Sanity check on a PA build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I’ve been looking into building a custom PA system lately, and I just wanted to make sure I’m at least somewhat in the ballpark before I start cutting wood and buying components.

I intend to be able to use this system in small/medium clubs and event spaces, mainly for techno & experimental music. As a basis of comparison, in the past I’ve used a Matrix B-52 1000 system as well as a single QSC K181 sub with K12 tops, and both were adequate in terms of output for most rooms, although more headroom would be welcome.

Right now, I’m looking at the following:

Sub: Faital 18XL1800 in a sealed enclosure, probably ~3ft^3 or slightly larger (qtc between .707 and .5)
LF: Eminence Kappalite 3012HO in modified JBL4530 ‘scoop’ enclosures
HF: Radian 850PB in EV HP9040 horns

The sub and sealed enclosure have been selected with extended low frequency response in mind. The 18XL1800 is a pro driver with good power handling and a particularly high xmax (20.17mm). From what I understand, after EQ’ing and in rooms small enough to have room gain, such an arrangement will likely yield the deepest low frequency extension practical. I think this is an uncontroversial design, but if I’ve overlooked something I’m open to input.

The LF enclosure design is the biggest open item right now for me... I started by redrawing the JBL 4530 plans available online in a parametric CAD package, then scaling and stretching the design up and down. A few things of note during this exercise:

First, the peaks/dips inherent to the RLH design seem to get out of hand with the speaker radiating in full space, so it seems like having them sitting on the floor is a must. With that said, if the direct-radiating portion is supposed to play up to 500-800hz, and we’re talking about a PA system here, I probably want to build relatively tall cabinets to get the driver up high, right? So, assuming we want to fit more than one panel onto a single sheet of plywood, we’re talking about something like an 18” wide by 26” deep by... say 54” or so tall scoop cab. Making the horn this narrow means the aspect ratio gets appreciably different from the original JBL cabinets. Something like a 18” wide by 36” tall mouth with the same 30* exit angle still seems reasonable though, correct?

Next, I noticed that the original 4530 design has a somewhat large volume inside the chamber behind the driver, which necks down into a smaller throat before expanding back up. My initial impression is that it seems better to eliminate that throat and shuffle around the folds so that there’s a constant reduction in section the whole way down, so you turn the rear horn into a sort of offset horn design instead of the chamber / throat adapter configuration of the original plans. Would this be a warranted design change, or am I going in the wrong direction here? Simulating this change in hornresp seems promising, but maybe I’m missing something as I’m not well-versed in loudspeaker design by any means.

Finally, the highs: I can get the EV horns semi-locally, and I’m extremely tempted to pick them just based on cool-factor alone. At the same time though, PA usage in reverberant/untreated environments isn’t such a bad application for these large-format horns, right? Like, it’s overkill based on my needs, but it’s not as absurd as using them in a living room hi-fi system, which is definitely a thing that people do. For PA usage, would scoops plus a sub to fill in the low end pair well with the horns?

Any input is much appreciated, thanks in advance!
 
I would never use a scoop, ever. Never heard one that sounded any good and you don't need the LF loading because you have the Faital 18's.

As for the EV's, do you really want to hump in/out a huge old diffraction horn? There are better modern designs in JBLs own catalogue that can be picked up as spare parts.
 
I'd think very very carefully before using sealed boxes for PA systems.

Having used the same drivers (Beyma 15P1200Nd) in both, ported boxes eat the sealed boxes for breakfast. It's not even a comparison until you've got 3x sealed vs 1x ported. Give each cabinet a kilowatt, and it's pretty close. Of course, that means the sealed boxes have 3x 15" drivers and 3KW while the ported box has 1x 15" driver and 1KW with comparable output.
That's not even considering tapped horns, which can do the work of 2x ported boxes with one driver.

+1 on Brett's comments.

You've got some good drivers lined up, but your implementation needs work.

What do you have in the way of power amplifiers and processing?

Chris
 
Your sub should be a reflex or tapped horn, this bandwidth is where the most output is required so this box should have the highest potential output in the system, this also usually means this is the biggest box in the system. Don't cripple your system by selecting the most inefficient design possible.

Your low/mid should be sealed or reflex, there is no point at all to using a scoop here it will just make a total mess of the response. Consider using a stack of dual 10" or dual 12" drivers to get more height for the CD.

