Hello. I am starting to form an idea for a speaker design using Peerless SLS 830668 subwoofers and Usher 8836AC2 midwoofers. Here are the spec sheets on the two:
https://audioalchemy.ro/difuzoare/usher/296-608.pdf
https://www.parts-express.com/pedocs/specs/264-1110-tymphany-sls-p830668-spec-sheet.pdf
How should I go about choosing a crossover point?
My intuition told me that I should aim for a crossover point around 100-200Hz, but looking at a speaker design using the same Peerless subwoofer (Paul Carmody's Tarkus) I saw that he crossed his speaker over at around 400Hz.
I have historically thought of subwoofers or bigger woofers as having the function of picking up the slack that a midrange/midwoofer driver struggles with. I have intuitively thought that so long as a midrange or woofer driver is capable of producing a frequency at high SPL's without much distortion, that it should be given the task of producing those frequencies as it will simply do a better job of reproducing those sounds given that it is smaller and presumably more controlled, with better directivity.
I have read that you should cross a tweeter as low as you can, because if it is capable of producing lower frequencies at a high SPL without distortion, then it will almost invariably do a better job than whatever larger driver it is crossed over with, on account of its smaller size. Here is a quote from Danny Richie of GR Research:
"You are much better off allowing the tweeter to cover as wide of a range as possible and cross as low as the tweeter will allow. And it doesn't matter how good you think your full range driver is. Any tweeter from 2kHz and up will outperform it in every way. It is just flat out going to have better dispersion, speed, and resolution due to its smaller size and much less moving mass"
I know Mr. Richie has a talent for ******* people off, but this makes sense to me. Taking this as true, does this concept not apply to crossing over a midwoofer and subwoofer? Could Mr. Carmody have yielded better results by lowering his crossover point? What reason might he have had for crossing over his Tarkus speaker so high? Or am I just completely confused here? He chose quite a large midrange driver for his Tarkus, so it just seems odd to me that he wouldn't take advantage of its broader bass extension.
Here is Mr. Carmody's writeup on his Tarkus speaker. I bet it kicks *** 😎
https://sites.google.com/site/undefinition/floorstanding-speakers/tarkus
I am new to speaker design, but I am very curious to learn! Thank you in advance for your responses!
https://audioalchemy.ro/difuzoare/usher/296-608.pdf
https://www.parts-express.com/pedocs/specs/264-1110-tymphany-sls-p830668-spec-sheet.pdf
How should I go about choosing a crossover point?
My intuition told me that I should aim for a crossover point around 100-200Hz, but looking at a speaker design using the same Peerless subwoofer (Paul Carmody's Tarkus) I saw that he crossed his speaker over at around 400Hz.
I have historically thought of subwoofers or bigger woofers as having the function of picking up the slack that a midrange/midwoofer driver struggles with. I have intuitively thought that so long as a midrange or woofer driver is capable of producing a frequency at high SPL's without much distortion, that it should be given the task of producing those frequencies as it will simply do a better job of reproducing those sounds given that it is smaller and presumably more controlled, with better directivity.
I have read that you should cross a tweeter as low as you can, because if it is capable of producing lower frequencies at a high SPL without distortion, then it will almost invariably do a better job than whatever larger driver it is crossed over with, on account of its smaller size. Here is a quote from Danny Richie of GR Research:
"You are much better off allowing the tweeter to cover as wide of a range as possible and cross as low as the tweeter will allow. And it doesn't matter how good you think your full range driver is. Any tweeter from 2kHz and up will outperform it in every way. It is just flat out going to have better dispersion, speed, and resolution due to its smaller size and much less moving mass"
I know Mr. Richie has a talent for ******* people off, but this makes sense to me. Taking this as true, does this concept not apply to crossing over a midwoofer and subwoofer? Could Mr. Carmody have yielded better results by lowering his crossover point? What reason might he have had for crossing over his Tarkus speaker so high? Or am I just completely confused here? He chose quite a large midrange driver for his Tarkus, so it just seems odd to me that he wouldn't take advantage of its broader bass extension.
