New Project Diary, M-165x/XT25 2 way bookshelf

Vapor Audio · « **on:** 27 Jun 2007, 05:42 pm »

I'll post details here as I progress in case anyone else is interested. The parts are a single GR M-165x, Vifa XT25TG30-04, and .75cu/ft Parts Express box. I'll be running these sealed, since the 165x needs more volume to gain any significant extension vented - and sealed they'll be easy to integrate with a sub, or would make good surrounds.

One pair of speakers are mounted in the cabinet, below are the results I got measuring the drivers in the box. That's the ideal way to measure drivers, in the box they'll be used, the results are more accurate than a simulated baffle response.

First up, the Frequence Response of both drivers measured at the same volume. These will be the traces I'll use for crossover simulations. The 165x is quite efficient, thanks to a light cone and strong motor. That means I won't have to pad the tweeter down much, which is always a good thing. Also note the 165x breakup that begins around 3300hz. One could probably use the driver with no notch filter, but I plan on implementing one in my design to tame the peak.

Here's the Impedance/Phase graph of the 165x in the sealed box. Of course the Fs will be significantly higher in a sealed enclosure. Hey Danny, what's up with that little squibble in the phase at 500hz that also shows in the frequency chart? Driver is NOT broken in.

Impedance/Phase of the XT25, driver not broken in. I plan on using a series resonance trap to tame the huge imp peak, nice low Fs!

M-165x CSD graph, again shows the breakup starting in the mid-3000hz range. I think it definately needs to be dealt with for ultimate performance.

Now to trace the curves for import into passive crossover designer and come up with an initial network attempt, and then of course lots of listening, measuring, and tweaking!

Vapor Audio · « **Reply #1 on:** 27 Jun 2007, 10:03 pm »

Alright Danny, what do you think about this for the 165x network?

First order electrical - 1.5mh
Series RLC - .25mh, 5uF, 1.67ohm

That gives me a near perfect LR4 rolloff at 2000hz, and puts that breakup 30db or more down.

However, if I do a Zobel with something like a 12uF and 8ohm, the phase and impedance looks much better, but the breakup is only about 12-13db down. So what's the better way?

Danny Richie · « **Reply #2 on:** 27 Jun 2007, 10:39 pm »

In the designs that I have used it in, I used a simple second order network with it, and it made any break up in the upper ranges gone. The break up is pretty mild and the decay rate in that range is fairly quick compared to the break up on other woofers. It behaves more like a peak in amplitude with little resonance issues. I don't think it will be a problem for you. Check your spectral decay after you add your network to make sure that it is well down in output in that range.

Quote

Hey Danny, what's up with that little squibble in the phase at 500hz that also shows in the frequency chart? Driver is NOT broken in.

If it is in the box, then it could be a reflection from the back wall of the box. 500Hz is about 26" long. Is your box about 13" deep?

Also, see how your spectral decay in the lower ranges dies and then picks back up again? This is a reflection from something and it is likely the cause of the peaks in the lower ranges. You can also see from the spectral decay that you need to adjust your start time. Check your impulse response and start your time window to begin right at the beginning if the impulse response and not at zero or no delay. From 1 meter away that will typically mean starting your time window at about 2.7ms. Then gate your ending time window to 6.7ms (catching 4ms total). You may have to gate it sooner if you see any reflection in the impulse response.

Vapor Audio · « **Reply #3 on:** 27 Jun 2007, 10:44 pm »

Alright, thanks Danny ... so I assume you keep the Zobel with a 2nd order electrical?

I've never done a series network, but most of the XT25 networks I see online use series. I have one question about that, all the series networks I've tried modeling have no suckout at all when I invert the polarity of the tweeter. Is that just the way series networks are, or should I see a suckout in the modeled response?

Danny Richie · « **Reply #4 on:** 27 Jun 2007, 10:53 pm »

Quote

Alright, thanks Danny ... so I assume you keep the Zobel with a 2nd order electrical?

There should be no need. try one and see how the response looks. You may also try crossing it in the 2kHz range with that tweeter. You might could even get away with 1,800Hz with that tweeter if using a 3rd order network.

Quote

I've never done a series network, but most of the XT25 networks I see online use series. I have one question about that, all the series networks I've tried modeling have no suckout at all when I invert the polarity of the tweeter. Is that just the way series networks are, or should I see a suckout in the modeled response?

With a series network that is properly in phase, you will get the same suck out when you reverse the polarity that you get with a parallel network.