Nice compression driver, I really like the performance delivered by one of these and that performance does require a big horn but the one you seem to be all hot for is going to cause a lot of headaches in a mobile application as it won't stack easily.
 
Great tip on purchasing modern JBL horns/waveguides as replacement parts, I hadn't realized how inexpensive they are. A local seller had the EV horns available at a remarkable price, but after considering alternatives the price seems not to be so remarkable after all.

For amps, I was planning on something like the QSC RMX2450a bridged for the sub (1300w FTC @ 8ohms), then another for the mids (450w FTC @ 8ohms both channels driven), then maybe something like a samson servo for the compression drivers. Anything beyond the 2450a seems to become cost prohibitive, and at that point power consumption seems to become sketchy anyway (15A circuits are standard in the US, and may be all that are available in some venues).

Processing would be something like a DBX driverack or that one behringer unit, or a similar 19" rack unit. Planning on electronic crossover and EQ, in any case.

I suppose maximum SPL isn't as important to me as detail and tone, and the best low extension possible would be nice. I'm not going to be trying to outdo every other event in terms of volume, most parties around here are way too loud anyway. Basically everyone dancing is going to be easily within 10m of the front of the speakers, so maybe I don't need 130dB, you know?

I had selected a sealed box for the sub due to the smoother rolloff, I had done some reading on here (albeit in the hifi section) about sealed subs being able to reach very low with limiting / EQ'ing, particularly in smaller rooms where room gain comes into effect. Most venues available to me are going to be relatively small rooms, so this seemed enticing. I had shyed away from tapped horns as it seemed like to get down to like 30hz with them the horn length just gets unreasonable, but maybe this isn't so.

Looking on the JBL site, the mid/high section from the 3732 which is sold in the Marquis Dance line as the MD2 seems really enticing, but then again, maybe I should just stick to the mid/high section out of the MD49 which is a combo waveguide that takes a single 8" cone woofer and a 1.5" compression driver.

Maybe build a 1/2" birch ply box to just house the mid/high section from the MD49's, and that up on a pole stuck in a Klipsch La Scala or Belle cabinet on each side? Would a short cabinet like that on the floor playing all the way up to 400hz or so suck with a crowd standing in front of it?

The Eminence Alpha-8MRA sealed-back 8" drivers are so cheap that it's tempting to get the dual 8" waveguides, but maybe there's not much point if it's unlikely that I'm going to put an appropriately-sized stack underneath them.
 
Decided to press on regardless and knock out most of the model for the sub cabinet today after work. All the parts is able to be scaled up/down from the assembly per desired front baffle size and enclosure volume, so I can easily switch it up to be vented instead. Got some penn-elcom hardware imported too which will come in handy for the other enclosures, and I'm getting in the swing of things (cabinetry is not my day job).

Once I finalize drawings, I have a buddy who can CNC route 18mm birch ply for me, which is sort of the impetus for this whole science experiment.
 

Attachments

  • Capture.PNG
    Capture.PNG
    92.5 KB · Views: 219
I suppose maximum SPL isn't as important to me as detail and tone, and the best low extension possible would be nice. I'm not going to be trying to outdo every other event in terms of volume, most parties around here are way too loud anyway. Basically everyone dancing is going to be easily within 10m of the front of the speakers, so maybe I don't need 130dB, you know?
The less a driver moves from rest, the lower the distortion, sso if you want it clean, you want max volume displacement and use as little of it as you need. My standard rig for smaller rooms was a pair of ported JBL18s, a mid box (DR or horn) and a largish 1" horn driven by a CP380M Beyma. My system was modular so I could just add more/different enclosures as required, but it always sounded better with a pair of 18s, running easier. I didn't even need to add another amp to the rack.

I had selected a sealed box for the sub due to the smoother rolloff, I had done some reading on here (albeit in the hifi section) about sealed subs being able to reach very low with limiting / EQ'ing, particularly in smaller rooms where room gain comes into effect. Most venues available to me are going to be relatively small rooms, so this seemed enticing. I had shyed away from tapped horns as it seemed like to get down to like 30hz with them the horn length just gets unreasonable, but maybe this isn't so.
Nope; you can always EQ the LF down, but turning it up beyond it's performance envelope sounds bad. And bad sound turns off punters. Go with a ported 18, better yet 2/side. If you tune to say 40Hz add a 2nd order HPF just below tuning to protect the drivers from overexcursion due to unexpected LF signals below tune. This is extremely important.