Here is Mr. Carmody's writeup on his Tarkus speaker. I bet it kicks *** 😎
https://sites.google.com/site/undefinition/floorstanding-speakers/tarkus
I am new to speaker design, but I am very curious to learn! Thank you in advance for your responses!
Last edited:
It's not unusual to look at this situation with novice eyes and try to get feedback on which extreme direction to take things, to simplify matters. Take a step back, there are many things at play. Don't believe everything you hear. Those Peerless woofers don't appear to be "subwoofers" and shouldn't have a problem playing higher.
Hmm alright. That sounds like good advice. In lieu of all that I mentioned in my post, how then would you recommend choosing what driver might do a better job at playing that frequency band?
EDIT: rereading your comment, I seem to have slightly missed the point of what you have told me. Thank you for your advice 🙂
EDIT: rereading your comment, I seem to have slightly missed the point of what you have told me. Thank you for your advice 🙂
This is key part of the problem. What does it mean that a tweeter 'allows' a crossover point?
Every driver has two types of distortion. I don't know how they're called professionally, but let's try to imagine it together:
1. "relaxed" distortion - it's the level of distortion that a driver has in it's whole usable band. It's very low, mostly second harmonic.
2. "stressed" distortion - it's the level of distortion that happens at the bottom of driver's frequency response. it starts showing above it's "relaxed" distortion at some frequency and then rises as the frequency falls. You can see it in every distortion plot. What is crucial here is understanding where this "stressed" distortion comes from. The driver's motor and suspension are together very linear and nice in a very narrow range of motion. For a typical, affordable, popular midwoofer it's ok to have 2mm of linear excursion one way, but the very clean music only happens at much lower travel, let's say 1mm. When the driver exceeds this sweet spot, the "stressed" distortion starts creeping in before going suddenly really bad, but even before it get's completely haywire, it shows in the plots as gradually elevated harmonic from, say, 200hz down. There are two things that can make it do so: low frequency and high SPL. The worst scenario is when your music plays low and those lows are loud in the recording and at the same time the high pitched content plays along with the rumble. For example when vocalist is singing while the drummer is going ham. Conversely it doesn't happen when you listen to singers snapping their fingers 🙂
You can deduct from this two things:
1. Our music dictates two paradigms: low crossover point for music that is easy on the speakers and high crossover point for music that takes no prisoners.
2. The question is not only how low, but also how loud. How loud do you want the speaker to play? If quietly, then it's ok to cross low.
But there's more! Baffle step compensation. If you cross your woofer much below baffle step compensation corner frequency, then in some cases you'll end up with lower sensitivity set of speakers. It's worth considering two scenarios and seeing how they compare at various levels.
I like to cross low, because I like to listen quietly. I also like waveguides, which help a lot.
There is a very good presentation of the problem here:
That was incredibly informative! Thank you so much for that writeup! This all makes sense to me, and you bring up stuff I've never thought of before. I really appreciate you taking the time to write this, and I will watch the video you've shared ASAP
I am also giddy to hear another proponent of waveguides! The design I am working on utilizes one 🙂
I am also giddy to hear another proponent of waveguides! The design I am working on utilizes one 🙂
Last edited by a moderator:
I reckon the peerless 10"could go to near 800~1000 judging by the factory specs frequency graph with a steep roll off, probably wouldn't need a 7"mid could go much smaller like a nice higher quality 3~5" mid then bring a tweeter in much higher at something like 3.5~4Khz
This is something that I have thought of, but I just happen to have these drivers kicking around, and for my first speaker design I want to keep it cheap! Plus, I will be using a super cool 1.3" tweeter that I will 3D print a waveguide for, and cross it over at approximately 1.5KHz.
When I eventually move on to bigger and better things, you can be certain that you won't be seeing any 7 inch drivers in any 3 way designs of mine!
When I eventually move on to bigger and better things, you can be certain that you won't be seeing any 7 inch drivers in any 3 way designs of mine!
