Designing filters for full range drivers

0 Members and 1 Guest are viewing this topic. Read 1786 times.

Danny Richie

  • Facilitator
  • Posts: 12115
    • http://www.gr-research.com
Re: Designing filters for full range drivers
« Reply #20 on: 29 Jun 2018, 08:29 pm »
Danny we high jacked your thread. What would you do to solve this problem?

Wide band driver, with support at top and bottom?  No crossover in the critical regions with good phase coherence?  Perhaps V2 bass as all agree it works and is economical.

This is going to be a bit of a complex answer.

Here is the way I look it speaker design. It really is about speed. Speed can mean two things. It can mean how fast a driver returns to rest or how fast it can move.

How fast a driver returns to rest is what people are referring to when they say the bass is fast. This can be achieved several ways. Multiple smaller drivers can be used like in a line source or an array like the NX-Treme. With many drivers handling the lower end they each move less so there is less inertia or stored energy to deal with. And smaller drivers means less moving mass. Less moving mass can also speed settling times up. And then for really low end duty there is servo control.  :thumb:

Speed can also refer to how fast a driver can react or track an input signal. If you want to know how fast a driver is just look at how high it will play. If it will react fast enough to play the shorter wavelengths in the top octave then that's pretty fast. If it is not fast enough it simply won't play up very high.

Now I said all of that to get to another point.

Longer wavelengths (low bass) requires more surface area and longer exertions to maintain SPL levels down there. If a single driver is having to do everything (meaning play from 20Hz to 20kHz) then it is having to handle long exertions and everything else all at the same time. Imagine a driver having to handle the top octave while at the same time the cone movement is being dominated by the first two octaves. It becomes a compromise of everything. And the longer exertions mean longer settling times. So speed is hard to maintain down low.

The solution of coarse is to limit the lower end that the driver has to cover. If you just remove the first octave then you have relieved it of about half of the demand on it. Move up one more octave to 80Hz and now you are really taking a load off. The driver can now much more effectively handle everything else.

The upper end of the sweet spot tends to be just below 200Hz. If you take that part off of a driver then you have taken over 80 to 90 percent of the exertion requirements away from it. And along with that a lot of potential stored energy is gone. Setting times are much faster and cleaner and they can more effectively play everything else.

Now take that same line of thought and move it up the scale. Tweeters are much lighter and faster. They have a lot less moving mass as well. Smaller diameter also means better dispersion and off axis response.

No matter how great a full range driver really is it will still get clobbered by a good tweeter in the upper ranges.

So let the tweeter do what it can do. And the lower the tweeter can play the more speed you bring into the lower frequency ranges. This is why I like lower tweeter crossover points. And a Neo 3 in a waveguide crossing as low as 1.3kHz is hard to touch. Now the mid-bass driver doesn't have to be as fast.

And the real range to avoid a crossover is the mid-range. The heart of it is in the 300Hz to 500Hz region. I like to stay below 200Hz and just above 1kHz ideally.

Again it is all about speed and you can see it in the spectral decay.

And yes, drivers like the Neo 3 and Neo 10 are super fast. There' s nothing like them. The Wedgie design using four 3" LGK drivers and the Neo 3 is also super fast. The NX-Otica and NX-Treme models use the M165NQ drivers as mid-bass drivers and don't give up much speed in exchange for a little more body. All of those are favorites of mine.

So as for full range drivers. Some are really good. The LGK's are great as they are small (3" driver) and fast with decent off axis response for a full range driver due to its small size. The compromise is a limited SPL.

The TB has an incredibly smooth response after the filter. But again there are compromises. Off axis response is of coarse very limited. It can play up so high because of the smaller wizzer cone. The main cone is too heavy and damps out the upper ranges. If you had them already in use and wanted to take them up a notch then you could do so by properly implementing a rear firing or upward firing tweeter as a super tweeter. But if adding a tweeter you might as well configure a real crossover and let the tweeter do what it does best. Going at it from the beginning to add a tweeter would have me going for a better mid-bass driver.