Maybe build a 1/2" birch ply box to just house the mid/high section from the MD49's, and that up on a pole stuck in a Klipsch La Scala or Belle cabinet on each side? Would a short cabinet like that on the floor playing all the way up to 400hz or so suck with a crowd standing in front of it?
Two points, both from experience.
1: A LS is going to get you about 125Hz in 2pi space; I know because I've owned a number and measured them. The Belle may go a but lower, but nowhere near the 40Hz LF limit that covers most music. If you don't need 30Hz for the music intended, don't go that low. It will often set off rattles and other problems you don't need. EVen with my low B (31Hz) tuned basses 2H is about 15dB above the fundamental and if you get the rest of the spectrum right, no one will notice the missing fundamental.
2: It generally sounds better over a wider spread if the MF and HF are higher and even angled down if necessary: think about the desired area coverage and the vertical dispersion of your HF horn and possibly your midrange horn. Bodies absorb sound, less in a smaller room/numbers, but why add work when you can get above them. They make negligible difference at LF.

The Eminence Alpha-8MRA sealed-back 8" drivers are so cheap that it's tempting to get the dual 8" waveguides, but maybe there's not much point if it's unlikely that I'm going to put an appropriately-sized stack underneath them.
Are they the specified driver for the WG? If not, then the performance may not come up to expectations. If using a given WG/horn, use the drivers specified for it.

I'm not personally keen on Eminence Alpha drivers as in my experience, they are low performance.

Last hard earned tips. Think about the form factor of your enclosures. You are going to have to load them in/out of some really crappy access at times so make sure they will bet through any doorway and tight turns/hallways etc. Assume the worst because you'll find it one day. Never let anyone ever touch your rig. Learn how to roll and store cables properly and carry some spares. Another unused amp in the rack can be a lifesaver, so you can patch around an unexpected fault. No piece of gear is perfect.
 
For amps, I was planning on something like the QSC RMX2450a bridged for the sub (1300w FTC @ 8ohms), then another for the mids (450w FTC @ 8ohms both channels driven), then maybe something like a samson servo for the compression drivers. Anything beyond the 2450a seems to become cost prohibitive, and at that point power consumption seems to become sketchy anyway (15A circuits are standard in the US, and may be all that are available in some venues).

If you've got limited power available, I'd be looking at class D amps for sure. They're considerably more efficient than the class AB/H RMX series, and lighter weight, too.

I suppose maximum SPL isn't as important to me as detail and tone, and the best low extension possible would be nice. I'm not going to be trying to outdo every other event in terms of volume, most parties around here are way too loud anyway. Basically everyone dancing is going to be easily within 10m of the front of the speakers, so maybe I don't need 130dB, you know?

I had selected a sealed box for the sub due to the smoother rolloff, I had done some reading on here (albeit in the hifi section) about sealed subs being able to reach very low with limiting / EQ'ing, particularly in smaller rooms where room gain comes into effect. Most venues available to me are going to be relatively small rooms, so this seemed enticing. I had shyed away from tapped horns as it seemed like to get down to like 30hz with them the horn length just gets unreasonable, but maybe this isn't so.

Take it from someone who has used the same drivers in ported and sealed boxes. It's just not a comparison.
Relying on room gain to get the LF response and output is a surefire way of building a system that works really well in maybe one or two venues. It's perfectly fine for HiFi people - they put the sub in a (hopefully optimised) location in their room, and then play music at fairly sensible volumes.

You might not want 130dB, but will your clients?
What do you do when you're at the red lights, and someone says "turn it up!"?

FWIW I also have a 1800W 21" driver in a sealed box. It looks louder than it is. I'm thinking of putting that in a ported box, too.

Chris
 
So, after going through JBL's current catalog and calling up a dealer for pricing/availability of certain parts like gaskets and phase plugs, for the tops I definitely want to basically duplicate the JBL MD49, so the PT-K95MH waveguide loaded with the 2169H from 270 up to 2.7khz and a 1.4" throat compression driver going up from that, possibly 2452H from the higher-end MD2.

After the very helpful feedback in this thread, I began looking into higher output/efficiency for the sub. I stumbled upon the PPSL design, and began doing some sims in winisd/hornresp. Switching from the extremely expensive 18XL1800 woofer to two woofers that cost half as much each, the 18RBX100, and doing a vented box with slightly higher tuning than I had been hoping for initially seems to yield interesting results.

