Any designs for an RS125 MT?

[orbus]

Hi all,

Well, this weekend I took the plunge and bought a bunch of MDF for the Natalie P towers I'm planning to build. Only thing is - I bought too much wood! After seeing just how huge the things were going to be (it looked a lot bigger in person than it did in sketchup...), I've decided to scale things back a good deal. Since I find myself with a bunch of extra MDF around, I was thinking of maybe trying to squeeze some wall-mount surround-type speakers out of the leftovers.

I was thinking maybe an RS125 + RS28a MT in something like a 12x10x10 box. Assuming 3/4 inch mdf, that's about 12.4L. I modeled that size up for a sealed enclosure with an RS125 in WinISD and came back with an F3 of about 105 Hz, and an F6 of 76 Hz, which really might not be too bad for surrounds backed by a sub. Adding a 1.5 in. port tuned to about 40 Hz (5.5 in.) didn't look too bad either.

Where I fall down is on the crossover - don't know the first thing about designing one. I didn't find any reference designs about for this particular combination, so I took a stab at making my own in Speaker Workshop. I'll be the first to say that I have no clue what I'm doing.



This is (ostensibly) a parallel crossover design using 3rd order highpass and lowpass filters, both at 1800 Hz. The graph Speaker Workshop produced seemed to indicate this actually produced a pretty big dip at 1800Hz, so it might be better to overlap the frequencies a bit - say 1900 LP and 1700 HP or something. I don't really trust the FRD files I used, since they appeared to be taken at wildly different levels (retrieved from here).

If anyone has suggestions, comments, or (best of all) a reference design, please let me know. ;x(

EDIT: Playing with it a bit more, I'm really not convinced I did this right at all, so any comments are definitely appreciated. This is a learning exercise for me.

[JonMarsh]

You've gotta understand, speakers have different senstivities- that is, amount of acoustic output per volts in. They need to have been measured under equivalent conditions.

It doesn't look like you did anything about the upper breakup mode of the RS125- while not as bad as the RS150 or RS180, it still needs some attention (i.e., extra attenuation). Managing the phase in the crossover region is critical to getting a smooth on axis response and tolerable power response. The NatalieP uses 3rd order all pass targets, for comparison. From your component values, did you include any baffle step compensation? Or would you build these in flush to a wall? If the original data you borrowed was IB, then it's not of much use unless you run it through BDS or a similar tool and simulate the baffle step fall off of your intended enclosure design.

Measuring the drivers is even better.... :W

[orbus]

Yeah, not having the drivers actually present makes that difficult. I guess I could buy some to test. I was going to get the RS28a anyway, for the Natalie P.

It was brought to my attention that the design I posted was actually in series, rather than parallel, which was not what I was aiming for at all, and I think why my graph looked so screwy. Like I said, no clue at all what I'm doing.

Second stab at a diagram:



I hadn't even gotten into thinking about BSC yet. I'm currently just trying to figure out if the diagram even looks right. Walk before I run and all that.

As far as the RS125's breakup, it doesn't look too bad to me on the graph, but I'll see about adding some more stuff to knock it down faster in the high end.

I'm really shooting for something very basic here, if I'm going to design this myself. I'm not aiming for perfection - just something that will divide the frequencies up adequately. I know next to nothing about electronics, so I'm mostly relying on Speaker Workshop to point me in the right direction.

I'm really hoping someone's going to jump in with a proven reference design. :W

[HareBrained]

If you've not bought the RS28, consider the Microbe SE design from Roman:

http://www.rjbaudio.com/Microbe%20SE/microbe%20SE.html

It uses the Seas 22TAF/G which is cheaper than the RS28 and IMO, is a nicer driver. And if you're talking about surrounds, the difference in sound will not be noticeable.

[orbus]

I'll take a look at it tonight. I'm not married to the RS28a, just thought it'd be nice to match the Natalies.

I'm quickly coming to the conclusion that designing my own crossover is way out of my league, so a proven design would be most welcome.

[dlneubec]

As one of my first projects, I built the Microbe SE to use as surrounds. They are excellent speakers and are still being used as surrounds in my system. I highly recommend them.

[---k---]

Dan,
Since you have both, would you agree with Hairbrained that the Seas 22TAF/G is better than the RS28?

[dlneubec]

I can't say I agree. They are quite different tweeters in terms of how they can be implemented. The RS28 is very clean and neutral sounding and has the low end grunt to crossover very low. Where I've implemented it, the design needed these things.

The 22TAFG, on the other hand, is probably even cleaner, but has a smaller flange and dome and needs a crossover from 2.5khz and up, I would expect. I can't really compare the implementation in the Microbe SE to any of my speakers, so its hard to say which I prefer.

[orbus]

Thanks for the feedback guys. I was looking around on rjbaudio and noticed he's actually got an alternate design out there for the original Microbe using the shielded version of the RS28a. Link here. Not quite sure which I'm going to do yet.

Also, since I'm still trying to understand how crossovers work, I'm looking at the crossover for the Microbe SE and trying to figure out what all the different bits do.

First, the woofer section, since it's simpler. It looks like the 1.8 mH inductor, and the 4.7 uF cap would be a 2nd order lowpass filter. From modeling in speaker workshop, it looks like .05 mH inductor introduces a notch at about 10K Hz.

The tweeter section looks like a 3rd order highpass filter, with an L-Pad thrown in to attenuate the output some.

This seems understandable enough, even with my very limited knowledge. What I don't understand is why in the description he mentions the crossover as being 4th order, since as far as I can see, neither of the filters are. Unelss I'm just totally missing something.

I also don't see any attempt at BSC - just wondering if that's significant.

[BOBinGA]

The BSC is taken care of by the 1.8 mH inductor. If you were to run a calculation for a regular 2nd order filter using a 1.8 mH inductor, you would find that the crossover frequency would be about 1100 hz. That's where the BSC is starting, so the inductor is starting to roll off the high frequencies at 6 db/octave starting at 1100 hz which compensates for the BSC.

The order of the filter is determined by the electrical filter created by the crossover added to the natural rolloff of the drivers. In this case the fourth order filter Roman is talking about is the total (often refered to as the acoustical) rolloff of the driver and electrical filters.

-Bob