First attempt at designing a speaker and crossover. I'm sure that I've done it wrong.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey all,

I am going to be putting up a front projector, and in order to have the screen as big as I'd like, I'm going to have to use an acoustically transparent screen and place the speakers behind it. Rather than using in-wall speakers, I'd like to try my hand at designing and building an L/C/R set that will fit behind the screen, get decently loud, and sound good.

I am very new to diy speaker building -- having only just started seriously researching the topic in the last couple days -- so I'm sure that I've made mistakes in my initial design. I'm totally open to any advice/critique you can offer -- up to and including scrapping my current design/driver choices and doing something completely different (although, size and cost are concerns).

My idea is to make three identical MTMs that will mostly fit in cavities that I will cut into the wall, which is framed with 2x4s (there is some wiggle room, as I can build the projector screen to sit some ways off the wall). The drivers that I've designed the enclosure and crossover around are the Dayton Audio DS135-8 5" woofer and the Tymphany XT25BG60-04 1" tweeter. Will 5" drivers work for my needs, or should I be looking 6" or even 7" drivers? (Maybe the 6 1/2" Dayton Audio DC160 or the 7" RS180 -- although, with a larger woofer, would the super shallow depth cause more problems?)


First, the enclosure:

Z9jIwtC.png


I got the dimensions from the product page for the woofer on Parts Express, which lists the golden ratio ported cabinet dimensions as: 10-3/4"H x 6-5/8"W x 4-1/8"D, with a 1" dia. x 4"L port. I'm using a ported design, as I plan to cross these speakers over with my subs at 80 - 85 Hz, and it is the only way to get the low-end extension that I need with these woofers.

So, I just took the dimensions from Parts Express and essentially designed separate enclosures for each woofer within the cabinet. Is there a downside to building the cabinet this way? Also, am I going to regret going with a ported design?


The crossover:

I am sure that I am completely out of my league in trying to design a crossover, but I gave it a shot anyway. I'm just going to post a bunch of graphs and hope that you guys can tell the ways that I screwed up (and hopefully how to fix it :) ).

Schematic:

zMhFwJ9.png


Frequency Response with Phase Curves:

cvAL2bb.png

It drops down 2-3 dB at around 5 kHz. If I changed the resistors in the high-pass to two 2 ohm resistors in parallel, the line stays pretty flat at 90 dB; however, if I do that, I'm worried that I'll deliver too much power to the tweeter. Any suggestions?

Regarding phase: I left the phase source "as measured" (as opposed to "derived"), but I don't know which way it should be done. So, the phase may be inaccurate.

Impedance:

dMTSRx8.png

Am I going to be in trouble because it dips slightly below 4 ohms at 200 Hz?

Transfer Function (I think):

QSX0192.png

I have no idea what I'm looking at here, So I have no idea if this graph looks bad.

Component Power Dissipation (with 100 W fed through the crossover to the speakers):

Fg1oasV.png

This shows that the resistors are getting 20 W, but the resistors that I was looking at buying are only rated at 10 W; is this going to be a problem? Do I need to run four 8 ohm resistors in parallel instead?

----

Sorry for such a long post; if you made it all the way through, thank you.

I've been having a lot of fun learning about speaker building, and this forum has been a huge part of that. Any and all advice you can give will be greatly appreciated. Assuming that I can come up with a design that will work, I'm really looking forward to building my first set of speakers.
 
Regarding phase: I left the phase source "as measured" (as opposed to "derived"), but I don't know which way it should be done. So, the phase may be inaccurate.

I'm assuming your FR curves are from someone elses design or manufacturers specifications?

The only valid scenario to use "as measured" is if you actually measured the drivers in cabinet and did not move the mic between measurements

Otherwise you should use dervived, where you extract minimum phase and include a Z plane offset (to account for voice coil recess of the woofers).

So not quite back to the drawing board, but you have to get your data right for simulation to be correct.
 
The crossover and design looks good.... but!

You need to measure in cabinet. Since this is a two way, I recommend you stick to far-field, gated. Measure the drivers in place.