A 20x10" slot just big enough to fit the drivers into gives a 1.9:1 compression ratio, and sinking the plenum deeper into the box and adding a horn mouth actually seems to add some efficiency. With the back corners of the slot filled to reduce the volume in the horn before the drivers, it looks like it could actually play up to that 270hz crossover points well.

These would certainly be massive things to move, taking up something like a 4'x4'x4' cube each (although making them narrower in one dimension and taller in another may be beneficial to handling), but if the subs can play from 33-270hz, that saves me from needing a separate set of midbass enclosures.

Does this seem reasonable, or is 33-270hz too much to expect from a pair of 18's? The hornresp sim makes it seem like after EQ'ing those humps down, you could get 130dB max SPL from 33-270, is that too ambitious?
 

Attachments

  • Screenshot at 2018-12-16 13-54-56.png
    Screenshot at 2018-12-16 13-54-56.png
    7.9 KB · Views: 178
  • Screenshot at 2018-12-16 13-54-41.png
    Screenshot at 2018-12-16 13-54-41.png
    8.8 KB · Views: 168
  • Screenshot at 2018-12-16 13-57-24.png
    Screenshot at 2018-12-16 13-57-24.png
    8.1 KB · Views: 171
  • Screenshot at 2018-12-16 13-58-54.png
    Screenshot at 2018-12-16 13-58-54.png
    10.4 KB · Views: 180
A >500L back chamber is a huge box. I'd argue that's gonna be a 4-man lift. Minimum. If there are any stairs, forget it.

For the size of the rooms you're looking at, there's genuinely no need for a 4-way system.
4-way systems always end up big and complicated, and very difficult to get right.
3-way systems are easier to get right, and would be aligned well (IMO) for the sort of size gigs you're looking to do.

The biggest 3-way system I'd design would be something like:
1.4" HF, 2x12" midbass, 4x 18" subs.
If I was needed to go bigger than that, I'd bite the bullet and go 4-way, taking design cues from Nexo Alpha.

Chris
 
These would certainly be massive things to move, taking up something like a 4'x4'x4' cube each (although making them narrower in one dimension and taller in another may be beneficial to handling)
With that size, you're not going to get them through most doorways. Even if you change the form factor it will still make them unwieldy and hard to stack.

Otherwise, I basically agree with Chris' last post.
 
Forgive me if I'm gravitating to strange/impractical designs, but I spent some time looking further into the vented rear chamber + PPSL plenum + conical horn idea, but downsized greatly.

If I'm sticking (for now) with the JBL mid/hi progressive transition waveguide (8" mid, 1.4" CD high, bottom crosses at 270Hz), and we've decided stereo subs up to 270hz is a dumb idea, then we're into a 4-way system and we need a pair of midbass enclosures that can play 80-270hz.

The B&C 12CL64 weighs just 4.2lbs per driver, and seems to sim well in a 2:1 compression PPSL plenum feeding a horn with an 80Hz expansion truncated heavily back to just 36" horn length. Back chamber is 90.5L with a tuning of 52Hz.

The mouth measures 17.2"x27.8" and I assumed a 17.2" x 6" rectangular port underneath, so the box wouldn't be so large at all. Total system volume is only ~240L which is ~2'x2'x2', so I need to set aside some time to figure out the CAD but I think if bend the horn and add an access panel to load the plenum from the side, total size of the enclosures would be manageable, particularly given the predicted output.
 

Attachments

  • 12cl64 ppsl 36in spl.png
    12cl64 ppsl 36in spl.png
    15 KB · Views: 90
  • 12cl64 ppsl 36in power.png
    12cl64 ppsl 36in power.png
    14.2 KB · Views: 92
  • 12cl64 ppsl 36in params.png
    12cl64 ppsl 36in params.png
    14.7 KB · Views: 88
total size of the enclosures would be manageable, particularly given the predicted output.

I still think you're on the wrong track, sorry.

2x10"s (Faital Pro 10FH520) in a ported box would keep up with that horn in output, in a smaller box, and make it up to 1kHz easily. 1.4" HF unit above, add subwoofers, and you're done.
I'm using those drivers (and a HF horn) in a cabinet that's around 800mm x 300mm x 250mm (HWD), and it's enough output for outdoor festivals. FWIW, they'd match well to 4x QSC KW181 per main speaker, which would be around 18dB louder than what you originally said was enough.

Going 4-way with three of those sections horn-loaded will be a serious undertaking, and IMO a serious departure from what you set off to achieve.

