Quick & Dirty 4430-Inspired Two-Ways Part I

[4313B]

B380, 2426J on 2370A, 3110A. Attenuated HF 6 dB, not quite enough yet. Narrow vertical field, 40°, is apparent. Not 4430's, tho, I hear the horns.

A walk down memory lane - my first non-exponential horn system.

I'd forgotten how superbly competent B380's are.

I find that particular enclosure volume tuned up to ~ 29 Hz quite fantastic. I've built several custom pairs of 4430's exactly like that. It's basically Keele's "optimum" calculation for the 2235H (4.55 cubic feet tuned to ~ 29 Hz). Tuning that volume down to 26 Hz and applying the boost at Fb from the B380 is a mild assisted alignment (for a 6th order Butterworth assited alignment according to Keele one would tune that volume down to ~ 20 Hz, 29 Hz * 1/sqrt(2), and apply the high pass bump filter at that frequency). One can "cheat" with the stock 4430 by closing one of the ports and applying a high pass bump filter at ~ 24 Hz. The nice thing about the stock B380 is that it doesn't suffer much power handling loss at all above resonance like a true B6 alignment would. Probably why JBL went with it instead of trying to go solid to 20 Hz.

RED = Stock B380 without BX63
GRN = Stock B380 with BX63
BLU = Stock B380 B6 tuned and assisted

[Zilch]

I find that particular enclosure volume tuned up to ~ 29 Hz quite fantastic. I've built several custom pairs of 4430's exactly like that. It's basically Keele's "optimum" calculation for the 2235H (4.55 cubic feet tuned to ~ 29 Hz).

While I have a BX63A, I've never used it with B380. Natural room response and boundary effect have always made your red curve there work for me.

I'll run the numbers to see what I need to raise them to 29 or 30 Hz. I'm not ready to shorten the ports quite yet. Maybe a rolled insert will get it. And, I'll nose around the plumbing department at the hardware store. Most of my port tuning solutions find inspiration there. Do you have any suggestions?

(For a 6th order Butterworth assited alignment according to Keele one would tune that volume down to ~ 20 Hz, 29 Hz * 1/sqrt(2), and apply the high pass bump filter at that frequency.)

Like the Q=2 6 dB bump high pass in 5235 at 20 Hz, maybe? 5235 with 4430 cards is startin' to look like a good fit here....

[Mr. Widget]

I'm not ready to shorten the ports quite yet. Maybe a rolled insert will get it. And, I'll nose around the plumbing department at the hardware store. Most of my port tuning solutions find inspiration there. Do you have any suggestions?

I recently ran into that very problem with the Project May cabs. I tried putting dowels etc. in the ports, but after taking new measurements I couldn't get the frequency to shift predictably. I ended up cutting new ports that were shorter as I wanted to preserve the originals.

Widget

[Zilch]

1) It seems that LE85 has about the same sensitivity as 2426 (110 dB) when mounted on Biradial horns. I thought I was having to pad down (-6 dB or more) 2426 because it was more sensitive. Turns out it was the HL91 that was eating that SPL in early LE85 designs, apparently. Izzat right? (Note: tired old LE85 sounds mighty smooth and mellow in comparision to 2426. Could be my eBay 2426's have some bad history, tho....)

2) I also erroneously assumed low sensitivity of 2235 was responsible for a significant portion of the padding requirement. It's SOME part, but not so much. Best I can determine, LE15 is 95 dB, 136A is 94 dB, and 2235 is 93 dB. Do we know where LE15B (L200) would fit in there? 96 dB, maybe?

3) I tried the CD horn pre-emphasis on the M552 crossover. As others have observed, it doesn't get it in the last octave. The 80-Series settings are better than the 60-Series, but not good enough.

4) I cannot do this without RTA instrumentation. Even a two-way with these few variables is impossible to get balanced by "ear." After a couple of hours of listening, the ear (and brain) becomes habituated to the current setup and adjustment. Next morning, I turn it on only to discover it's total crapola. Mr. Widget has generously offered to loan me an RTA for the duration.

5) The two crossovers incorporating HF boost I have tried, 3110A and L200B (see pic below) work reasonably well. Can somebody run the sims (voltage drive) on the HF sections of these topologies (3110A at min mid and max HF boost) in comparison to N3134 (4430) please, or recommend what software I should use to run them?

6) N200B sounds pretty good, actually. Maybe I'll open them up and modify for biamping a'la N3134, installing bypass caps and also a rheostat in the HF bypass loop in lieu of the 5-ohms there. Do I need to change the cap in the Zobel to the 14 uF of the N3134, or just let the 8.25 uF ride? I can't believe the impedance peak of 2235 is that different from 136A. (Interestingly, the Zobel stays in when biamping 4430, but NOT when biamping 4435.) Even in detail, the topologies of N200B and N3134 are strikingly similar.

