Crossovers for DIY JBL horn system

Hi chaps

This is my first post to the forum so please be gentle with me.

Over the last several months I have been gathering components for an all-JBL horn system.The plan for each stereo channel is:

2 x 2235H woofers in approx 10 cu ft reflex boxes to cover up to 500Hz.
2440 compression driver (with 2440 or 2441 diaphragms) on DIY conical horn to cover 500Hz to 5kHz.
2404H tweet to cover 5kHz up.

I now need to concentrate on the crossovers. Any suggestions? I was thinking of the Marchand XM126 active tube crossover. This is available as a 3 way with a choice of plug-ins which set the slope and crossover frequencies. I'm a tube freak so I'd like to avoid solid state in my system (no offence to the SS crowd).

Forgive my ignorance, but would cards for a 24db slope and 500Hz and 5kHz be OK? Is there any benefit in using shallower slopes? Will the JBL drivers need any other EQ? Is there any benefit in setting up the drivers with asymmetrical slopes and/or frequencies? If so, the Marchand won't be suitable without mods.

Any suggestions welcomed. Thanks already to list members John and Ted (locals).

Many thanks. Doug (Canberra, Australia)

Well, you're certainly asking the right questions. :)

24 dB/oct. at 500 and 5000 Hz should work OK with the drivers chosen... though I might fudge the upper crossover point a bit higher (say, 6000 Hz). You shouldn't damage the 2404 at 5000 4th order (the majority of the JBL PA designs that used the 2404 used it with a second-order slope at 5000 or so), but I would imagine it might sound a bit more "effortless" and "airy" not having to go as low. In the last design I did with one, I was using a 2410 on a flat-front biradial horn with it, and I wound up with a crossover at around 11000 Hz. With the bigger horn driver and flare, you won't want to take the main horn (the mid) THAT high, but I'd experiment a bit with higher crossover points. Keeping the mid horn as the "focal point" of a lot of the bandwidth up top, can help the "coherence" of the system... so letting it do as much of the upper bandwidth as it can do gracefully, can oftentimes be helpful.

As for asymmetrical slopes- yes, that indeed CAN help with the blend between drivers. A higher order lowpass can increase delay time at crossover, which might help the blend between 2235 and 2441. I've found that fourth order woofer lowpass and second order horn highpass can be a good combo, to blend things together.

One thing to keep in mind- while the 2441 diaphragm itself should have no problem with 500 Hz crossovers, make SURE the horn flare ITSELF is capable of supporting 250 Hz (preferably lower horn cutoff than that, actually). The flare should not be "run out" at frequencies below 1 octave above flare cut-off frequency. This can dictate a LARGE flare... but it definitely will sound better. Also, the lower cut-off will allow the use of shallower slopes (ie, 2nd order, 12 dB/oct) at lower frequencies (ie, 500 Hz)...

As for the use of tubes- I concur that tube amps should be very good on the horn drivers (both mids and tweeter), but I'd very strongly advise using a very stout, high-current solid-state amp (with a big power supply) on the 2235s. They can be very "hungry", and the big current capacity of the big transistor amps can be very helpful in keeping them in control. There's a lot of moving mass there, and the additional damping of solid state can go a long way towards making them perform at their best. Plus, the bigger solid state amps can go a long way toward balancing the dynamic output capabilities of the dual 2235s (which will be like, 98 dB/1W sensitivity together) and the horn drivers (which will be 107 dB or so sensitivity for the 2440 and 104 dB sensitivity for the 2404). Given the difference in sensitivity, I'd use an amp about 4 times the power (into a 4 ohm load) on the bass than will be used on the mids and tweeters... for amps that big, solid state is usually a LOT more practical...

As for horn EQ- it'll just be a matter of the characteristics of the mid horn you construct. The 2404 is almost ruler-flat from 5000 Hz and up, and the 2235s should also be very flat. If the mid horn compensates for the rising response inherent to the 2440, then you might not need any EQ at all...

Regards,
Gordon.

Hi Gordon, your reply was very helpful indeed. Do you think I could "get away" with just one 2235 instead of two? The conical horn is going to be quite large - I'm not sure I'm going to have room for a pair of 2235s! Thanks again. Doug

Quote:

---

Originally Posted by Doogster

Do you think I could "get away" with just one 2235 instead of two? The conical horn is going to be quite large - I'm not sure I'm going to have room for a pair of 2235s!

---

One 2235 will play quite loudly and in moderate sized listening rooms I bet you could get all the bass you would require out of one per channel. If you are seeking subsonics that is another story.

Widget

Quote:

---

Originally Posted by Doogster

I now need to concentrate on the crossovers. Any suggestions?)

---

Hi. I have started on a loosely similar system. Well - actually every component is different! - but they're all comparable in each class.
2226 for the 15" bass
2447 for the 4" compression mid
2405 for the compression tweeter.
The best value for money in crossovers in Australia seems to be Behringer.
They make a JBL DSC260 equivalent and a JBL M553 equivalent.
I may try to build my own - for reasons explained below.

Quote:

---

Originally Posted by Doogster

I was thinking of the Marchand XM126 active tube crossover. This is available as a 3 way with a choice of plug-ins which set the slope and crossover frequencies. I'm a tube freak so I'd like to avoid solid state in my system (no offence to the SS crowd).

---

Quote:

---

Originally Posted by Doogster

Forgive my ignorance, but would cards for a 24db slope and 500Hz and 5kHz be OK?
Is there any benefit in using shallower slopes? Will the JBL drivers need any other EQ? Is there any benefit in setting up the drivers with asymmetrical slopes and/or frequencies? If so, the Marchand won't be suitable without mods.

---

24dB/octave Linkwitz/Riley is a very common choice - practically a default standard - because it provides a decent balance of performance. Unfortunately 2 characteristics tend to be somewhat in conflict. The benefit of shallower slopes is that the transient response looks better. The down side is less driver roll-off (not least to protect your MF and HF drivers from undesirable interaction and overload distress and/or destruction).
JBL/Harman researchers feel that the benefits of shallow crossover slopes are minimal - (the waveforms _look_ like you should be able to hear a difference but the _evidence_ is that it's minimally audible at best) while the benefits of steeper slopes are definitely substantial. Their own speakers have mostly moved in the direction of steeper slopes. The older units were typically 12 or 18 dB/octave. The newer, admittedly exceptional, "spare no expense" DMS-1 is 90db/octave!

As to asymmetric slopes and frequency - (I assume you mean, say the upper frequency of the LF and the lower freq. of the MF - not the upper and lower bounds of a particular driver). It is usually considered desirable to keep the _actual_acoustic_ slopes of the 2 drivers the same. Otherwise the relative phase of the 2 drivers will alter in the crossover band and this will lead to variable cancellation/reinforcement and hence response anomalies.
If you cross the speakers well with in their usable bands then the x/over will dominate and the electrical response should also be symmetric. If you cross near the roll-off of the driver then an asymmetric electrical response may be needed to ensure the acoustice response behaves well. i.e if the driver has started to roll of at 12dB/octave then the x/over response should be 12dB/octave also to produce 24dB/octave. This may explain the example in the first reply posted.
Ideally you want to have a measurement system to see how the acoustic response interacts with the x/over. The ideal electrical response is unlikely to be simple - hence why I may have to build my own x/over.

I'm in Canberra - drop me a PM if you want to chat or borrow my measurement microphone.

One point - every conical horn / compression driver combo I have ever heard of or built and measured myself has needed eq to get flat response. Usually a lot of eq - not a little. This will probably be one of the largest influences on the final sound of the speaker if you don't address it. The only suggestion I can make is to measure the response and then go about flattening it as desired.