-way first. Learn where and why should you use the drivers, like I have never seen two almost similar tweeters used with xover @7500 Hz (usually another tweeter is super tweeter for frequencies the main one cannot reproduce). Think in octaves, not linear scale. When deciding the crossover (=xover) frequencies you should take into account the directivity of the drivers so that directivity of the loudspeaker is as constant as possible which means that in room its sound is balanced overall.

Best way to design is to prototype, measure and then simulate the crossover. I recommend VituixCAD (it's free for non-commerial use) for the simulation, and Room EQ Wizard for measurements (Arta should work also). After you have learned how and where on the front baffle you should put the various drivers and what's the optimal distance between them, and considered the need for waveguide (for tweeter, or the mid in some cases) build the prototype (or final if you are commited enough) box and start measuring the reponses of the individual driver at various angles. Combine the measurements and import them to CAD program and start the crossover simulation. You can use the values you calculated earlier as starting points (please ditch the other tweeter, and consider WG for the other as it will give you lower crossover frequency which will be better match for the mid). Good luck, it will be fun!
 
There should be a standard made for each type of driver (woofer, sub, mid, tweeter etc.)
A woofer optimized for a two-way system is quite different tha one designed for a 3 or 4-way system, and just the same, a woofer for a closed box is very different than one designed for bass-reflex or a 6-th order box, so you would end up with many different types of "standards". Speaker manufacturers and designers don't want to go into this rabbit hole.
The standard industrial procedure starts with selecting potential candidates based on a few parameters, which, in addition to those already mentioned by JR, include Qts and Bl.
Distortion (harmonic or intermod) is evaluated by the designer in the target box.
Even the basic frequency response cannot be trusted, because it's done on a standardized open baffle, which gives a very different response than most commercial boxes.
 
I have done the following calculations and selections, and now I need help correcting any errors and would also appreciate any other inputs.

First, if you really want an idea what your speaker drivers and crossovers will do you need to use a proper simulation software. I really like Vituixcad:

VituixCAD Loudspeaker simulator

Second, in my experience "textbook calculated" crossovers are ALWAYS A TOTAL DISASTER.

If someone could also help make a pcb/schematic of the crossover, pads and BSC it would be amazing.

You do not need a PCB for a crossover. Given the physical size of components it is best to hardwire them on a suitable piece of plywood (not MDF). Keep it external in order to be adjust things like L-Pads etc easily.

Speaker build:
4-way speaker build:

Very ambitious. Every single way adds complexity. Doing that with passive crossovers is challenging.

**Bass Driver: Scanspeak 32w/4878t00**
- Price: 630 euro

Outch. 12" is the size for lower midrange, not a bass driver and 630 Euro can buy a lot of drivers that are better suited to operate as "bass". And you may be able to buy several drivers per side for the price of one SS driver.

I would go as far as suggesting that making at least the LF section active (like was on some Harman Sub Brands, like Revel & Infinity) and adding some parametric EQ to deal with room modes will be a MUCH better choice.

Using non-traditional enclosure systems (dipole or cardioid woofer systems) should be seriously considered.

**Mid Woofer: Scanspeak 18w/4531g00**
- Price: 200 euro

Not a bad choice, I liked the Wavecore WDF182BD0X series better. They have a more advanced linear magnet structure. Given the low sensitivity and small diameter multiples will be needed, maybe 4pcs in an array? This will also give improved directivity control over single drivers.

Alternatively, a suitable 10 - 12" driver may be better suited for the lower midrange role.
**Mid Tweeter: Scanspeak d3004/662000**
- Price: 240 euro

Overpriced, underperforming. Poor directivity control.

It is better to use a spherical wave waveguide (some might say "horn") with a dimeter similar to the woofer, if using a large driver (10-12") coaxial arrangements may be considered.

Otherwise, I like the ring radiators with a large waveguide. Note that these will need considerable equalisation, because the constant directivity causes a non-flat response that will be maximum SPL at the lower cutoff of the waveguide.

**Tweeter: Scanspeak d2104/712000**
- Price: 220 euro

Again, over priced, under performing and probably unnecessary. Using a decent waveguide loaded tweeter will likely extend high enough.

These are the driver specifications with the correct crossover points.

What makes your chosen crossover points "correct"?

