Input on driver selection

[Face]

Being a big fan of Vifa's XT series, especially the XT-19, I want to build a full range speaker around it.

For the top end, obviously the XT-19. Mids, I'm looking at the Seas Excel W12CY, and bottom end the Seas Prestige L26RFX/P. Madisound will be drawing up a crossover for me, but I will be picking up some modeling software to touch things up after everything is together.

Some more details. The L26's will have their own cabinets made of 3/4" MDF with heavy bracing such as a two layer front, shelf, and 1" dowels. The W12's will share a smaller cabinet with the XT-19. I'll have to wait to spring, but they'll receive some sort of veneer.

For the price range(s), any better suggestions?

Driver Links:

XT19: http://www.madisound.com/catalog/pro...oducts_id=1179

W12CY: http://www.madisound.com/catalog/pro...roducts_id=700

L26: http://www.madisound.com/catalog/pro...roducts_id=822

Thanks,
Mike

[b_force]

Very nice drivers, although, I would go for another mid-woofer.
Something more in the price-range of the Seas Prestige and the Vifa XT.
The L26's perform very well.

Maybe something like SB Acoustics (SB15NRXC30-8 ) or so?
The Peerless HDS Exclusive is also a great driver.

[Jed]

Those are all great drivers. I'd go with an XT25 and cross around 2k. The XT19 won't go low enough. The Seas Excel driver has some peaks in HD above 2k and can sound harsh if not crossed low. I've got some L26/XT25s laying around here not used at the moment. PM me if you want them.

Also, I tested various 10" drivers in this post:

http://www.htguide.com/forum/showpost.php4?p=427778&postcount=41

The Seas L26 is excellent, especially in the bass. I use it in my Duet10ss design.

Jed

[b_force]

I was also thinking of the XT25.

Do you have a link to a THD test of the W12 Jed ?

[Jed]

I was also thinking of the XT25.

Do you have a link to a THD test of the W12 Jed ?

It was in a Klang and Ton magazine. Everything is clean up to 2.2k, until the spike of 3rd order distortion that is typical of the Seas Excel drivers.

[Face]

I will ask about the XT-25 when I speak to the engineer, thanks all.

[Face]

Jed, I came across your page the other day, linked from the Polk forum. I've had the woofers for a couple of weeks already since I planned on using them somewhere else, but thanks for the link.

[Jed]

Jed, I came across your page the other day, linked from the Polk forum. I've had the woofers for a couple of weeks already since I planned on using them somewhere else, but thanks for the link.

No problem. Let us know what you decide to build. Should be a great system.

[Face]

Speaking of high dispersion tweeters, what does everything think of the Scan-Speak Illuminator D3004/602000 verses the XT19/25?

http://www.madisound.com/catalog/product_info.php?manufacturers_id=148&products_id=8497

Thanks,
Mike

[Jed]

Speaking of high dispersion tweeters, what does everything think of the Scan-Speak Illuminator D3004/602000 verses the XT19/25?

http://www.madisound.com/catalog/product_info.php?manufacturers_id=148&products_id=8497

Thanks,
Mike

I doubt anyone here has heard the Illuminator in a system. It just came out. I can say the SS6600 is an upgrade from the XT25. The timbre match of instruments is more natural with the Air Circ than the XT series. Though for the money the XT25 double magnet tweeter is a value leader.

Jed

[dlr]

Those are all great drivers. I'd go with an XT25 and cross around 2k. The XT19 won't go low enough. The Seas Excel driver has some peaks in HD above 2k and can sound harsh if not crossed low.
Jed

I disagree on the XT25 recommendation. I've used the XT19 LR2 @3k. It's nearly the equivalent of the Hiquphon OW1 in this application. At 2k it needs to be LR4. I had it at that point once, but I didn't spend much time with it as I didn't like the midrange I'd coupled it with. I used it for many months as indicated and found little difference with the OW1. Even the crossover changes needed were minimal.

As far as the W12CY, I've not used it but considered trying it. Some of the specifics of measured distortion characteristics is a bit over-rated as I see it, but that's just my opinion. What I see in the driver indicates that it's easy to use LR2 @3k as well, the way I'd likely use it, but I haven't actually heard it. I would think it to be a very good choice in a 3-way.

The SB is going to be much more problematic in the crossover in comparison, due to the breakup. Even crossing LR4 @2k, there's the issue of linear distortion of any signal harmonics that coincide with it, even though the crossover is low and steep, as long as distortion is being considered. Any motor distortion that might coincide with that range won't be filtered, of course.