WC

Re: Designing filters for full range drivers
« Reply #21 on: 30 Jun 2018, 04:29 am »
Would the 12" Eminence woofers work in the same H-frame as the 12" servos? $300 for 4 woofers is a lot more achievable than the $1500 for the servo woofers and amps, since I already have an amp that should work for the woofers and am already using DSP. How close do they get to the servos in the H-frames? Can they do 20-200 Hz like the servo amps?

jparkhur

Re: Designing filters for full range drivers
« Reply #22 on: 30 Jun 2018, 01:11 pm »
Would the 12" Eminence woofers work in the same H-frame as the 12" servos? $300 for 4 woofers is a lot more achievable than the $1500 for the servo woofers and amps, since I already have an amp that should work for the woofers and am already using DSP. How close do they get to the servos in the H-frames? Can they do 20-200 Hz like the servo amps?

WC. This is what I just did for a friend and it works perfect. See my above post. 

Jonathon Janusz

  • Full Member
  • Posts: 755
Re: Designing filters for full range drivers
« Reply #23 on: 30 Jun 2018, 02:15 pm »
JP,

Because you just did this, and have done both to know for sure, are the driver hole cutouts and bolt spacing the same between the Eminence drivers and the servos?  If so, it might make an even more attractive path for folks to get in on the action - get set up with the Eminence package now to meet the current budget and the servos become a bolt-in upgrade later if someone wants to go all in?  This would also mean that folks could go with Jay and Don's flat packs if they don't have access to the woodworking stuff with the Eminence drivers.

(Aside, nice to know you are still (back?) into the speaker building hobby!)

Danny Richie

  • Facilitator
  • Posts: 12115
    • http://www.gr-research.com
Re: Designing filters for full range drivers
« Reply #24 on: 30 Jun 2018, 03:52 pm »
The Eminence woofers and the servo subs fit the same 11" through hole. The screw holes might be a little off, but they can always be rotated to a new set of screw holes.

Tyson

Re: Designing filters for full range drivers
« Reply #25 on: 30 Jun 2018, 04:27 pm »
This is going to be a bit of a complex answer.

Here is the way I look it speaker design. It really is about speed. Speed can mean two things. It can mean how fast a driver returns to rest or how fast it can move.

How fast a driver returns to rest is what people are referring to when they say the bass is fast. This can be achieved several ways. Multiple smaller drivers can be used like in a line source or an array like the NX-Treme. With many drivers handling the lower end they each move less so there is less inertia or stored energy to deal with. And smaller drivers means less moving mass. Less moving mass can also speed settling times up. And then for really low end duty there is servo control.  :thumb:

Speed can also refer to how fast a driver can react or track an input signal. If you want to know how fast a driver is just look at how high it will play. If it will react fast enough to play the shorter wavelengths in the top octave then that's pretty fast. If it is not fast enough it simply won't play up very high.

Now I said all of that to get to another point.

Longer wavelengths (low bass) requires more surface area and longer exertions to maintain SPL levels down there. If a single driver is having to do everything (meaning play from 20Hz to 20kHz) then it is having to handle long exertions and everything else all at the same time. Imagine a driver having to handle the top octave while at the same time the cone movement is being dominated by the first two octaves. It becomes a compromise of everything. And the longer exertions mean longer settling times. So speed is hard to maintain down low.

The solution of coarse is to limit the lower end that the driver has to cover. If you just remove the first octave then you have relieved it of about half of the demand on it. Move up one more octave to 80Hz and now you are really taking a load off. The driver can now much more effectively handle everything else.

The upper end of the sweet spot tends to be just below 200Hz. If you take that part off of a driver then you have taken over 80 to 90 percent of the exertion requirements away from it. And along with that a lot of potential stored energy is gone. Setting times are much faster and cleaner and they can more effectively play everything else.

Now take that same line of thought and move it up the scale. Tweeters are much lighter and faster. They have a lot less moving mass as well. Smaller diameter also means better dispersion and off axis response.

No matter how great a full range driver really is it will still get clobbered by a good tweeter in the upper ranges.