7) And in a final surprise, JBL has acknowledged my order for PT waveguides and OASR horns. Operatively, OASR (Optimized Aperture Symmetrical Radiator) will require CD compensation, no? It's 85° x 85°, probably conical, from the looks of it. How low do it go? We'll see....

[Zilch]

31xxA Curves as published by JBL:

[Zilch]

5235 came with 800 Hz 18 dB cards. From the component list in the manual, it looks like 4430 cards, actually 12 dB, are built on the 18 dB blanks with a couple of jumpers in lieu of poles. Cards are NLA, apprently.

Does anyone have a pair of 4430 cards they're willing to loan or sell to me for this project so's I don't have to tear into these? Or blank 18 dB cards, maybe?

Purdy-please?

If so, PM me to advance this endeavor....

[Zilch]

From the 5235 manual:

"Whenever a midrange or high frequency compression driver is connected directly to a power amplifier, it is recommended that a capacitor be wired in series with the driver in order to attenuate unwanted low frequency and switching transient signals, and to block DC components, any of which can damage the driver. This applies to most biamplified and triamplified systems employing compression drivers. (In conventional systems with passive, high-level dividing networks, the network comes between the power amplifier and the driver(s), so the protection is already built-in.)

Notes:

1) In all JBL monitors which are designed for biamplification, the dividing networks already incorporate the required attenuation capacitors between the input(s) and the compression driver(s).

2) Series capacitors are not required for low frequency and midrange cone or dome type transducers."

The manual specifies a 40 uF non-polarized capacitor for crossover frequencies of 500 to 5000 Hz with 8-Ohm driver, 20 uF for 16-Ohm, and a chart for more precise selection of capacitor values. Then:

"Notes:

1) The optimum value cited in Table 3 assumes that the capacitor is active at approximately one octave below the listed crossover frequency, assuring minimum acoustical interference with crossover region performance while maintaining a good degree of protection.

2) Somewhat smaller capacitor values may be required for additonal protection in high-power sound reinforcement systems.

3) To shunt the reactive component of the compression diriver's impedance below horn cutoff, a 50 W resistor having a value of two to three times the rated impedance of the compression driver should be connected across the driver's terminals.

Caution: Below the cutoff frequency of the capacitor, the power amplifier will be unterminated. If the power amplifier has an output transformer (typical of vacuum tube amps and some solid state amps designed to drive constant voltage lines,) a 20 W resistor equal to 10 times the compression driver impedance should be installed across the amplifier output terminals."

Ummmm, do we actually DO this stuff, or is it boilerplate? Do I need to insert this humungous cap in line? I assume that a capacitor an octave below the crossover frequency is not gonna alter the crossover characteristics, but do we also gotta hang a 50 W resistor across the driver terminals as suggested here for inductive reactance? I DO see banks of shunt resistors in some JBL crossover designs....

[Robh3606]

I have a pair of 4430/35 cards you can borrow to check things out. I am looking for 4434/45 cards for my 5235. Didn't want to rip into those to make them. I know how you feel. Why mess up stock cards!! Send me a PM. I can get them out on the way early next week unless you get a better offer before.

Rob

[Zilch]

Thanks, Rob! Maybe we can track down some 4344/4345 cards to swap here.

[Earl K]

Hi Rob / Zilch

I think that you guys could really benefit by getting in with Ians' PCB program and learning how to do "photo-resist" layouts. The actual core info is still back in his main thread ( it's about the photo-resist kit that you need to buy ). I'm gravitating this way myself for small runs of pcb cards. This way one can simply make up their own 18db cards or whatever. I think the multipin connector on the card is made by "Molex" (?) .

[Zilch]

I think that you guys could really benefit by getting in with Ians' PCB program and learning how to do "photo-resist" layouts.

Thank you, Earl. I'll look that up!

Very likely, we'd be able to build these on perf board, as well. Once I have the 4430 ones in hand, I'll try to replicate them....

[Zilch]

0°, 15°, 30°:

[Zilch]

0°, 15°, 30°:

[johnaec]

'Any idea of what adapter he used to go from 1" to 1.5"?

John

[Zilch]

As I read it, Tim confirms Widget's findings: PH-316's 20 kHz response is down 10 dB at 30° off-axis horizontally, but not so bad at 10 kHz. Alas, for 4430, it's VHF that most matters.

As Ian suggested, the vertical behavior measures substantially better.

Mount 'em sideways.