The SB looks to be really good for the money, but if I had the budget I'd go for the Seas unit.

dlr

[Jed]

I disagree on the XT25 recommendation.

dlr

Why? Have you used the new double magnet XT25? It works great down to 2k, which is where the W12CY would do well crossing LR4 given it's harmonic distortion peaking around 2.3k I haven't used the XT19, so maybe it can reach 2k LR4, but I can say from experience pairing the OW1 with W15CH was a stretch, when I used it in a 3 way years ago. That system always sounded strained in the upper mids/ lower treble. Maybe the XT19 will fair better, but not worth the chance when XT25 can do it without question.

Also, have you used the SB 5"? It has lower harmonic distortion above 2k than Seas Excel and I managed LR2 just fine. Plus it's a great deal cheaper. Although with an extra notch compared to Scan Speak 15W, which is really smooth above 1k.

[Jonasz]

Jed, can you give some additional info about the OW1/W15 3-way?

[dlr]

Why? Have you used the new double magnet XT25? It works great down to 2k, which is where the W12CY would do well crossing LR4 given it's harmonic distortion peaking around 2.3k I haven't used the XT19, so maybe it can reach 2k LR4, but I can say from experience pairing the OW1 with W15CH was a stretch, when I used it in a 3 way years ago. That system always sounded strained in the upper mids/ lower treble. Maybe the XT19 will fair better, but not worth the chance when XT25 can do it without question.

Also, have you used the SB 5"? It has lower harmonic distortion above 2k than Seas Excel and I managed LR2 just fine. Plus it's a great deal cheaper. Although with an extra notch compared to Scan Speak 15W, which is really smooth above 1k.

I've only used the original XT25. The double magnet is, I assume, to increase senstivity. I've had not need of that.

The issue that I have with the XT25 is the same no matter the magnet. It is too directional in the upper range. This is due to the open ring concept. An equivalent dome has better dispersion. I prefer any of several 25-28mm domes to the XT25. The price of the XT25 is hard to beat, however. One other somewhat questionable benefit of the XT25 is that its directionality greatly reduces its diffraction signature on a baffle.

The XT19 is a tad bit more directional than the OW1, but close enough that I really couldn't hear a lot of difference. To those for whom it matters, the XT19 also has somewhat better motor distortion characteristics, IIRC.

I have an SB17 2-way in progress, no 5". I may buy one (or a pair) to test for my midrange tweak. I do think that it's FR can be improved. I'd treat the SB17, but I don't want it tweaked for this design. That will likely be a later alternative. I will be tweaking the tweeter as it's reversible, the midrange tweak is not. Maybe when this 2-way is done I'll try the 5" and tweak it, but it will have to be as a midrange in a 3-way for how I'd use it.

Dave

[Jed]

I'd take better harmonic distortion profile over a bit of top end air. Different taste I guess.

As for the OW1/ W15CH system; Jonas, what would you like to know? I worked on it with Roman B years ago. Crossover was 2.2k LR4 and LR2 at 300hz to an Excel W22 and then later a RS225. I don't have the files for it anymore... speaker workshop days of long ago :-).

[dlr]

I'd take better harmonic distortion profile over a bit of top end air. Different taste I guess.

If the distortion profile guaranteed it would sound better, so would I. In the case of the XT19 vs. OW1, it doesn't, at least not in my experience. As for the XT25, there are domes with better dispersion and distortion profiles for similar or lower cost such as the Seas 27 series anyway that can also cross lower.

But to each his own.

Dave

[Jed]

If the distortion profile guaranteed it would sound better, so would I. In the case of the XT19 vs. OW1, it doesn't, at least not in my experience. As for the XT25, there are domes with better dispersion and distortion profiles for similar or lower cost such as the Seas 27 series anyway that can also cross lower.

But to each his own.

Dave

Try crossing a Seas 27 series LR2 below 2.2k. Not much low end to work with is there!

For LR4 slopes, yeah the Seas are great and I've used plenty of them. I agree the Seas 27 series are great performers, but the XT25 double magnet will sound good in this system and has very low harmonic distortion to boot, and allows the possibility for shallow slopes (used with SB 5"?). That's a lot for $54. Well, I guess that's why I designed the M5W which uses these concepts.

How do you feel about ribbons? :B

Jed

[Face]

I already started to build the bass bins, but now I'm not sure if I want a modular speaker or a single cabinet for channel.

These two speakers ruined it for me, the recent Trillium thread: http://www.htguide.com/forum/showthread.php4?t=31184


And Tony Gee's Soup Sandwich: http://www.humblehomemadehifi.com/Soup_Sandwich.html


I do like the versatility of bass bins though, in case I decide to go with a different mid and try open baffle, but the speakers above are downright sexy.