Make sure you re-measure the impedance of the woofer.

Best,


E

Thanks for the response. That seems like great advice. It will be awhile before I can take these measurements, as I'll have to wait until the room is built. I suspect that I should also have the screen in place in front of the speakers before I measure them.

Is using a multimeter to measure the impedance accurate enough?
 
Administrator
Joined 2004
Paid Member
Other than the tweeter levels coming up a few dB it looks like a starting point. You could consider MiniDSP hardware and a couple of amps and go fully active which would allow you to easily tune the system response to the drivers, room and your preferences.

Subwoofer may be a worthwhile investment for FX. (Kits exist)
 
I'm assuming your FR curves are from someone elses design or manufacturers specifications?

The only valid scenario to use "as measured" is if you actually measured the drivers in cabinet and did not move the mic between measurements

Otherwise you should use dervived, where you extract minimum phase and include a Z plane offset (to account for voice coil recess of the woofers).

So not quite back to the drawing board, but you have to get your data right for simulation to be correct.

I got the frd/zma files for the woofer from the product page on Parts Express. For the tweeter, I used a program called FPGraphTracer, and got them from the spec sheet on the product page.

I assumed that I would need to use derived, but I'm not sure what I should input for the tails and frequency slope settings. I'll have to try to find some documentation or something.

I just guessed at 1/3 inch z plane offset, but obviously I'm going to have to buy the drivers and measure them.
 
Other than the tweeter levels coming up a few dB it looks like a starting point. You could consider MiniDSP hardware and a couple of amps and go fully active which would allow you to easily tune the system response to the drivers, room and your preferences.

Subwoofer may be a worthwhile investment for FX. (Kits exist)

I'll definitely look into going with a minidsp. I planned to buy one to use as a peq for my subs.

Can a single minidsp be used for all three L/C/R speakers?
 
One other thing. The woofers and tweeter are far to wide apart. You do not need to build a separate enclosure for the tweeter.

The two woofers may share an acoustic space so long as the signal to them is identical.

Best,

E

I actually initially had them all in a single space, but then thought that the golden ratio enclosure dimensions that were listed might be important. Putting them all in a single space, should I double the optimum enclosure volume listed for a single woofer, and use two 1" x 4" ports instead of one?

This was my original baffle:

l4wGQ9C.png
 
Just another Moderator
Joined 2003
Paid Member
Another very cheap option (provided you have a sound card with line out and line in) is to make up your own impedance cable as here Cables and use REW to do the impedance measurements. I used to use speaker workshop but found that it's accuracy with some drivers was great and with other drivers was way off. REW seems to be consistent regardless of driver.

Some people find that they need a bigger reference resistor (I'm using 10 ohms without problem, but I often see 100 ohms recommended, I guess it depends a bit on the sound card).

Tony.
 
Last edited:
Yes, if you intend to keep building speakers as a hobby, then investing in a good measurement mic and something like DATS is the way to go. But if these speakers are likely to be a one-off kind of thing, building an impedance measuring jig and just using the measurement mic that comes with most AVR's these days along with REW and your computer's sound card can work just fine.

Keeping in the same money range as the DS135, I think you can do better for your midwoofer.

Dayton RS100-8 - $28
SB 12NRXF25-4 - $29 (on sale @ almost 50%) Run these in series.
Silverflute W14RC25-8 -$21 (spec sheet here)

Ported in small boxes, all 3 give you LF responses well below 80Hz and good high SPL for home theater, especially mounted in-wall without baffle step loss.
 
I'm surprised that no-one has remarked on your low crossover point on that Vifa XT25-60 tweeter. Usually you go a lot higher, say 3kHz, with ring radiators:
Vifa PL14WJ-

I notice you started with near textbook values here, which is OK with wallmounting, since there is no bafflestep. 2-Way Crossover Calculator / Designer

But those sort of calculators usually need flattish drivers and impedance correction to work. Impedance correction is easy enough to estimate from the driver impedance curve. I usually just look at the impedance at 10 kHz and the DCR resistance. MTM is a good way to do home cinema, IMO. Avoids all that unnecessary loudness.
 