If you use a series network then you will need to use a higher order series network.

Vapor Audio · « **Reply #5 on:** 27 Jun 2007, 11:10 pm »

Right after I posted that I found a series network that has good phase tracking as well. I'm planning on putting together a few different combinations, since I have a crapload of crossover components to experiment with.

Vapor Audio · « **Reply #6 on:** 4 Jul 2007, 05:45 am »

Well I have to give big props to John (Zaph) for the idea of laying out the crossover on the floor connected with alligator clips for testing. I just finished about 2 hours of testing, changes, testing, changes, etc etc etc. I was able to go through about 30 crossover revisions and came up with this.

Yellow = without notch filter on woofer, Pink = with the notch filter. No other changes were made between those two lines, and as you can see the notch definately smooths things out. As for which one sounds better, I can't say. I'll decide during voicing by listening to some stereo music with and without the notch.

Green line is with notch and many other revisions, bigger inductor - more padding - smaller cap on the tweeter (volume was raised for easier visibility), more baffle step was added by going from a 1.5mH to a 2.2mH, and from there it's pretty much within 1db flat. This is looks to be a crossover design ready for assembly and voicing! Expect a final x-over schematic soon!

Danny Richie · « **Reply #7 on:** 4 Jul 2007, 02:16 pm »

Quote

Well I have to give big props to John (Zaph) for the idea of laying out the crossover on the floor connected with alligator clips for testing.

It's good to know he was helpful to some one. That's what I do as well except that I have the speaker a foot to 4 feet off the floor depending on the size of the speaker, and the network is assembled behind it on a small platform.

Is that a parallel or series network that you are using? Also how far away are you taking the measurements from?

Vapor Audio · « **Reply #8 on:** 4 Jul 2007, 06:03 pm »

Quote from: Danny on 4 Jul 2007, 02:16 pm

Quote
Well I have to give big props to John (Zaph) for the idea of laying out the crossover on the floor connected with alligator clips for testing.

It's good to know he was helpful to some one. That's what I do as well except that I have the speaker a foot to 4 feet off the floor depending on the size of the speaker, and the network is assembled behind it on a small platform.

Is that a parallel or series network that you are using? Also how far away are you taking the measurements from?

This is a bookshelf speaker, so I did the testing with it on it's stand. It's a parallel network, 2nd order with a notch on the woofer - 3rd order with a resonance trap on the tweeter - and tweeter connected out of phase. All the measurements were taken semi-nearfield, about 9-10" away with the mic on the tweeter axis and no gating (I wanted to see the bass response).

Danny Richie · « **Reply #9 on:** 4 Jul 2007, 06:29 pm »

I was wondering how you were getting that bottom end extension.

Better move out to at least a meter to see how the drivers really are summing.

For a parallel network, I'd try second order on the woofer and third order on the tweeter with one resistor in line with the tweeter at the front of the network. This should keep both drivers in phase electrically.

Vapor Audio · « **Reply #10 on:** 4 Jul 2007, 07:40 pm »

I did of course measure at around a meter with some gating, and the curve where the drivers sum looks pretty much the same. Phase tracks fairly well as it is now, with a swing down about 45 degree at the xover point and up about 30 degrees before and after it. Connect the tweeter in phase as opposed to reverse and I get a very nice 30+ db suckout at the xover poin. Minimum impedance is around 4ohm at 2500hz - below that it never drops below 6ohm.

Vapor Audio · « **Reply #11 on:** 5 Jul 2007, 08:47 pm »

More progress, got them up and running in stereo. I've made about 20-25 crossover revisions thus far through a combination of measurements and stereo listening. The big changes have been to the tweeter crossover, I removed the resonance trap because the impedance/phase plot taken with a Woofer Tester 2 was nearly identical with and without it. The 3rd order electrical gets off the tweeter fast enough that it's simply not needed - now all I have is a series resistor after the crossover. I also lowered the 2nd cap value significantly, from 20uf to 11.5uf. Now the tweeter is much smoother through 1500-3500hz, and the sound has a bit more life. The woofer has remained pretty static for the last 10 or so revisions, 2.2mh - 6.8uf - then notch filter.

I've always liked the XT25 sound, it has a bit of snap ... a kinda crisp and lean presentation. The combination of that (and a very slight intentional tip up in the higher octaves) and the full sound shaped into the M-165x woofer gives these bookshelves a very engaging, powerful sound. Right now they're about 6-7 feet out from the wall and sound great, they image very well, have loads of detail, and a very full non-fatiguing sound.