**Main Cabinet Size:**
- Total Height: 187 cm
- Width: 35 cm
- Depth: 38 cm

This is a VERY LARGE single enclosure. I would consider splitting this into multiple boxes to stack.

- Thickness of MDF: 24 mm

Very, very seriously consider to use a different cabinet material, or if you must use MDF, consider an "applique" layer of artificial or real stone.

So-called "black galaxy" granite from India is affordable in 12mm and gives a polished high gloss surface and makes for a very inert cabinet. These are used for kitchens and bathrooms, so making cutouts with a champher or roundover can be left to the Stone shop.

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

The higher the LF driver is placed, the greater the influence of the so-called floor notch.

This should be analysed and normally placing the woofer as low as possible and placing the midrange at about ear height and setting the crossover to keep the floor notch out of the lower midrange drivers operation range is a good choice.

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

This seems relatively high for a seated position.

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

This leaves a large distance between lower midrange and upper midrange.

This will lead to comb filtering and places the tweeter axis nearly at standing height (defo for my GF).

Normally I would aim for a seated height of 1m and place the HF Axis at ~ 1m height, unless there are very good reasons to place it elsewhere.

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

Yeah, don't do that, see above.

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

Lower midrange needs a separate enclosure.

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

That's a lot of wasted space.

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

This is near impossible. Each driver is a 2nd order bandpass device. So unless you create huge overlap between drivers, you cannot ever achieve a 2nd order acoustic crossover. And the acoustic crossover is what matters.

I would suggest to spend some time looking into speaker design and reading a few good books and websites on the topic, before continuing.

IF you were to consider my advise, I would do the following to come somewhat close to your original idea of a massively tall speaker that would have impressed the people who erected stonehenge (which I wholeheartedly approve of).

Go fully active with a DSP crossover and get a measurement microphone (from MiniDSP). It is much much more likely you will end up with a well working system.

Use high quality linear (non-switching) Amplifiers for MF & HF - my recommendation would be a few of the pre-build LM3886 based Amplifiers you can find on Aliexpress. Woofers, Class D will be fine, you will need a lot of power.

Make a Mid-Hi MTM array with a waveguide equipped tweeter similar to for example Amphion:

1691145969703.png

Use two woofers and place one at floor level and one at the top of your enclosure.

Split your actual ~200cm tall enclosure into three sections, the middle one with tweeter height at 100cm and 60cm tall and the two woofer sections 70cm high each.

You can make this a narrow "tallboy" by mounting woofer in the sides, but instead I suggest that a wider and not so deep box is preferable if you can accommodate it in terms of decor.

An old Dynaudio design (Myrage) implemented a similar design with 2 X 12" Woofer, 2 X 3" Dome and a 1" Tweeter. I heard some recently, they are still an excellent speaker.

1691146793752.png

I would probably suggest much more radical designs, if you were to ask me.

But something like a Myrage or similar, using a fully active system and suitable DSP crossover is 99% assured to give great results and is actually manageable as build.

Finally, a website to consider reading about speakers (not mine):

DIY-Loudspeakers

There are also many excellent projects completely worked and reasonably easy to replicate. Here a scan speak 4-way similar to your ideas, FWIW:

Ellipticor-A50-mkII

Thor
 
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.

OMG, everything sounds the same and audio-orthodxy central. I'd consider it the last place where you might get useful information. That's pure cargo cult over there.

Thor
 
The calculations may be correct, but your actual "design" clearly is not. So you most likely have a very correct answer to a totally wrong question.

Thor
Yes, right - I already decided more research is needed, and while I still like the idea of the tall cabinet and two tweeters, I am for sure gonna look at cheaper drivers, the scan speak are overpriced even if they are fantastic! I would be more proud if I could build a great speaker from lesser or rather cheaper parts. I have already begun making a list of possible drivers from SBA, Dayton... Haven't checked SEAS catalogue yet, but they may well be expensive too.
 
First, if you really want an idea what your speaker drivers and crossovers will do you need to use a proper simulation software. I really like Vituixcad:

VituixCAD Loudspeaker simulator

Second, in my experience "textbook calculated" crossovers are ALWAYS A TOTAL DISASTER.



You do not need a PCB for a crossover. Given the physical size of components it is best to hardwire them on a suitable piece of plywood (not MDF). Keep it external in order to be adjust things like L-Pads etc easily.



Very ambitious. Every single way adds complexity. Doing that with passive crossovers is challenging.



Outch. 12" is the size for lower midrange, not a bass driver and 630 Euro can buy a lot of drivers that are better suited to operate as "bass". And you may be able to buy several drivers per side for the price of one SS driver.

I would go as far as suggesting that making at least the LF section active (like was on some Harman Sub Brands, like Revel & Infinity) and adding some parametric EQ to deal with room modes will be a MUCH better choice.

Using non-traditional enclosure systems (dipole or cardioid woofer systems) should be seriously considered.



Not a bad choice, I liked the Wavecore WDF182BD0X series better. They have a more advanced linear magnet structure. Given the low sensitivity and small diameter multiples will be needed, maybe 4pcs in an array? This will also give improved directivity control over single drivers.

Alternatively, a suitable 10 - 12" driver may be better suited for the lower midrange role.


Overpriced, underperforming. Poor directivity control.

It is better to use a spherical wave waveguide (some might say "horn") with a dimeter similar to the woofer, if using a large driver (10-12") coaxial arrangements may be considered.

Otherwise, I like the ring radiators with a large waveguide. Note that these will need considerable equalisation, because the constant directivity causes a non-flat response that will be maximum SPL at the lower cutoff of the waveguide.



Again, over priced, under performing and probably unnecessary. Using a decent waveguide loaded tweeter will likely extend high enough.



What makes your chosen crossover points "correct"?



This is a VERY LARGE single enclosure. I would consider splitting this into multiple boxes to stack.



Very, very seriously consider to use a different cabinet material, or if you must use MDF, consider an "applique" layer of artificial or real stone.

So-called "black galaxy" granite from India is affordable in 12mm and gives a polished high gloss surface and makes for a very inert cabinet. These are used for kitchens and bathrooms, so making cutouts with a champher or roundover can be left to the Stone shop.



The higher the LF driver is placed, the greater the influence of the so-called floor notch.

This should be analysed and normally placing the woofer as low as possible and placing the midrange at about ear height and setting the crossover to keep the floor notch out of the lower midrange drivers operation range is a good choice.



This seems relatively high for a seated position.



This leaves a large distance between lower midrange and upper midrange.

This will lead to comb filtering and places the tweeter axis nearly at standing height (defo for my GF).

Normally I would aim for a seated height of 1m and place the HF Axis at ~ 1m height, unless there are very good reasons to place it elsewhere.



Yeah, don't do that, see above.



Lower midrange needs a separate enclosure.



That's a lot of wasted space.



This is near impossible. Each driver is a 2nd order bandpass device. So unless you create huge overlap between drivers, you cannot ever achieve a 2nd order acoustic crossover. And the acoustic crossover is what matters.

I would suggest to spend some time looking into speaker design and reading a few good books and websites on the topic, before continuing.

IF you were to consider my advise, I would do the following to come somewhat close to your original idea of a massively tall speaker that would have impressed the people who erected stonehenge (which I wholeheartedly approve of).

Go fully active with a DSP crossover and get a measurement microphone (from MiniDSP). It is much much more likely you will end up with a well working system.

Use high quality linear (non-switching) Amplifiers for MF & HF - my recommendation would be a few of the pre-build LM3886 based Amplifiers you can find on Aliexpress. Woofers, Class D will be fine, you will need a lot of power.

Make a Mid-Hi MTM array with a waveguide equipped tweeter similar to for example Amphion:

View attachment 112673

Use two woofers and place one at floor level and one at the top of your enclosure.

Split your actual ~200cm tall enclosure into three sections, the middle one with tweeter height at 100cm and 60cm tall and the two woofer sections 70cm high each.

You can make this a narrow "tallboy" by mounting woofer in the sides, but instead I suggest that a wider and not so deep box is preferable if you can accommodate it in terms of decor.

An old Dynaudio design (Myrage) implemented a similar design with 2 X 12" Woofer, 2 X 3" Dome and a 1" Tweeter. I heard some recently, they are still an excellent speaker.

View attachment 112674

I would probably suggest much more radical designs, if you were to ask me.

But something like a Myrage or similar, using a fully active system and suitable DSP crossover is 99% assured to give great results and is actually manageable as build.