Last edited by a moderator:
Myriad ways apparently, judging by some of the complex ones I've seen here and on other forums; me I just studied the pioneers that believed in the KISS system using the fundamentals of acoustics to guide them.
No clue how the various forum's multi-way DIYers arrive at XO points, but just looking at the driver's specs, cab widths, the point source drivers' pistonic BW, then doing a basic acoustics calculation:
upper cab width = ~13543"_sec/pi/9" = ~479 Hz
Lower cab width = ~13543"_sec/pi/13" = ~332 Hz
acoustic mean cab width = (479*332)^0.5 = 399 Hz, so 400 Hz a 'no-brainer' 😉 based solely on cab width.
The pioneer's (my) way:
SLS-P830668:
Vc dia = 3.843 cm
Fs = 32.4 Hz
Vc frequency = ~34400/pi/3.843 cm = ~2849 Hz
acoustic mean for equal power distribution = (32.4*2849) = ~304 Hz
8836AC2:
Vc dia = 3.8 cm
Fs = 31 Hz
Vc frequency = ~34400/pi/3.8 cm = ~2882 Hz
acoustic mean for equal power distribution = (31*2882)^0.5 = ~299 Hz
acoustic mean XO point = (304*299)^0.5 = ~301 Hz, so absent any other considerations, 300 Hz works for me to ~preserve the pair's directivity overlap, equal power distribution.
The better question in my mind then is why the widths he chose? BSC is the only obvious choice I can think of ATM.
That said, considering a more accurate way to use cab width is to include the sides impact on its effective acoustic frequency, which is ~1.4142x cab width (the actual math is somewhere on the net, but lost the link):
mean cab width = (9" x 13")^0.5 = ~10.82"
~10.82" * 1.4142 = ~15.3"
~13543"_sec/pi/~15.3" = ~282 Hz, making my 300 Hz much closer, so really don't have a clue ATM why the 400 Hz XO and/or the cab widths were chosen, at least from a technical POV.
Waveguides! So you care about off-axis dispersion, which is also relevant when choosing the xover points, especially between mid/tweeter. Make sure the xover point is where the two drivers have equal (or at least similar) dispersion.
To be honest with you, the extent of my understanding of factoring in dispersion has been to use a wavelength calculator to try to estimate when the mid would start to beam, and then cross somewhere below that. This makes sense though, and I actually have heard from a man name Matthew Poes in a PowerPoint presentation he gave on the Audioholics YouTube channel basically exactly what you said. I will do some research into learning how to do this. Thanks!
Last edited by a moderator:
Bryguy, wellcome to the forum!
To me it looks like you have set yourself a too high task trying to design and make passive 3-way speakers as first project! First of all please study basics of loudspeaker design by reading available websites - asking random specific questions here will just get you confused.
Loudspeakers are bunch of compromises and first of all one should set a goal including specs like like size, price, spl max level, dispersion characterics. Then start to think about how to reach the goal - bass box type, driver number, driver specs, box shape, locating drivers etc. details, matching dispersion and distortion characteristics and so on
And nowdays it might wise to start learning with a good and freeware simulation like VituixCAD2. Lots of info readily in there and with help file and videos, including hundreds of driver measurements.
Some good sources for basics
http://www.mh-audio.nl/PC.html
http://www.2pi-online.de/html/speaker_design.html
http://www.troelsgravesen.dk/Diy_Lo...m#Loudspeakers_in_general/FAQ/tips/ideas,_etc.
http://www.zaphaudio.com/mantras.html
To me it looks like you have set yourself a too high task trying to design and make passive 3-way speakers as first project! First of all please study basics of loudspeaker design by reading available websites - asking random specific questions here will just get you confused.
Loudspeakers are bunch of compromises and first of all one should set a goal including specs like like size, price, spl max level, dispersion characterics. Then start to think about how to reach the goal - bass box type, driver number, driver specs, box shape, locating drivers etc. details, matching dispersion and distortion characteristics and so on
And nowdays it might wise to start learning with a good and freeware simulation like VituixCAD2. Lots of info readily in there and with help file and videos, including hundreds of driver measurements.