Here are some shots of how I have everything laid out on the floor for testing, like I said this makes things MUCH easier! Changes can be made in seconds, and fortunately I have a massive stockpile of x-over components

JohnR · « **Reply #12 on:** 6 Jul 2007, 03:30 pm »

Clipleads! I love it already!!!

Vapor Audio · « **Reply #13 on:** 6 Jul 2007, 05:24 pm »

Quote from: JohnR on 6 Jul 2007, 03:30 pm

Clipleads! I love it already!!!

Ha, whatever works! And surprisingly they sound pretty darn good even with all those crappy alligator clips barely holding the crossover component leads together.

I'm about 95% done with the crossover tweaking, I should have a final schematic this weekend.

Vapor Audio · « **Reply #14 on:** 12 Jul 2007, 04:21 am »

Crossover is final! I needed a break from the OB2X which is driving me nuts, so spent all day today finishing up the details for this project. I made a few more changes to the network today based on listening and measurements. After I decided on final values, assembled the boards and buttoned it all up.

I really have to say, these speakers sound incredible. I'll hold off on the adjectives until I have somebody else listen to them, but my initial impression is simply stunning. Fast, articulate, musical, with a very deep and wide stage. They have a presentation that instantly draws you in. Being familiar with the drivers prior, I honestly didn't expect results this good ... but I got 'em anyway! 3db down point should be around 53hz in a sealed box, so not tons of bass but what is there is clean and full. As you can see in the pictures, I used all high-end components, ClarityCap SA and AudioCap Theta caps, Goertz Foil inductors - except for tweeter because I didn't have the right value foil, North Creek resistors. I'm awfully proud of how I assembled the boards too

Just FYI, this schematic is after probably 30 revisions, definately not a whatever looks good in the crossover simulator only final design.

Trafford · « **Reply #15 on:** 12 Jul 2007, 11:11 pm »

Turbo,

Excuse my novice ignorance but if you ported the cabinet or increased the size could you get better base results? What if this package were placed in a small floor stander, like the size of the OB2X? I have been very interested in this combo of speakers, as well as the Relevator/XT combo, but I had hoped for a little lower response perhaps around 40-45Hz so a sub would not be necessary for a nice 2-channel setup that can really sing the blues.

Are my expectations unrealistic? I know enough to be dangerous but not enough to answer my own questions or even come close to an answer.

Vapor Audio · « **Reply #16 on:** 12 Jul 2007, 11:47 pm »

Sure you could do this exact build in a vented floorstander. Around 1.7 cu/ft, 3" x 8" port both ends flared should give you about a 35hz tune.

Hank · « **Reply #17 on:** 13 Jul 2007, 04:54 pm »

Nice big breadboard area!

Yes, clipleads! From Radio Shack, I bet - I've had some for at least 15 years and they still come in handy.

Vapor Audio · « **Reply #18 on:** 16 Jul 2007, 11:37 pm »

Another update after a week of listening. I've found these to be quite enjoyable, they work best with a bit of boundary reinforcement - either sit with your head within a few feet of the rear wall, or place the speakers within a few feet of a wall. They image incredibly well, and have loads of detail. The M-165x has a wonderfully musical sound, and the XT25 has a unique character that anybody who's heard one properly implemented can attest too - it has a certain crispness to the leading edge of notes, it does bring a bit of attention to itself, but measures flat and never becomes fatiguing.

This honestly is my kind of sound, a wide and deep stage, detail in spades, fullness to the vocals, and crisp upper octaves. I originally built these to be used as surrounds, but that'll almost be a shame. They will probably end up replacing my Dayton RS180/RS28a MTM's as my office system.

More measurements, can be found in my DIY Projects Gallery. I tried to add as much info to image descriptions as needed http://www.audiocircle.com/index.php?action=gallery;area=browse;album=980&page=3

MerRev · « **Reply #19 on:** 23 Jul 2007, 03:37 am »

TurboFC3S brought over his not so little bookshelfs this past weekend and I think all that attended nodded in agreement- very musical speakers. I'm biased- I lived with the Vifa xt25 tweeter for more than a year and never grew tired of it. It has a unique delivery- atleast with the 3 designs I'ver heard it in. We had the speakers about 18-24" from the front wall and I think everyone was surprised just how much bass we were getting from a sealed two way. Thanks again for bringing them over as everyone really seemed to enjoy them.

Merrev