So let the tweeter do what it can do. And the lower the tweeter can play the more speed you bring into the lower frequency ranges. This is why I like lower tweeter crossover points. And a Neo 3 in a waveguide crossing as low as 1.3kHz is hard to touch. Now the mid-bass driver doesn't have to be as fast.

And the real range to avoid a crossover is the mid-range. The heart of it is in the 300Hz to 500Hz region. I like to stay below 200Hz and just above 1kHz ideally.

Again it is all about speed and you can see it in the spectral decay.

And yes, drivers like the Neo 3 and Neo 10 are super fast. There' s nothing like them. The Wedgie design using four 3" LGK drivers and the Neo 3 is also super fast. The NX-Otica and NX-Treme models use the M165NQ drivers as mid-bass drivers and don't give up much speed in exchange for a little more body. All of those are favorites of mine.

So as for full range drivers. Some are really good. The LGK's are great as they are small (3" driver) and fast with decent off axis response for a full range driver due to its small size. The compromise is a limited SPL.

The TB has an incredibly smooth response after the filter. But again there are compromises. Off axis response is of coarse very limited. It can play up so high because of the smaller wizzer cone. The main cone is too heavy and damps out the upper ranges. If you had them already in use and wanted to take them up a notch then you could do so by properly implementing a rear firing or upward firing tweeter as a super tweeter. But if adding a tweeter you might as well configure a real crossover and let the tweeter do what it does best. Going at it from the beginning to add a tweeter would have me going for a better mid-bass driver.

Bolded the part about the Neo10 and Neo3.  Yes, this is true.  Especially when used together.  There's a "seamless" quality to the sound that just cannot be matched by drivers that have different shapes or that use different materials in construction.  It's like the Martin Logan electrostatics, but with way better soundstage, dynamics and detail.  Of course I'm speaking from my own experience with the Super 7s, where 4 Neo 10s are used per side. 

HT cOz

  • Industry Participant
  • Posts: 1179
Re: Designing filters for full range drivers
« Reply #26 on: 1 Jul 2018, 06:53 am »
Yes it’s $1,700 for the BG Drivers.  I understand it’s great speaker but the steep price of admission keeps many people away.  What about T MM WW instead of MM T MM WW, especially for those of us who are going active? 

Jonathon Janusz

  • Full Member
  • Posts: 755
Re: Designing filters for full range drivers
« Reply #27 on: 1 Jul 2018, 12:31 pm »
Yes it’s $1,700 for the BG Drivers.  I understand it’s great speaker but the steep price of admission keeps many people away.  What about T MM WW instead of MM T MM WW, especially for those of us who are going active?

Here is a link to just about exactly what you're looking for:

https://www.audiocircle.com/index.php?topic=152039.msg1690550#msg1690550

(the latest news from a thread right now two below this one in the GR circle - talked about passive by Danny and active by Rich)

morakot

  • Jr. Member
  • Posts: 23
Re: Designing filters for full range drivers
« Reply #28 on: 9 Aug 2018, 05:20 am »
I spent little over a week listening to these speakers...
I can say It’s better with the crossover.  I like the pace better, the speakers are more coherent and the high is more relax and easier to listen to.










Danny Richie

  • Facilitator
  • Posts: 12115
    • http://www.gr-research.com
Re: Designing filters for full range drivers
« Reply #29 on: 9 Aug 2018, 01:39 pm »
That's a pretty cool paint job on the TB drivers model.

WC

Re: Designing filters for full range drivers
« Reply #30 on: 9 Aug 2018, 04:21 pm »
Morakot,

How do you like the Eikonas?

I have some Eikona 2s that I was going to try in a MLTL cabinet, but I now I was thinking of using it in an open baffle over a pair of 12” eminence woofers in H-Frames. If I like the open baffle presentation I can easily upgrade to the 12” servos subs. If I don’t like OB I can always put the Eikona 2 in a MLTL enclosure and use it in my home theater. I was looking for a not overly expensive way to try out OB to see if I like it like so many of you do.

morakot

  • Jr. Member
  • Posts: 23
Re: Designing filters for full range drivers
« Reply #31 on: 10 Aug 2018, 02:10 am »
Regarding to what you do I would consider the filters.
My Eikona needs the filters more then the Tang band to sound good.
I like my Eikons it lacks bass and bloom. They sound like a very good mini monitor.

