Hello everyone. This is a great site, I enjoy prowling around.
I'm rebuilding my Mission 737R's; a LF/MF driver has failed. So I'm replacing those with SEAS P21RF/P, and replacing the HF units with Morel MDT30.
For the crossover, I plan to use a 2nd order Linkwitz-Riley.
My question is this: should I calculate the component values using the "nominal impedence of 8 ohms for each driver"; or "the supplied R(e) of 6.4 for the SEAS and 5.2 for the Morel"; or "the value from the impedence curve of each at the crossover point I would like to use"???
Also, is there a free/shareware program that I could use to simulate these different ideas.
Cheers!
Mark
I'm rebuilding my Mission 737R's; a LF/MF driver has failed. So I'm replacing those with SEAS P21RF/P, and replacing the HF units with Morel MDT30.
For the crossover, I plan to use a 2nd order Linkwitz-Riley.
My question is this: should I calculate the component values using the "nominal impedence of 8 ohms for each driver"; or "the supplied R(e) of 6.4 for the SEAS and 5.2 for the Morel"; or "the value from the impedence curve of each at the crossover point I would like to use"???
Also, is there a free/shareware program that I could use to simulate these different ideas.
Cheers!
Mark
Member
Joined 2003
Welcome to DIYAudio, Mark.
I think you are getting over your head. Crossover design is not simple, it is what makes the speaker, considered more important than the drivers themselves by many.
I had recently tried to help someone understand the complexities in designing a good crossover, but he did not want to listen. Just put a couple numbers in an online calculator and viola, a great speaker. Not so much.
You may end up with something that makes sound, so it "works" but is far from optimal, and in many cases can even make excellent speakers sound bad. It is much too often that people think that speakers are just that - speakers in a box.
I suggest some research first in the following areas:
-Baffle Step Compensation
-T/S parameters
-Frequency response, impedance
-electrical and acoustic phase
-polar response
Maybe electrical filters as well, some study there after looking at an impedance plot should show you why "nominal impedance" when applied to a filter will not create the expected result. Speakers are not resistors.
What a crossover creates is a transfer function, a frequency response that is subtracted from the driver's frequency response to create your target response. Location of the drivers on the baffle, their "acoustic centers" will affect this response as "phase" is important to determine the frequency response at any axis.
You can design a loudspeaker with software only, with very good results if done right, but don't expect it to be easy. There are lots of resources online, with the help of Excel and some spreadsheets, much can be simulated:
The resources:
FRD Tools
Guides:
RJB Audio guide
Jay_WJ guide
The spreadsheets are not very user friendly, but they include documentation - read it thoroughly. Some of the numbers you enter are not what you will think they should be off the bat, so it is important that everything is done correctly.
Good luck.
I think you are getting over your head. Crossover design is not simple, it is what makes the speaker, considered more important than the drivers themselves by many.
I had recently tried to help someone understand the complexities in designing a good crossover, but he did not want to listen. Just put a couple numbers in an online calculator and viola, a great speaker. Not so much.
You may end up with something that makes sound, so it "works" but is far from optimal, and in many cases can even make excellent speakers sound bad. It is much too often that people think that speakers are just that - speakers in a box.I suggest some research first in the following areas:
-Baffle Step Compensation
-T/S parameters
-Frequency response, impedance
-electrical and acoustic phase
-polar response
Maybe electrical filters as well, some study there after looking at an impedance plot should show you why "nominal impedance" when applied to a filter will not create the expected result. Speakers are not resistors.
What a crossover creates is a transfer function, a frequency response that is subtracted from the driver's frequency response to create your target response. Location of the drivers on the baffle, their "acoustic centers" will affect this response as "phase" is important to determine the frequency response at any axis.
