Constant Directivity speaker #3: Peerless 830869 + BMS 4540ND on JBL mini-PT

Built over the same three-day weekend as the recent DBR fullrange, here is my third brush with CD speaker building. Midbass is a Peerless 830869 8" Nomex-cone in a ~53L tower reflex, better known as a mass-loaded transmission line, MLTL. I'm using a 3" Precision Port. HF is handled by a BMS 4540ND 1" compression driver mounted on the small 6"x6" JBL progressive-transition waveguide. Here are the drivers:



Here they are on my #2 CD project, now sporting its phase-plug, for a few months now:



From the spec sheet response of the Peerless midbass, it looks like it would be suitable for a ~2kHz crossover to the small HF driver, which I think should hold directivity close to that. My actual measurements will tell me the exact story. Worse case, I'll have a small pinch in this area.

Here is a naked enclosure and one with everything mounted:







Construction is fairly simple, butt-jointed 18mm BB-ply with a holey brace down the line and a dowel brace between both drivers. External dimensions are 37.5" height, 13.5" width and 9" depth. I went for a ~golden ratio internal cross-section. Finish is simply waterborne varnish, as I wanted to keep the wood as light as possible to contrast the black drivers and port. I'm using Neutrik Speakon connectors on this one, wanted to try these for a while.

Having one speaker assembled, I'll begin taking some measurements to begin work on the crossover. I'm thinking of using a Crites 3636 autoformers to pad-down the much more sensitive compression driver and already have a crossover layout figured out for that, with staggered impedance at every stage via swamping resistors, thanks to a massive impedance boost at the autoformer primary. I'll post more as things progress.

az

Here is my first set of measurements. I moved some furniture around, set the drivers about halfway to ceiling ~50" up and put the mic 1m away. I think it's the lowest gating I've been able to achieve in my home so far. I rigged a lazy-Susan to set angles more precisely.

Midbass near-cone response to show tuning:



Landed at 35Hz, pretty much what I was going for. I was afraid the port was coming too close to the back of the enclosure and would be mass-loaded, lowering Fb, but it was not the case fortunately.

1m response of the midbass:

Blue trace: 0°
Red trace: 15°
Green trace: 30°



Blue trace: 30°
Red trace: 45°
Green trace: 60°



Here are measurements for the tweeter. These were taken with -10dB pad on the microphone.

Blue trace: 0°
Red trace: 15°
Green trace: 30°



Blue trace: 30°
Red trace: 45°
Green trace: 60°



az

Here are combined midbass and tweeter responses at the same polar angles:

0°



15°



30°



45°



60°



The midbass has been offset by -10dB to represent the actual level difference between drivers. Looks like I won't need a ton of pad-down after the tweeter will have its CD-compensation implemented. I think my original idea of a ~2kHz crossover looks pretty good unless I read things wrong. The midbass is -6dB at 45° at 2kHz and it seems the waveguide's directivity is fairly constant above this, seeming to collapse below this point, where its output is droping fast below ~1.5kHz anyway. Directivity will narrow a bit in the top octave, but it's pretty good up to 30° and not bad at 45° still.

az

I measure the woofer to have a 3.6cm advance on the Z-axis with both drivers mounted on the same plane. With 17.78cm between driver centers, this gives me a vertical lobe ~11.5° above horizontal. This will change when the crossover is implemented, but a bit of elevation to the lobe might not be so bad as the drivers center-point is somewhat below ear level. I ought to try PCD if I can manually produce enough data points for some impedance graphs.

az

With a 11.5° main lobe, it will hit ear level pretty much at the distance my couch is from the speakers, so it might be a good thing not to alter this too much.

az

I have a a pair of Peerless 830491 8" drivers and I wonder if they would work in the same enclosure.

oldskl said:

I have a a pair of Peerless 830491 8" drivers and I wonder if they would work in the same enclosure.

Click to expand...

Specs are not too far off, a bit of tweaking would probably be required though. Looks like it's meant to be more of a subwoofer too, it will likely not make it up to 2kHz as the 830869.

az

They are meant to be used as small sealed home theater subs. My plan was to pair them with Econowave top end but they would need to get up to 1200 hz.

oldskl said:

They are meant to be used as small sealed home theater subs. My plan was to pair them with Econowave top end but they would need to get up to 1200 hz.

Click to expand...

The on-axis response for that woofer looks like the 60° response for mine, but I cannot find other angles for it. I fear it might be a stretch taking it to ~1200Hz with good response at 45°, but I could be wrong.

What could be really cool though, is to simply make flanking subwoofers with these for use under smaller E-waves comprised of the following:

Peerless 830875
JBL 6"x6" PT waveguide
Whatever compression driver hits your fancy, D220Ti being a cheap and known performer

This would be the 6.5" version of the midbass I'm using here. It will hit 2kHz with good 45° response. A small 15L enclosure could be left sealed, with f3~100Hz (fairly low Qtc) or vented to 50Hz with f3~50Hz. The small waveguide is also what I'm using, but my crossover work can only be applied to the HF driver I'm using, BMS 4540ND. I don't know if Zilch or anyone else ever worked out a crossover for D220Ti on this small PT waveguide. Some have used it with the standard E-wave crossover, but some tweaking would probably be needed to really dial things in properly.

Hopefully I'll receive some parts on order before the weekend and be able to start moving things forward.

az

Started to work on the crossover, here's the best I produced so far:





The 120Hz dip is likely floor bounce.

az

You're welcome oldskl. Look around for proven projects, you might find some inspiration too.

az

I think you might be about 90° out of phase at the crossover, the 600Hz hump I'm guessing is the on-axis effect of the baffle width. Go for the impedance measurements! Just set up a jig for LIMP (it's in the manual).

Here is a closeup with phase:



The summed response is calculated by Holm. I don't have enough gator clip lead anymore to do both Xovers at the same time. :thumbsdn: Gotta grab some tomorow.

EDIT, deleted due to error on my part.

az

The phase shouldn't be heading in opposite directions like that. It's probably right for each driver alone, with some time shift error in either direction (which is fine). One way or another, though, the measurement is not indicating the relative phase between them correctly. Maybe you need to lock time zero when measuring the closer driver, or something like that? I haven't used holm in a while.

Yeah, I did not lock time zero with the added plots. :bash: Here is it with time lock.





Probably need a sharper slope on the woofer.

az

Yeah, that's more like it, those are the lines you want lining up with each other (on the listening axis, or an axis near midpoint between driver centers, which I think is the best strategy for these kind of speakers, or really any 2-way).

Thanks for pointing out the blunder, it'll be more productive with that reminded.

That last response is not terrible, but I can probably better it. Trying to get some BSC from the low-pass might be a lost cause as my 13.5" wide cab should have its step around 330Hz.

az

Think I'm done for tonight. How's the following look?





Best on-axis response I could produce so far. Pretty much fits a ±4dB window, midrange and up. There might be a pinch on a sonogram of this between 2khz and 3kHz, but I don't know to what extent yet.

With this arrangement, the tweeter has a 0.7cm advance on the Z_axis, which means a vertical lobe 2.25° below horizontal.

I'll try some more on Sunday.

az

Response looks great, but if that relative driver phase is right, you'll see response issues 2-5k if you measure up and down vertically. Looks like maybe adding an order to the woofer filter and reversing tweeter polarity could get the phase closer to aligned, though I see you're already pretty asymmetrical there.