Seem like you want to build the open baffle with H-frame. I think that's the way to go.
It needs help with the bass. Even with MLTL there's so much air 5.5”  Eikona can push out.

The Eikona is very good and worth the time and effort.

WC

Re: Designing filters for full range drivers
« Reply #32 on: 10 Aug 2018, 06:38 am »
I was going to start out using DSP so I should be able to add some filters digitally along with applying crossovers for the drivers. I am looking forward to hearing the Eikonas in a speaker soon.

skeeter99

Re: Designing filters for full range drivers
« Reply #33 on: 11 Aug 2018, 01:45 am »
Danny, have you ever heard or measured the Mark Audio drivers? I had the Alpair12's in SuperPensil12 cabinets before and really enjoyed their involving nature. I'd be curious how the newer Generation 3 drivers measure and what could be done with them. They seem to be the cream of the (non-exotic) crop of full range drivers.

Scott

Danny Richie

  • Facilitator
  • Posts: 12115
    • http://www.gr-research.com
Re: Designing filters for full range drivers
« Reply #34 on: 11 Aug 2018, 04:52 pm »
Danny, have you ever heard or measured the Mark Audio drivers? I had the Alpair12's in SuperPensil12 cabinets before and really enjoyed their involving nature. I'd be curious how the newer Generation 3 drivers measure and what could be done with them. They seem to be the cream of the (non-exotic) crop of full range drivers.

Scott

I have had some of those sent to me as well. Someone sent me a popular model in a TL design cabinet with little dots painting on the cone.



They measured pretty rough and had a lot of break up and ringing. This was the on axis response.



And the spectral decay.



I designed a filter for it to take care of those peaks.



I can reduce output in the range with the break up and ringing, but can't fix it. it's still there and it was pretty rough.



I didn't think how it sounded was in the ballpark with other full range drivers that I have tested.

I haven't measured or tested the Alpair 12 but did find a manufactures response for it. It looks pretty choppy. That scale is a 20db scale so those peaks and dips are 8db or more.



The real key is to see what the the spectral decay looks like. But they don't post those.

skeeter99

Re: Designing filters for full range drivers
« Reply #35 on: 12 Aug 2018, 01:56 am »
Danny, those look like some of the very old CHR series of drivers which are the most inexpensive ones. I'd be curious how the newer Alpair10.3 and 7.3 drivers measure on a 5db scale! The 5db measurement you took sure doesn't look good!

Is the picture next to a LGK? How did the sound compared to each other?

Danny Richie

  • Facilitator
  • Posts: 12115
    • http://www.gr-research.com
Re: Designing filters for full range drivers
« Reply #36 on: 12 Aug 2018, 03:38 pm »
Danny, those look like some of the very old CHR series of drivers which are the most inexpensive ones. I'd be curious how the newer Alpair10.3 and 7.3 drivers measure on a 5db scale! The 5db measurement you took sure doesn't look good!

Send them to me. I measure and test them for free.

Quote
Is the picture next to a LGK? How did the sound compared to each other?

Yes. No comparison. The LGK is smooth, relaxed, and natural sounding, with really good detail. No stored energy or break up either. They are really light weight and settle really fast. See the spectral decay.



skeeter99

Re: Designing filters for full range drivers
« Reply #37 on: 12 Aug 2018, 06:49 pm »
Send them to me. I measure and test them for free.

Yes. No comparison. The LGK is smooth, relaxed, and natural sounding, with really good detail. No stored energy or break up either. They are really light weight and settle really fast. See the spectral decay.

I don't currently have any on hand but was thinking of picking up a pair and building the Frugel Horn XL for fun (if I can find a local resource to build the cabinets). If I do, I'll send them your way!