Quick & Dirty 4430-Inspired Two-Ways Part I

[Guido]

PM me your list with vendors, specs, and part numbers, so we can put together a preliminary parts list to post here, please. Are you using air-core inductors? Are they custom? N/A from Parts Express, for example.

I do order everything at intertechnik, a german company where I have dealer conditions. Therefore it makes no sense to put their parts on the list as you do not have an importer in USA.
Caps:
I'll use MKP (metalized pp) Caps. Bypassed with Polyester and PP Film/Foil caps.
As I bypass I'll use 2 pcs 12uF for C2 and 2 pcs 3.3uF for C1.
Inductors:
I'll use air core types. 0.04mH, 0,15ohms as L4 and 0.56mH, 0,23 ohms as L1

I'll add also a L-Pad for adjustment of the overall sensitivity

2.2 mOhm 0,25W Resistors for the bias voltage (9V)

[Zilch]

If I run 8-Ohm 2431H on a 16-Ohm crossover like 3110A or 3120A with an 8-Ohm non-inductive resistor in series with the driver to make the impedance correct for the high pass, is there any significant consequence other than the 6 dB attenuation (which we need, anyway)?

I don't think we give a whit about damping factor at these frequencies, right?

[This works, by the way. More on it later, after measurements....]

[Zilch]

Our earliest measurements of an all-original LE85 here compared to a pair of 2426J's showed it had a smidgeon more HF extension than those newer design drivers. We also have seen many users assert that the older aluminum-diaphragm LE85 is somewhat smoother and warmer sounding than the titanium 2426, which allegedly has an element of "harshness" by comparison. We wanted to use LE85 in our L200 upgrade, if possible, since it comes with the package.

On initial examination, we doubted that the deep-throated LE85 would play well on the new Progressive Transition (PT) waveguides, especially with two intervening throat adapters required to make the connection. Not so, apparently. The two engineered diffraction constrictions, one in the throat adapter and one in the waveguide itself, seem to allow it to work just fine.

This affords an opportunity to upgrade L200B to constant directivity at little expense. A major feature of 4430's (which, notably, used the pro version of LE85 in their initial iteration,) is their uniformly controlled dispersion of mid- and high-frequencies over a 100° x 100° sound field provided via the trademark 2344(A) "butt-cheek" Biradial horns. The character of the listening experience is very different from that of systems using exponential horns with lenses and "focused" UHF drivers. While there is still a "sweet" spot with 4430's, it is broader and more diffuse both horizontally and vertically than provided with these other systems. The literature also highlights constant directivity's enhanced reverberant field as more like "live" musical performance.

For those who strive to achieve audio nirvana with speaker systems behaving as room-sized "headphones" reproducing and beaming the source material to the listener precisely as recorded, broad-field constant directivity is a bad deal. But, before deciding to convert your L200B's to even more focused three-ways, give constant directivity a try. It can be accomplished for about $225 without doing anything irreversible to your L200B's, and if you decide you like it, an additional $150 will get you quite a decent-sounding system. And if you REALLY like it, we're working on additional upgrades that carry it further.

Here's what you need:

1. A pair of PTF1010HF Progressive Transition waveguides, the rectangular one, 100° x 100°, JBL Part# 338786-001, $83 each.
2. A pair of JBL thread-on 1" to 1.5" throat adapters Part# 339308-001, $19 each.
3. A pair of P-Audio Part# PC-35 or equivalent 1" bolt-on to 1.375"-18 thread-on adapters, $10. http://www.loudspeakersplus.com/html...oadapters.html

And here's what you do:

1. Lay the cabinets on their backs and remove your woofers to gain access to your LE85's and horns. Write down the color codes of the wire connections so you get them correct when mounting them back in.
2. Remove your LE85's and horns. If L200B is like L200, they're held in by the four screws on the front of the baffle and a steel bracket attached at the LE 85/horn connection wood-screwed to a brace across the rear of the cabinet. Look in there to write down the wire color codes, and disconnect them before detaching and wrestling the horn/driver assembly out through the woofer opening.
3. Extend the leads for the LE85 (red and white/black on mine) with about 3 Ft. of additional wire. Use small wire nuts from the hardware store to make the connection. Run the leads out through the port.
4. Block the horn opening with a piece of plywood on the inside (about 6" square will do it) secured with wood screws from the front. Mark and drill pilot holes in the plywood first to make the screwing easier.
5. Reconnect and reinstall the woofers. Stand the cabinets back up in their listening position.
6. Assemble the waveguides to the LE85's using the adapters. The P-Audio adapter mounts to the LE85's first using 1/4"-20 flat-head screws, 3/4" long (I used hex socket screws). Then, the waveguide, bolted to the JBL adapter with 1/4"-20 hex bolts, spins easily on.
7. Set the LE-85/waveguide assemblies on top of the cabinets on something to protect the finish, and reconnect using the wires you've run out through the ports.
8. Set your crossover level control to the 9:00 position (20% on) and play. The adjstment is very sensitive; move anywhere outside the 8:00 to 12:00 range, and the requisite CD compensation is essentially lost.

Exponential horns such as are standard in L200B's automatically compensate for the inherent falling power response characteristic of compression drivers like the LE85 by progressively concentrating the beam width with increasing frequency. Constant directivity horns and waveguides don't do that, rather, they spread all frequencies evenly throughout the defined sound field, and thus, the crossover output must compensate for the progressively falling response by providing increased power at higher frequencies. The L200B crossover was not designed to do that, BUT, as documented earlier in this thread here, at these lower settings of the level control, it produces an approximation of the desired compensation curve, good enough for assessing the decidedly different character of constant directivity listening.

Surprisingly, you'll find that, even at such low settings, the mid/high vs. bass response is easily established; the LE85's are being driven with about 4 - 6 dB less power in the midrange than originally (though, at the highest frequencies, it is the same). The lens formerly consumed that much acoustic output doing its dispersion thing. Net result, there's lowered distortion in the midrange with constant directivity.

Alas, the "approximation" provided by the L200B crossover is not ideal, but it'll give you a different perspective on what is possible with constant directivity on this two-way system. It has a smooth and mellow vintage character, much easier listening, at least for me, than the standard L200 configuration. If you find that you like constant directivity better than what you're used to L200B's sounding like, there's more you can do to "sweeten" the system coming. See pics and RTA curve, below. On the other hand, if you really hate it, another forum member may want the "gear" you bought to try for themselves.

Recommended test disc: "Red Hot & Blue," Chrysalis #F2 21799, a vintage 1990 compilation recording of 22 Cole Porter songs by 20 different artists and styles from Tom Waits to U2, including the Neville Brothers, Sinead O'connor, David Byrne, KD Lange, others. Get it if you don't have it already. There's enough unmitigated sizzle here to singe your eyebrows.

[Zilch]

You'll spend another $150 to acquire a couple of vintage JBL crossovers which WERE designed to provide compression driver compensation, either 3110A (800 Hz) or 3120A (1250 Hz) shown connected in the photo above, and a pair of Parts Express 16-Ohm L-Pads and knobs/faceplates, Part#'s 260-254 and 260-270, respectively, also shown above, front.

Either crossover will work, but they're both compromises. At 800 Hz, the lower limit of the waveguide is pushed. At 1250 Hz, the upper limit of 136A/2231/2235 woofer is pushed (I'm using Giskard's 2235's). Alas, there's no 1 kHz version, though from the sim curves (see link below), 3110A actually operates in that region.

Both of these have a "HF Boost" switch (the "A" designates that,) to adjust the amount of compensation provided to one of three levels, and, similarly, three levels of separate midrange adjustment. Additionally, though, you'll need the L-Pads as a final adjustment of the balance between the mid/high and woofer to compensate for the LE85's having about 4 -6 dB excess sensitivity.

