Help with MTM build using w5-704d and DX25

[kevinm]

Hey guys,

Long time lurker, but first time poster. I've tried to do as much research as I can using existing boards, but I need a little extra help.

For the situation, I met a new family friend (he's well-versed and experienced, but not like you guys) who introduced me to DIY audio and I've loved it. We started this build a while back before I discovered this site, so I hadn't seen the existing builds. I know you guys always tell newbies to start with existing designs, and I definitely will for my next build. But I'd love to do what we can with this existing design if you guys are willing.

My family friend is old school (learned from experience, leverages loudspeaker books and uses formulas for xovers), so I haven't been exposed to the more detailed design using tools and design programs. For my experience, I understand the basics, but there are just more variables than I understand and I'd love to have some help.

We built an MTM design using 2 Tang Band W5-704D woofers and a Vifa DX25. They're crossed at 2k hz LR4 in a 0.77 cubic ft vented box. We are playing around with a 4th order so we could cross the dx25 lower (we liked its lower range detail).

I have been testing it as a center channel, but I'd say we are split 50/50 music vs movie for use. In my not so expert opinion and for the scope of the project, I like how it is turning out. But I would like to tighten up the design to further improve it. In the first crossover run, I didn't use resistors (I have no idea why now that I look back). As a result, the tweeter is much too bright because of the higher senstivity than the two w5's.

Let me know any way you guys can help and what more information you need!

Thanks!

- Kevin

[JonMarsh]

You're at a point where there is really no substitute for measurement. If you know another successful design with the same or similar drivers, you can infer some things from that. I don't mean to denigrate your friends efforts, but using textbook network designs without driver measurements including impedance is going to be hit or miss. I've been doing computer aided design since the late 80's, and have always relied on actual driver measurements in the target enclosure design since the early 70's, though obviously calculating initial crossover networks by hand.

Also, don't forget the impact of cabinet position in your playback environment, as well as the need for baffle step compensation. Check out our reference sections for some design site links. And if you would like inputs on your current design, we need to see specifics, including the crossover network- that at least can be simulated, perhaps using driver data from a similar design. You might try searching on the internet for published measurement data or FRD files, for example.

The advise you've heard about the steps to doing DIY is predicated on the concept that the first step is to get experience doing actual speaker construction- the second, and considerably bigger and more difficult step is doing speaker design, which requires both the craftsmanship skills from the first step as well as a grounding in engineering fundamentals.

[kevinm]

Hey Jon,

Thanks for the reply! No denigration at all. Trust me, I understand the difference in the DIY builds on here and what we/he's done in his past. I'd definitely love to learn the speaker design skill as you're referring to. While I don't have an engineering degree, I grew up with a mechanical engineer and am very comfortable working with my hands. I have the tools to build cabinets (I've built a few now, so I'm comfortable with measuring, cutting,sanding, bracing, and clamping. I have a table saw, a few other saws, router, and sanders. I'm also comfortable with soldering.

I work in the computer industry, so I'm also comfortable with electronics as well.

I have looked online for FRD/ZMA files, but I can't find any. In addition, the guys who have used the same drivers before don't seem to be active online anymore. I can't get ahold of them.

So would you suggest me getting measurements using the box?

[kevinm]

I've found reliable response charts from Zaph that I can trace. So I'll work on those and get back.

I noticed Zaph found different T/S parameters than listed on PE, and it looks like the parameters in winisd suggest an optimal box size of 1.28ft^3 with a lower tuning (about 48hz). So it looks like I should redesign the box.

Here is where I found the info. http://www.diyaudio.com/forums/multi...dy-used-2.html

[kevinm]

Alright, so I've been playing around with PCD and response modeler using the current box setup with the current cross parts. PCD gave me pretty drastically different values for the crossover, but the ones we used were relatively close (relative being a relative assumption on my part).

I added some resistors to crossover to smooth out the response some. There is still a nasty roughness around 3hz on the tweeter. Thoughts? Overall advice? See anything I should change?

[johngalt47]

One thing is to add the DCR to the coils.

[kevinm]

Thanks! I added the values. It looks like it made it slightly worse. What do you think is causing it? Any chance its the baffle? I've seen many members here add a strip of wool underneath the tweeter, do you think that would help?

[cjd]

can you post impedance, and frequency response with phase?

Something doesn't look right here (not to mention the actual component use looks complex for what it's trying to do)

[kevinm]

Is this what you're referring to?

[cjd]

Impedance is very low - too low

Phase, I want to see response graphs with phase, not transfer functions. Hit the "Driver Phase" button - however, the transfer functions show something is off, regardless.

How did you extract phase? What z-axis offset are you using for the W5? Does your impedance data have phase?

[chrisn]

Have you seen this build thread? It uses the same drivers you have: http://www.htguide.com/forum/showthr...818#post413818

[kevinm]

Hi Chris,

Sorry for the late reply, I was busy preparing for my statements build.

I did look at the thread. I tried to contact augerpro to see if he has any FRD or ZMA for either of the drivers. I was going to test out his TM design with the drivers I already have and fiddle with them.

[rj45]

It so happens that I am working on a MTM (RS180s) with a DX25.
Try this - it should roll off the DX25 around 2100 Hz.
+ in --- 6.8 uF ------------------- 4 Ohm ----
|
| .35 mH (20 gauge) to ground.
- in -----------------------------------------------

The DX25 seems easy to work with, and not harsh or finicky.

Hope this helps,

don