Now, how do I make such a pretty face? I would imagine I would need some more bits for my router table and plenty of patience.

Thanks,
Mike

[Face]

Upon more research, it appears that style cabinet is beyond my skills at the moment.

As for now, I'm still waiting for a call back from Madisound...


Btw, what's the general consensus on PE's Woofer Tester 3 and WinSpeakerz compared to Madisound's Leap software? Would they even come close?

[Face]

Well, Madisound came through yesterday, everything is on it's way from them and PE.

About driver placement, how important is: 1/2 wavelength @ x-over freq = driver spacing?

Woofer cab is just about done, just need to finish the face, then it's off to the mid/tweeter enclosure. I'm psyched!

Off to my celebrating...everyone have a safe new years!

[Paul Ebert]

Upon more research, it appears that style cabinet is beyond my skills at the moment.

If you've got a table saw, a sliding (or large capacity) miter saw, and a means for cutting the driver holes (jasper jig or the like), it's not that bad. It is a lot of work, however. Oh, and you'll need a handsaw.

Note that the front baffles on the Trillium are slabs made of 4 pieces of MDF. I think this is easier to make than trying to cut all of the angles required to make the facet pieces. I got this idea from Jon Marsh, who used it for his M8ta's.

Here's what's involved in making the baffles:

1. Cut the 4 baffle pieces to size plus a bit extra for trimming later.
2. Lay out and route the driver holes. Do the holes on the front piece normally and then a bit wider for the woofers on the other pieces.
3. Align and glue the 4 pieces together.
4. Trim them to final length and width on the table saw and miter saw. Doing so now gives you nice clean sides without having to do precision cuts and carefully aligning them as you would if you trimmed them to exact size earlier.
5. Cut the top facet with the miter saw.
6. Cut the other facets with a handsaw as described in the thread.

Like I said, it is a lot of work, but I think it's worth it in the end. :T

[ThomasW]

Btw, what's the general consensus on PE's Woofer Tester 3 and WinSpeakerz compared to Madisound's Leap software? Would they even come close?

LEAP is great if it's using measurements taken with the drivers mounted in the baffles where they're going to be used. If a LEAP design is generated from the mfgr's plots (this is what Madisound does) it's not worth much.

WT3 is a toy, and I know this because I have one. If things like this interest you get WT2

http://www.woofertester.com/wt2product.htm

With devices like WT2 you'll still need a microphone based testing system to measure the frequency response of the drivers in the baffles

[Face]

Thanks for the responses.

I have a router, router table, and table saw, but not miter saw yet. That'll be up next on the tool list.

Madisound did try to compensate for the difference when the drivers are mounted in a actual baffle.

Any recommendations for testing mics?

Thanks again,
Mike

[Face]

Crossovers are up and running. Cabinets should be ready for the drivers in the next few days.

[Face]

Everything is now together.

I'm still using aligator clips for some connections though, but I can't get over how fantastic they sound as is.

The first hour or so was a little rough though. As expected, bass was a little soft and mids were a little muffled. About 5 hours later of various music and sound levels, they're really opening up.

The top enclosure isn't completely done though. It's just a baffle with a mid range enclosure attached to the back at the moment. I've been playing with the angle of the tweeter and mid and believe I found the angle that works best already. I'm going to hate to have to break these down again to finish the top cabinet, install spikes, and round over the edges, but I wanted to hear them before going past this point. I'm glad I went large on the tweeter/mid cabinet though, otherwise I would have no idea how to fit all the crossover parts inside.

Thanks again for everyone's help. I'll post pictures as I continue with the build.

[Face]

Any thoughts?(I feel like I'm talking to myself)

I'm still waiting on my Mic for testing, but just form listening it appears I have a dip in upper bass response. I reversed the phase on the woofer and some songs sound better, and others sound worse. According to the box program I used, I shouldn't have any problems with box or port resonances, but who knows. I'm going to hold off doing any more work on the cabinets till I get some near field and far field FR graphs, just in case I have to build new cabinets.

[Face]

Here's one in all it's ugliness.

[Dave Bullet]

Are those your measurements you posted or Madisounds simulation? Have you got any off-axis curves?

Did Madisound design around a slanted baffle, or have you just decided to put that in the design? Did you give MAdisound your baffle dimensions and driver placement?

If you are spending that much on drivers and premium crossover components, for about 10% the total cost of your system you could probably get a frequency response measurement setup which you would find invaluable for the inevitable tweaking you might want to do.

Dave.

[Face]

Dave,

I have been measuring on my own, I will post the graphs tomorrow. The Madisound crossovers measured ok, but I ended up scraping them for my own design and they measure even better.