The RTA curve below shows the 3120A with HF boost on max, and the mid level at min, the scenario illustrated earlier in this thread here for 3110A. The L-Pads are at "+2 dB," a 4 dB absolute attenuation (the center "Normal" position being -6 dB). Result: +/- 2.5 dB to beyond 20 kHz.

These settings are too "hot" for my preference. I set the HF boost back to "mid," and the L-Pads at -6 dB "Normal." There's plenty of adjustment in either direction for most tastes and spaces, probably. I pushed this combination to 95 dB SPL, the limit of my tolerance in this listening space, with no apparent deterioration in sound quality.

Frankly, I'm about to cut wood here. I can cut again if I decide to use the larger square PT waveguide later, and the additional upgrades (drivers and crossovers) we're working on all use these waveguides. My L200's already have the cutout on the rear panel for the 31X0 crossovers (to which I'll add bypass caps, and concurrently separate the LF and HF sections for biamping), and the L-pads will go into the front baffle location of the original level switch. For now, I'm gonna play these sweet old LE85's while we continue working with the B380's and 2431's.

[Guido]

Good news Zilch! Thanks for the work.

When do we proceed with the 2431?

[Zilch]

Hi, Guido!

We're waiting for your horns and my additional drivers. I'm actually listening for the UPS truck.

I get NO communication from JBL regarding shipments. Stuff shows when it shows, is all, as other forum members have experienced.

Do you have your AM crossovers together yet and your drivers? I'm anxious to know what you think upon listening, i.e., whether you find this to be as significant an improvement as it appears to me here. You'll have to file the mounting holes on your P-Audio horns just a bit to mount the drivers.

[Guido]

I still wait for the drivers. I have an agreement with JBL here THAT they will be delivered but not WHEN!

Nope, the xovers are not finished. I'm still waiting for some parts for the biasing.

How is Widget going with the driver measurements?

[Zilch]

How is Widget going with the driver measurements?

We also await the additional new 2431H drivers for measuring.

Seems Mr. Widget is doin' more good stuff with Project May right now. Maybe I get a little listen to that at the same time!

[majick47]

I'v been following all the quick & dirty 4430 posts and saw mention of lowering the bass from 30hz to 26hz closely matching the B380. I read that one of the two ports in the L200 could be blocked to get the results. The L200b has only one port and if I read correcly the port tube could be extended to get the same results useing plumbing pvc. Didn't see anything more on the length of the port extension for the L200b. The L300 might also be up for the port extension for 26hz since it shares the same woofer/136a and almost identical cabinet volume. Figure the ports should be the same size in the L200b and L300.

[Zilch]

I'v been following all the quick & dirty 4430 posts and saw mention of lowering the bass from 30hz to 26hz closely matching the B380.

I'm refreshing my recollection as to how we got where we are with respect to bass tuning from early on in this project. It derives from Giskard's post on the subject:

http://audioheritage.org/vbulletin/s...page=4&p=40770

I enhanced the low end by applying the BX63's 6 dB, 26 Hz "bump" filter assist to B380 using the 5235 crossover's built-in HPF boost provisions. Giskard suggests there (I think ) that the same approach could be applied to actual 4430's after tuning them lower by blocking one of their ports.

L200 volume and tuning are similar to 4430. The volume I calculated for each is 5.25 cu.ft. (gonna have to recheck that, tho). Both use 4" dia. ports, L200 being 7.25" long, 4430 being 8.25". (4.5 cu.ft. B380's 4" dia. ports are 9" long.) So, I figured that blocking one of the L200's ports would get me into the same 25 - 26 Hz ballpark.

I've not applied the bump filter to the L200's, but, even unassisted, this may be an appropriate extended-bass tuning for the 2235H woofers installed there. I can always remove the plugs and see how they play. In my view, it's kinda like tuning a passive radiator. You can either have deeper bass or tighter, less extended bass, depending upon the tuning you choose. We're making both options available for application as desired.

