Final Measurements and comments

[poneal]

I have finished measuring my RS150 + MDT20 two way and posted results on the PE board here:
http://www.pesupport.com/cgi-bin/config.pl?read=224210
Not the smoothest curves but considering the tweeter costs less than $30 its not bad and sounds nice. You may notice that the tweeter phase is not aligned. I did this on purpose because these are mounted on the wall and the woofer is at ear level. This offset brings the phase back into alignment at my seating position. A non-fatiguing speaker and well worth the $200 I put into it. Regards, Paul.

[kexodusc]

Looks great Paul...I'd love to know how you arrived at the crossover...perhaps a little write-up - did you just start with a text-book 4th order LR and go from there? How many different models did ya run?

I'm messing around withs some dual-chamber reflex designs right now - sub and speakers - I highly recommend looking into this for your next project...very "sealed" like bass from a ported design.
Not hard to build, but predicting excursion behaviour is a bit tricky.

[poneal]

went through multiple iterations. There was a guy that I haven't heard from in awhile that came up with that series notch. I was trying 3rd or 4th orders on the woofer but that notch worked out great. I will keep that in my bag of goodies as it works great for metal coned drivers. The final change was made after measuring them. The tweeter didn't look near as good as PEs data. It's not bad but definitely not the smoothest tweeter. After measuring them, I changed out the woofer inductor to a 1.7 vs the 2.2. The 2.2mH was just too much. Great bass but it affected the crossover and midrange regions. The 1.7mH brought this back up to a flatter response. This design took the longest mostly because I was learning to take measurements. No small feat in and of itself. I thought that I would post this here so that people could see what a real measurement looks like. Not as pretty as what you see in the magazines LOL. They doctor those up with smoothing and farfield with both drivers working. So what your really seeing in most of those graphs are in room response. I measured each driver separately (both impedance and response) and then used this data in the crossover. You may notice that the amplitude is not what you expect. This is because SpeakerWorkshop does things differently. Everything is relative. Well, enough rambling. It was a fun project with a hugh learning curve. Regards, Paul.

[kexodusc]

Yeah, magazine plots have more air-brushing than the covergirls on the magazines in the grocery stores. Oh well. They can be a good point of reference if the magazine does the same procedure for all, though. Did you only try the 2 inductors? How far off were SW's modelling and your actual measurments?

[poneal]

Of course I used the supplied PE data for both drivers so that helped out a lot. It also confirms that PEs clio files are pretty accurate. It also proved to me that modeling software does work. I had 3 inductors in at one time. Had two on the tweeter side. I found it wasn't needed and just took up space so I took it out. Overall, its about a 4th order LR because each driver is about 6db down at the crossover point. As previously mentioned, that series notch is what does all the work on the woofer side. This driver has two nasty breakup nodes to deal with and this filter took care of both of them. What I did was center it between the two breakup nodes so that both end up being attenuated.

My next project will be a paper project. I've done Poly, Metal but not paper. I'm looking at the Seas Line using a 27TDFC tweeter that everyone is raving about and a CA21RLY woofer. I would like to do that scanspeak 8545 but at $165 each, it's a little out of my price range. The Seas woofer is around a 1/3 of that.

Keep us informed on the dual chamber reflex. It sounds interesting.