The only issue I'm having though is a dip from 1200-2000. No matter what HP I use on the mid, it's still there. Even without an inductor it's still there. I attached pieces to the sides and top of the baffle and it didn't affect it, so I'm starting to believe it could be mid range's sub enclosure. It's only 6" deep, which would somewhat correspond to the frequencies where I'm seeing the dip.

I planned on building new top enclosures anyway, but this time I will try a deeper mid enclosure and see what happens.

To ensure it isn't the mic itself, I tested it on another pair of speakers and there was no dip in that range.

Thanks,
Mike

Mike

[Dave Bullet]

Dave,

I have been measuring on my own, I will post the graphs tomorrow. The Madisound crossovers measured ok, but I ended up scraping them for my own design and they measure even better.

The only issue I'm having though is a dip from 1200-2000. No matter what HP I use on the mid, it's still there. Even without an inductor it's still there. I attached pieces to the sides and top of the baffle and it didn't affect it, so I'm starting to believe it could be mid range's sub enclosure. It's only 6" deep, which would somewhat correspond to the frequencies where I'm seeing the dip.

I planned on building new top enclosures anyway, but this time I will try a deeper mid enclosure and see what happens.

To ensure it isn't the mic itself, I tested it on another pair of speakers and there was no dip in that range.

Thanks,
Mike

Mike

Is the dip related to destructive phase interference from one or other drivers? I can't recall nominal sensitivity of the L12 - but is the level matched with the other drivers? Unless your xo points are wide enough, you will lose sensitivity across the passband for your mid... which can be offset by the tweeter and woofer sections (assuming shallower slopes and good phase alignment). The baffle shape and driver position on it would have more effect on ripple than bafflestep at those frequencies. Is there an inherent dip on the Seas manufacturer data? What does the driver look like naturally (without crossover in place) compared to Seas factory measurements?

The L12 appears to be is equidistant from all baffle edges which will probably provide the widest ranging ripple. An offset might smooth this.

CAn you post some measurements?

[Face]

The red line is the mid mounted on the baffle with just an LP filter(1st order @ 2000hz).

The green line is a -3db LCR filter that is supposed to work from 400-1000hz.

This was taken about 2" from the driver.



I did a little experimenting with baffle width and shape by adding pieces of duct tape to each side, then both sides of the baffle. Then added ripples of duct tape around the mid, but they only appeared to affect higher frequencies(1500hz+).

Work is kicking my butt this week, hopefully I'll have time to build a few test baffles this weekend. FYI, my current TM baffle is approximately 12" wide, 18" tall.

Thank you!

[Face]

Just a quick update.

My FR is almost flat as a board now. I built a 2 piece front baffle, the top half is recessed 3/4", aligning the T and M voice coils. It's kind of ugly since it's just for testing, so pics will come later.


Jed, or anyone else...Before I build a final baffle, I'm still on the fence if I want to use the XT-19 or not. You mentioned the 6600 earlier, is it a no brainer if I have the money?

Thanks all!

[JonMarsh]

Just a quick update.

My FR is almost flat as a board now. I built a 2 piece front baffle, the top half is recessed 3/4", aligning the T and M voice coils. It's kind of ugly since it's just for testing, so pics will come later.


Jed, or anyone else...Before I build a final baffle, I'm still on the fence if I want to use the XT-19 or not. You mentioned the 6600 earlier, is it a no brainer if I have the money?

Thanks all!

If you have the money, yes- no brainer- lower distortion, better off axis dispersion. Just my 0.02.

[Face]

They sounded great before, I'm anxious to see how they sound after.

Thanks Jon!

[Dave Bullet]

Time for more pictures of your work in progress! :T

[Face]

These graphs may be a little premature since I haven't listened to them with the current configuration, but here goes.

Here are my filters:
Xover
2nd HP on tweeter
2nd LP on mid
1st HP on mid
3rd LP on woofer

Measurements were taken .5" from phase plugs. Total FR and impulse are from 6' away.





Is it possible that the midrange peak @ 500hz could be from driver resonance? I'm using Mortite as a gasket and don't have any foam around here to try instead.

Also, other than needing a little more of a drop between the M and T, any suggestions?

[dlr]

These graphs may be a little premature since I haven't listened to them with the current configuration, but here goes.

Here are my filters:
Xover
2nd HP on tweeter
2nd LP on mid
1st HP on mid
3rd LP on woofer

Measurements were taken .5" from phase plugs. Total FR and impulse are from 6' away.


Is it possible that the midrange peak @ 500hz could be from driver resonance? I'm using Mortite as a gasket and don't have any foam around here to try instead.

Also, other than needing a little more of a drop between the M and T, any suggestions?