You can design a loudspeaker with software only, with very good results if done right, but don't expect it to be easy. There are lots of resources online, with the help of Excel and some spreadsheets, much can be simulated:
The resources:
FRD Tools
Guides:
RJB Audio guide
Jay_WJ guide
The spreadsheets are not very user friendly, but they include documentation - read it thoroughly. Some of the numbers you enter are not what you will think they should be off the bat, so it is important that everything is done correctly.
Good luck.
Anyway, after what was said, impedance should be at about the one at xover band.
I noticed the online calculator thing as well, and after what I heard/read on this forum, I couldn't really believe the online calculator is so simple. As many explained to me, the x-over function is influenced by the enclosure, the drivers, the music you listen to, the room you are in, etc, none of which is included in those calculations.
Troels uses a program called LspCAD quite effectively. Although it doesn't take into consideration the environment or music type, it does use driver specific data, and enclosure data to model the effect of a x-over you plan to use.
This seems an itterative process, but it allows you to simulate the effects of changes to your x-overs to match what output you desire.
I'm reading up on x-overs now, and it will take a while to get my head around it.
Note: LspCAD6 allows use of real measured data instead of manufacturers data, which means you can plot measured data to see the actual response of your speakers.
Troels uses a program called LspCAD quite effectively. Although it doesn't take into consideration the environment or music type, it does use driver specific data, and enclosure data to model the effect of a x-over you plan to use.
This seems an itterative process, but it allows you to simulate the effects of changes to your x-overs to match what output you desire.
I'm reading up on x-overs now, and it will take a while to get my head around it.
Note: LspCAD6 allows use of real measured data instead of manufacturers data, which means you can plot measured data to see the actual response of your speakers.
re:"the value from the impedence curve of each at the crossover point I would like to use" 
Crossover Design Calculators
L pad calculator - attenuation dB damping impedance decibel loudspeaker - sengpielaudio Sengpiel Berlin
Crossover Design Calculators
L pad calculator - attenuation dB damping impedance decibel loudspeaker - sengpielaudio Sengpiel Berlin
Thanks
Thanks for your advice/inputs, I really appreciate it!
I have done a fair bit of reading on loudspeaker design but that was several years ago, so I will definitely be checking out those suggested links from Dcibel and PeteMck.
Knowing full well that a voice coil is an inductor and that its impedence changes with freq, I thought "what if" about using the Z at the x-over point for the calculations. So thanks for your collaboration Inductor.
LspCAD is now on my hit list ksporry.
I'll eventually report back on my success, or potential lack of!
Cheers,
Mark
Thanks for your advice/inputs, I really appreciate it!
I have done a fair bit of reading on loudspeaker design but that was several years ago
, so I will definitely be checking out those suggested links from Dcibel and PeteMck.Knowing full well that a voice coil is an inductor and that its impedence changes with freq, I thought "what if" about using the Z at the x-over point for the calculations. So thanks for your collaboration Inductor.
LspCAD is now on my hit list ksporry.
I'll eventually report back on my success, or potential lack of!
Cheers,
Mark
Member
Joined 2003
Member
Joined 2003
The best thing someone can teach a "newb" is how to do something right the first time, so that mistakes like using some silly online calculator are not repeated.
For reasons why one should not tell someone to use a simple online calculator, whether they are entering impedance at the crossover freq or not, see post #2.
When I first joined this forum I did not know what I was doing, and made some silly speakers and mistakes...learned from mistakes and made better speakers, eventually came to something that I feel can be described as a great speaker...If only I had done it right or somewhere close to it the first time.
Most will just tell a "newb" to build something designed by someone more knowledgable, like Zaph or Mark K, which is fine if all you are after is a great speaker, but IMO that's more DIY assembly than DIY audio. You don't learn anything from it, and if you are interested in loudspeaker design it will likely get you nowhere.
For reasons why one should not tell someone to use a simple online calculator, whether they are entering impedance at the crossover freq or not, see post #2.
When I first joined this forum I did not know what I was doing, and made some silly speakers and mistakes...learned from mistakes and made better speakers, eventually came to something that I feel can be described as a great speaker...If only I had done it right or somewhere close to it the first time.