Some good sources for basics
http://www.mh-audio.nl/PC.html
http://www.2pi-online.de/html/speaker_design.html
http://www.troelsgravesen.dk/Diy_Lo...m#Loudspeakers_in_general/FAQ/tips/ideas,_etc.
http://www.zaphaudio.com/mantras.html
What GM is talking about is directivity,. He calculates how you can transition from one driver to another in such way that their own peculiar dispersion patterns fuse together into one smooth pattern that doesn't suddenly narrow or widen. It makes the frequency response of the speaker smooth no matter where you are in front of the speaker. You mentioned that you want to use a waveguide in your designe. What if I told you that the reason we call it a waveguide is only because it's meant to manage directivity? Otherwise we'd simply call it a horn. A waveguide is a horn that shapes the directivity of a driver without making the thing particularly more efficient and loud (the thing that horns are most famous for). Waveguides add some gain, but they're specifically used in hifi and studio monitors not to make the speaker particularly loud, but to make the dispersion pattern the way we want it to be. This is why we make this distinction between waveguides and other horns.
As far as I know the term 'accoustic waveguide' was coined by Earl Geddes. If you want to know more about matching drivers' directivity, then suggest you read one of his articles.
By the way, Juhazi has a very nicely controled dispersion pattern in his avatar. You may want to learn how to make one of those heatmaps in VituixCAD with simple frequency response measurements.
As far as I know the term 'accoustic waveguide' was coined by Earl Geddes. If you want to know more about matching drivers' directivity, then suggest you read one of his articles.
By the way, Juhazi has a very nicely controled dispersion pattern in his avatar. You may want to learn how to make one of those heatmaps in VituixCAD with simple frequency response measurements.
Last edited:
^ Wavelength, from speed of sound 343 m/s, direct connection between length and Hz, basis for everything acoustic. Size of sound. I'm about 188Hz tall.
Thank you for your kind welcome! Yessir you are absolutely correct 🤣 I do seem to have a propensity for biting off more than I can chew! I guess it's just not fun for me if I'm not way out of my depth. You are probably right that I should be looking for a more methodical guide, rather than piecing together everything bit by bit by looking through the DIYA forums. I will read all of those resources today! Last night I did read DIYA user wintermute's post titled "So you want to design your own speaker from scratch!" and I learned a lot from that.
I started with using Jeff Bagby's PCD (at the time thinking crossover design was as simple as using the manufacturer provided impedance and FR charts), then tried XSim, but per another member's recommendation I have downloaded the VituixCAD2 software you mention, and I have also learned a lot from reading through its instruction manual. I like that it has box simulation built in.
Anyways, I am perfectly happy for my speaker to live in my mind for as long as it needs to, but I know at some point I will finally have to actually build the speaker. For now, I guess I'm just not comfortable with committing to a design until I feel confident that I've done my due diligence. That said, I have a few drivers kicking around that I am trying to piece another speaker together with, composed of parts that are basically worthless and that I got for free, so I am also giving myself the opportunity to paint with broad strokes without any stakes involved! Hopefully it will be a learning experience for me and help me with this design that I am more passionate about
Thank you Rhalf. I did not realize that was a feature of VituixCAD! I will look into that. I have learned how to read the heatmaps, but I imagine creating one for myself is an entirely different beast! I have heard the name Earl Geddes thrown around a lot, so I will look for some of his material. I guess the source of my confusion was not in what to do, but how to do it. I understood that my objective is controlled dispersion, but I am still a long ways from understanding how exactly to achieve it! I knew that I wanted a waveguide, and that I want to stay out of the range of beaming in the drivers that I use, but beyond that I have been (and am still) basically clueless.
Ok, so my next goal I think is to learn how to match the dispersion characteristics between drivers. I appreciate the help of DIYA members for pointing me towards the resources I need to learn. Have a good day!
There is no magic behind this. It is a bit of work to measure all the off-axis SPL response data. Plotting them in a heatmap or other type of plot is not a big deal, be it VituixCAD or another software tool.