Have you windowed those measurements? It appears not to me. I haven't used ARTA, but I see that it has a window function. I suggest a half Blackman-Harris if it's in the list. There's no way to know what's going on without windowing to the first reflection. This is going to limit the low end to about 300-350 Hz, but it's needed to analyze the driver responses properly.

Have you chamfered to back of the midrange opening? If the midrange is on a double thickness of baffle to offset it, it's even more important to do this to reduce or prevent resonance of the air in the gap between basket and baffle.

The midrange should not be peaking the way it appears. There should be a step bump, but not a step shelf from 1k down to 250Hz. A small midrange is going to exhibit a fair amount of diffraction, even above the step due to its better dispersion. I've got measurements of a design at my site that shows the SS 12m/4631, nominally close to the W12 in dimensions and dispersion in the midband.

My 3-way

I've not updated any pictures, but the similarity in one aspect may help. I used an offset midrange with the tweeter set back using 3/4" mdf as well. You'll need to use wool felt damping if you offset. However, looking back I would more likely not offset with a small midrange like these. I found that the 12m actually needed a bit of delay due to 3/4" being too far forward. This wasn't bad, because a small delay is easily added by rolling off the top end a bit above the Fc which benefits due to pushing any breakup down even farther as opposed to the more common relaxing of the woofer or mid slope to reduce delay.

Dave

[Face]

Sorry, I didn't window them. I'll look into that for next time.

Yes, the back of the midrange is chamfered. I should have also been clearer on something else, these measurements are from my slanted baffles. The stepped baffles didn't look much different in testing and I wasn't fond of the appearance either.

Thanks for your response David.

[Face]

I can't find a window function in Arta, maybe because I have the demo only?

Here's today's measurement. I raise the tweeter's crossover point and played with the mid a little. No matter what I do, full range, different points/slopes, I can't get rid of that hump at 1k...it's almost like a bad rash. :P
I'm almost at the point where I want to order a different mid range, possibly the Excel W15LY-001 or Excel M15CH-002(Hex magnet). Or I may build yet ANOTHER cabinet, a deeper one...and hope it goes away.

Edit: I can always just brand lines and pick up the ScanSpeak 15M/4531K Revelator 5.5 instead.

But at least I have the tweeter playing in a safe zone. Too bad the cap I'm using at the moment sucks. I can't understand why people think Mundorf caps sound good. :roll:

[dlr]

I can't find a window function in Arta, maybe because I have the demo only?

Here's today's measurement. I raise the tweeter's crossover point and played with the mid a little. No matter what I do, full range, different points/slopes, I can't get rid of that hump at 1k...it's almost like a bad rash. :P
I'm almost at the point where I want to order a different mid range, possibly the Excel W15LY-001 or Excel M15CH-002(Hex magnet). Or I may build yet ANOTHER cabinet, a deeper one...and hope it goes away.

Edit: I can always just brand lines and pick up the ScanSpeak 15M/4531K Revelator 5.5 instead.

But at least I have the tweeter playing in a safe zone. Too bad the cap I'm using at the moment sucks. I can't understand why people think Mundorf caps sound good. :roll:

Take a look at the second pic in your post 36. The third bar has FFT and Window, the latter set to Uniform. The windowing is done on the impulse. You set the end of the window just before the first reflection.

Dave

[Face]

Like this?



Honestly, I'm also baffled(get it?) as to why the mid is doing this. It looked similar on an open baffle too, so it doesn't appear as the enclosure is at fault.

Thanks for your help Dave.

[dlr]

It's getting closer, but I think not there yet. Not knowing ARTA, it's hard to know how it sets up the window. In LAUD you set start and stop time markers directly. In SoundEasy, you set the start marker that looks like yours, then set a window length that stops at the length specified. ARTA looks a bit like SE, you've got the start bar a bit before the initial response. I don't see a length for the window. That would be around 3ms +/-. Also, look for a Blackman in the list. Half-Blackman is what should be used. I guess the 50% means a half window.

Dave

[Face]

I set the window for 5ms, right before the peak on Impulse response. No Blackman in the list, only Hann. I set that to 50%.

Today I changed the mid's LP from LR to Butterworth, it smoothed out more of the dip at 1k. Still need to figure out the shelving though. If I get a chance, I'll pick up some foam gasket material to see if any of it is resonance.

[Face]

Foam did help a little, the peaks are a little smaller.

I built a few other baffles and cannot get rid of that dip at 1k, so I'm going to change things up a little bit, more to come later this week.

[Dennis H]

I'm a bit late to the party. Did you figure out how to set the window in ARTA yet? Left click before the impulse and right click before the first reflection.