You asked for a sanity check...

Chris

PS - if you've already bought the JBL horns for some reason, try a direct-radiating 15" instead of a 12" in a complicated horn.
 
Last edited:
Chris,

Appreciate your patience with me. I'm having difficulty replicating your empirical experience through the conceptual/analytical simulation-based approach, which I think means I must be doing something very wrong.

I have attached an overlay of the maximum SPL chart of the proposed 2x12" horn-loaded cab with an equivalent max SPL chart of a 2x10" vented cab of 29L tuned to 60Hz.

Probably worth going into the assumptions made:

  • 2x12" horn loaded radiates into 2pi space (270Hz LPF ok to set on floor?)
  • 2x10" faital vented radiates into 4pi space (1000Hz LPF needs to raised high?)
  • Pmax in hornresp set to 1/2 the rated power handling (300w for the faital driver, 125w for the B&C 12"
  • Xmax in hornresp set to the manufacturer's rated Xmax (7.25 for faital, 4.5 for B&C)
 

Attachments

  • 12cl64 ppsl overlay w faital 10fh520.PNG
    12cl64 ppsl overlay w faital 10fh520.PNG
    155.3 KB · Views: 81
  • hornresp faital params.png
    hornresp faital params.png
    18.7 KB · Views: 93
You've got a 4ohm resistor in series with the cabinet, and the Faital Pro drivers are happy with 1200w peaks per driver. IIRC I used 90v RMS input levels on my simulation, although my current amplification will go past 100v.

Why are you setting the power levels so low?

I also tuned the 10"s at more like 80Hz for a bit more mechanical power handling, although when you've put a 100Hz 4th order crossover in there, it's largely moot.

Chris
 
Ah, that's what I was missing, Rg should be 0 unless I want to simulate voltage drop in the cables / amp.

I had wanted some sort of safety factor, but I suppose two hours of pink noise is probably the worst case imaginable and there's no need to pad that number further.

Remember that the pink noise rating goes as follows:

They usually use pink noise clipped to a 6dB crest factor. ie, peaks are 6dB hotter than the RMS level.

The RMS power rating is what gets put on the drivers. During that power test, they are being subject to brief peaks that are 6dB louder than the average level.

So a 100w driver can stand repeated 400w peaks, so long as mechanical limits etc aren't exceeded.
If the 400w peaks occur more frequently, the average power will rise past 100w, and that'll burn the driver out.

To make it clear, when they're running the pink noise test for the example speaker, they'll need an amplifier capable of at least 400w to make sure the amp itself doesn't clip before the speaker reaches full power.

The thing to remember, though, is that using a 400w amp would be pushing the speaker to the ragged edge - if mic feedback or similar high-duty-cycle (low crest factor) signals occur, the amp would be capable of destroying the speaker in short order.

I tend to go for 2x RMS ratings where practical, which splits the difference between the average and peak levels in the pink noise stress test.

Remember, doubling in power levels will only get you 3dB more.

Chris

PS - I simulated the 2x10" in half-space to match the 12" horn, so there'll be a 6dB discrepancy there.
 
Hopefully I'm not beating a dead horse here but, I tried to combine some of the different ideas I had explored previously while looking at various enclosure types.

Here is a simplified La Scala type cab, internal dimensions 77.14cm*60cm*47.68cm, 18mm thick ply. The sides of the doghouse in the plan view now form an irregular heptagon, which means that all of the flare happens in that view (fewer 'ramps'), except for one piece. Drivers are still those B&C 12CL64, and they'll load into the doghouse via an access panel on the top (or bottom, probably, I think I've shown the cab upside-down here).

Before adding bracing, the volume inside the doghouse is ~44L.

I realized Hornresp can sim these types of horns where the rear chamber vents into the horn mouth, and after playing around a little, it seems like this design functions better with the doghouse sealed.

This seems like a relatively compact and efficient kick bin, and I'm surprised at how flat it sims.
 

Attachments

  • ppsl_lascala_iso.PNG
    ppsl_lascala_iso.PNG
    93.8 KB · Views: 38
  • ppsl_lascala_params.PNG
    ppsl_lascala_params.PNG
    19.4 KB · Views: 42
  • ppsl_lascala_power.png
    ppsl_lascala_power.png
    13.9 KB · Views: 53
  • ppsl_lascala_spl.png
    ppsl_lascala_spl.png
    15.4 KB · Views: 41
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.