Another very cheap option (provided you have a sound card with line out and line in) is to make up your own impedance cable as here Cables and use REW to do the impedance measurements. I used to use speaker workshop but found that it's accuracy with some drivers was great and with other drivers was way off. REW seems to be consistent regardless of driver.

Some people find that they need a bigger reference resistor (I'm using 10 ohms without problem, but I often see 100 ohms recommended, I guess it depends a bit on the sound card).

Tony.

Thank you so much for this info. I'm sure DATS or something similar is invaluable for those who build many speakers (and know what they're doing); however, I think that it's probably more than I need at this time.

I do have a sound card, and it looks like an easy cable to make. Is there a specific type of resistor that works best?
 
Yes, if you intend to keep building speakers as a hobby, then investing in a good measurement mic and something like DATS is the way to go. But if these speakers are likely to be a one-off kind of thing, building an impedance measuring jig and just using the measurement mic that comes with most AVR's these days along with REW and your computer's sound card can work just fine.

Keeping in the same money range as the DS135, I think you can do better for your midwoofer.

Dayton RS100-8 - $28
SB 12NRXF25-4 - $29 (on sale @ almost 50%) Run these in series.
Silverflute W14RC25-8 -$21 (spec sheet here)

Ported in small boxes, all 3 give you LF responses well below 80Hz and good high SPL for home theater, especially mounted in-wall without baffle step loss.

Thanks for the suggestions. Would there be any drawbacks to going with a slightly larger midwoofer? I was looking at the Dayton Audio RS150-4 6".
 
I'm surprised that no-one has remarked on your low crossover point on that Vifa XT25-60 tweeter. Usually you go a lot higher, say 3kHz, with ring radiators:
Vifa PL14WJ-

I notice you started with near textbook values here, which is OK with wallmounting, since there is no bafflestep. 2-Way Crossover Calculator / Designer

But those sort of calculators usually need flattish drivers and impedance correction to work. Impedance correction is easy enough to estimate from the driver impedance curve. I usually just look at the impedance at 10 kHz and the DCR resistance. MTM is a good way to do home cinema, IMO. Avoids all that unnecessary loudness.

My plan now is to measure the frequency response of the drivers in my room, and design the crossover from there.

Thanks for the warning that my crossover point was too low for the tweeter I chose; I had no idea. I'll definitely make sure to set a higher crossover point.
 
I've heard a couple of wallmounted MTM + Subwoofer setups, and liked them. Never built one, but how hard can it be? :D

I ran this Visaton Couplet Light projekte file up the flagpole in Boxsim (Visaton Downloads): Couplet Light – Boxsim Projektdatenbank

Not much margin on tweeter output level (here a 91dB unit) with two 86dB 5" woofers and no bafflestep. Impedance is predictably low. But 3.5kHz crossover looked OK with a butterworth 90 degree phase.

If it was me, I'd use two 4 ohm 5" low-inductance (like SB Acoustics) woofers in series in a common enclosure, and I think it would all be easier. Most woofers are really designed for boxes and bafflestep.

Concept results below. I used 15L closed box per driver, because the W130S happens to be made that way with a highish Qts near 0.5. A reflex would dig an octave deeper, of course.
 

Attachments

  • JustinRT Concept Circuit.PNG
    JustinRT Concept Circuit.PNG
    8.1 KB · Views: 207
  • JustinRT Concept FR.PNG
    JustinRT Concept FR.PNG
    26.6 KB · Views: 218
  • JustinRT Concept Phase.PNG
    JustinRT Concept Phase.PNG
    28.9 KB · Views: 228
Just another Moderator
Joined 2003
Paid Member
I do have a sound card, and it looks like an easy cable to make. Is there a specific type of resistor that works best?

I think that in the Walin Jig II that I made uses a 20W type. Though it is also used for other types of measurements. At the low levels that you are doing impedance measurements at I think that you should be able to use a normal metal film resistor, maybe a 2W to be on the safe side.

regards,

Tony.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.