Finally, a website to consider reading about speakers (not mine):

DIY-Loudspeakers

There are also many excellent projects completely worked and reasonably easy to replicate. Here a scan speak 4-way similar to your ideas, FWIW:

Ellipticor-A50-mkII

Thor
A lot to think about for sure
 
These effects happen typically at frequencies above 50kHz for electrolytics and much higher for film or ceramic caps. We're talking about audio frequencies here.

Of course, but typically, connecting two 100uF caps in series will result in the same ESR as that of a 50uF of similar technology. ESR is roughly inversely proportional to capacitance. Again for similar technology.

Don't worry, Kichhoff is taking care.
There has been a lot of discussion regarding the effects of parallel capacitors in crossovers to improve performance, including drawbacks of using different values in parallel. Example:

https://www.tnt-audio.com/clinica/caps_parallel_e.html
I’m not sure Kirchoffs loop rule applies so well to circuits with crossover inductors and voice coil inductors moving in a magnetic field - maybe Faraday might be a better choice:

 
I would be more proud if I could build a great speaker from lesser or rather cheaper parts.

I have a design I did for a friend. It's the radical opposite of you are looking at.

Drivers:
Eminence Omega Pro 15A as woofer
Audax PR170M0 as midrange
HiVi RT2H-A as HF

First order series crossover at ~ 300Hz & 1.2kHz

Crossover uses third order acoustic slope Highpass and first order acoustic lowpass.

Vented Box.

Big

Loud

Low

Clean

All a proper speaker should be.

I have already begun making a list of possible drivers from SBA, Dayton... Haven't checked SEAS catalogue yet, but they may well be expensive too.

Consider Wavecor. And Pro Audio drivers.

The Audax are classics, as are Eminence. Faital Pro has some excellent choicees.

I really like the HiVi planar drivers over all and any domes.

Here a few of my designs...

DIY, though a few were made semi-commercial with MDF/Granite trunkated pyramid shaped enclosures in Hong Kong.

DIYHIFISUPPLY CRESCENDO MANUAL Pdf Download

commercial

https://old-site.amr-audio.co.uk › ...PDF Abbingdon Music Research LS-77 standmount loudspeaker

Commercial



Commercial



There were some crazy prototypes that never made it out of development for a variety of reasons.

Thor
 
Last edited:
Then clearly you've never been there.

Actually, I got banned, for racism.

Apparently calling Chinese product designed by a Chinese engineer (I am arguably using the term "engineer" in the loosest possible sense) a "Chinese product designed by a Chinese engineer" is not stating a fact, but "underhand racism".

They could not find anything else to ban me for, because, unlike this bunch of audio cargo cult scientists, I get my facts right and indisputable (and I called their BS left, right front and center - on the principle that science exists to be criticised) and did make them look really bad.

So, kindly get your monkey of my back.

Thor
 
There has been a lot of discussion regarding the effects of parallel capacitors in crossovers to improve performance, including drawbacks of using different values in parallel. Example:

If we consider lumped models for anything you might use, any possibile interaction are at 100's of kHz or higher.

Usually they need serious effort to push them low enough for audio to care.

For linear/analogue audio the lumped model is commonly sufficient.


I’m not sure Kirchoffs loop rule applies so well to circuits with crossover inductors and voice coil inductors moving in a magnetic field

It absolutely applies.

Kirchoffs loop rule is a "law of nature" not subject to being repealed. But a speaker driver (especially in enclosures) is a complex impedance (and system). Still, energy escape as Soundwave is part of the electrical equivalent, if sufficiently complex to be useful.

And many things commonly assumed constants are actually signal dependent variables in this model. Thiele/Small is valid at 1mW....

Thor
 
There has been a lot of discussion regarding the effects of parallel capacitors

Be suspicious of any discussion which contains no numbers. Parallel resonance effects between capacitors are absolutely something you need to check when designing power distribution for digital circuits which have current draw into the hundreds of MHz range. Not so much for midrange to tweeter crossovers.
 
Be suspicious of any discussion which contains no numbers. Parallel resonance effects between capacitors are absolutely something you need to check when designing power distribution for digital circuits which have current draw into the hundreds of MHz range. Not so much for midrange to tweeter crossovers.
I was thinking of when you tack together capacitors to make up a required value and use a very small capacitance in parallel with a very large value, each having their own RC and LC values very different - would this not affect the resultant phase angle?
 

Latest posts

Back
Top