Most will just tell a "newb" to build something designed by someone more knowledgable, like Zaph or Mark K, which is fine if all you are after is a great speaker, but IMO that's more DIY assembly than DIY audio. You don't learn anything from it, and if you are interested in loudspeaker design it will likely get you nowhere.
re: 'silly online calculator ' - most online calculators work fine, and give perfectly acceptible results if you know how to use them. Sure, the more knowledge you have the better results you will get because you will understand more of the factors that need to be taken into account. But everyone has to start somewhere, as you yourself have found, you will learn from your mistakes...
Speaker design dosn't need to be rocket science (the 'only an expert can do it right' syndrome), and gross mistakes aside, there's plenty of margin for error, not to mention personal preference...
Speaker design dosn't need to be rocket science (the 'only an expert can do it right' syndrome), and gross mistakes aside, there's plenty of margin for error, not to mention personal preference...
Member
Joined 2003
To each their own...
I wouldn't touch that online calc with a 10 foot pole...if "perfectly acceptable" is "it makes sound" then sure, but not acceptable for me as with using nothing but the calc alone I would have no idea what the result would be.
I don't wish to discuss this any further, I don't believe it will lead anywhere.
Good luck with your project Mark, please share your results.
I wouldn't touch that online calc with a 10 foot pole...if "perfectly acceptable" is "it makes sound" then sure, but not acceptable for me as with using nothing but the calc alone I would have no idea what the result would be.
I don't wish to discuss this any further, I don't believe it will lead anywhere.
Good luck with your project Mark, please share your results.
Dcibel didn't explain well, but the reason he objects so much is because those calculators are designed to use with resistors. They are meaningless cr@p when applied to real world speakers.
Real speakers have impedance that are all over the place, and while the stuff you get from one of those calculators will enable your speakers to play music, they will never ever give optimum results except with ribbon drivers.
Even the voice coil is NOT an "inductor." Voice coils have an impedance component which is actually more like the square root of an inductor. That's because it has nothing to do with the wire being wound in a coil, it is due to eddy current effects in the magnetic structure.
So I think he feels it is a big disservice to newbies to provide links like that without explaining that really they are useless for real speakers, and just provide a basis for understanding classical filter theory, which is the basis for the hardcore-measure-and-crunch-numbers true emulators like LEAP etc.
Back to the original thread topic, I'm wondering why not replace the Mission drivers? Not available? If not, I wonder if the mid and tweet chosen really match the Mission woofer.
FWIW I don't think I've ever seen a crossover schematic from a highly regarded commercial or well documented DIY speaker that used text book filter values.
For example a lot of designs incorporate baffle step compensation absorbed into the low pass section. Also for most designs the needed electrical filter responses is cascaded by the in-box acoustical response needing tailor made Q's of the electrical circuit to achieve the desired order and result.
However that being said, for simple 2 way designs w/o baffle step eq built in, using small really wideband drivers with resonances far from the filter poles and a few other caveats the filter calculators can be used with acceptable results.
EDIT> BUT If people at DIY are going to point others to use filter calculators they should include a standard disclaimer paragraph, without legalese of course. LOL
Last edited:
If the replacement drivers are reasonably similar to the original ie cone size & materiel, Re Le Q Fs Vas I would NOT mess with the crossover. Otherwise if you had to, as a last resort you could just modify a few values on it. Try and gather as much info on the original drivers and post pictures of the old crossover.
Ah ha haha ha ha when does THAT ever happen?!?!?
(OK, once in a while, but not too often. I agree about your disclaimer-maybe there should be a thread to determine standard warnings...)
Re resonances far from filter poles.
IMO You should be aiming for that as a goal otherwise you're asking for increased $$ time and complexity ie zobels, notches, and higher order filters. Also If you pick metal drivers instead of silk, paper and poly then more power to ya.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.