Regarding the B380's, I'm not using the 5235 crossover right now, so I programmed what I assume to be a 6 dB Q=2 bump into the UltraCurve at 25 Hz. Looking at the 5235 specifications, Q=2 means it slopes to 0 boost two octaves higher, i.e., at 100 Hz. I may have my head in a dark place on that part, though.

If it's desired to similarly adjust the tuning downward on L200B, since it only has one port, the port must be extended, either by adding on to it with a sleeve on the rear inside the box, or by inserting a smaller diameter tight-fitting tube of the appropriate length in from the front (less desireable, since the possibility of port noise from higher velocity air movement is increased). So, we need to find something in the plumbing department that will fit conveniently over the outside diameter of the port tube at the rear, or something the same size and a coupling device to mate on an extension. You'll need to provide that measurement so we can come up with a plan.

You've already told me that the L200B port is 4.25" diameter (inside diameter, I presume,) and that it is also 4.25" deep from the face of the front baffle. I didn't know JBL used 4.25" diameter ports; I thought 4" and 4.125" were their standards in that range, actually.

WinISD says that's a 30 - 31 Hz tuning, near optimally flat according to the program. To move it down to 26 Hz, it says we need about 7.5" total length. I'm guessing that's a better extended-bass tuning, even without adding boost, but I hope Giskard and others with more experience than me tuning bass boxes (a favorite subject here) will weigh in on this, now that we have more of the specifics for them to work with. I have absolutely NO clue what a B6 alignment is, either, so a tutorial might also be in order for the benefit of all us pilgrims....

[4313B]

Here's what got me started -

Back then all we had were Public/University Libraries to go and read this stuff. Now you can just download it for $5.

*****

Probably better off just ordering the Anthologies though.
The Journals should be at a library too.

Google for more information too.

[Zilch]

http://audioheritage.org/vbulletin/s...ead.php?t=5209

[Zilch]

Tested L200's w/2235H, one port blocked, using 25 Hz, 6 dB, Q=2 (presumed) "bump" filter assist. Works fine, though maybe not quite so fine as using the same on smaller-volume B380. Reopen the port, and the bump filter doesn't do much of anything, as the box is no longer playing down there.

Recap:

1) Closing one port on 4430 lowers box tuning from 34 to 24 Hz (approx.)
2) Closing one port on L200 lowers box tuning from 35 to 25 Hz (approx.)
3) L200B is tuned to approx. 30 Hz, optimal, but the bass can also be extended lower by lengthening the port and applying the bump filter.

Pic shows the filter and how to implement it using UltraCurve GEQ

[AES Loudspeaker Anthologies ordered....]

[Don Mascali]

Zilch:

I use an Ultra Curve too and the PEQ seems to do a better job. Centered at 25hz with a Q of 2.0 and a boost of 6 db seems to match what they use in the bx63. When the GEQ is used there is some filter interaction when drawing a curve. There is also a way to shelve the GEQ that would probably work too.

Of course this is all theoritical on my part since I can only use the RTA function. I don't have any other testing software or instrumentation.

[Zilch]

Thanks, Don. I let U/C draw the curve itself using BW/OCT = 13/3 there, i.e., expanded to include 100 Hz (which I don't claim to understand).

Haven't used the PEQ yet (U/C is new here), but I'll try that now.

Q as used here seems to be a figure of merit relating to crossover performance, not the fc/delta f3 filter parameter that I can discern. BW/OCT behaves reciprocally.

Yup, PEQ works, and sounds better, too. The BW/OCT parameter is unfathomable, though. There's nothing in the manual detailing it to explain why it's different, or what it means, even.

I'm just trying to replicate the Q=2 curve shown for the bump filter in the 5235 manual with it. PEQ looks closer, for sure, if not spot on. BW/OCT=1. Good call: