Modular Mid-High

Mark Wilkinson

Sophomore
Feb 24, 2015
243
13
18
Been working on a modular, easily array-able, mid-high setup, designed for ground stack use.

Two 12” mid drivers are in one box, and a BMS CD is in a second box that fits on top of the mid box.
The drivers are the same as in the PM90 / 60 …. rcf MB12N351 and BMS 4594HE.
The two boxes together look like a large JTR 3TX.
Basically, I simply took things I’ve learned from studying Peter’s DIYs and Jeff’s 3TX, and put it into a form to meet some specific goals.
Goals being: getting a ground stacked horn as high as possible, separating the horn from the mid to allow easy swapping of horns and compression drivers, easy arraying, ...and keeping anything I have to lift up high below 50 lbs.

The boxes are 40 degrees (as like pie wedges), which came from best splay angle I was able to measure for the xti-1464 .
The size of the high boxes was designed to barely hold the xti-1464 with the 4594HE. The other horns and CD’s I have are all smaller, and can easily be interchanged.
A pict with different horns is below.
horns R 500.jpg

The mid box matches the dimensions of the high box, other than height of course.
Put together, its 47”H x 16.5” W front x 14.5” D (119 x 42 x 37) 6” wide at rear
Mid box w/o grill 47lb. High box 23lb.

The high box has front rubber feet that nest into routings on the mid box, along with a rear jack foot on the high box to vary downward tilt..

On top of my labsubs or orbitshifters, the horn ends up about 7 ½ feet high at center.
I can gain 2 more feet of height pretty easily, by adding some diy 18” subs into the middle of stack.
(The diy subs sum almost perfectly with the hornloaded subs.)

A pict of 3 of the modular stacks is below. I figure first stack (with xti1464) is 60 degrees, and each incremental stack adds 40 degrees.
3 stack R 500.jpg
So I can go 60, 100, 140 degrees H pretty easy.

I haven’t yet measured multiple stacks outside. So far, indoors is looking like really smooth H coverage. And it's crazy powerful....I mean it's almost like 3 DIY60s arrayed....I measured efficiency, modular mid box vs the DIY60 mid, and was pleasantly surprised to find the modular was down only 0.3dB....100-650Hz.
I know there’s no way the vertical coverage/directivity can match the DIY 90 /60, but that’s a tradeoff I decided I could live with…..
Can't wait to test/listen outdoors.

Anyway, I realize this is a project that’s unlikely to generate a lot of interest, given its unusual nature…but I wanted to post it on this forum now that the new bulletin is running, and kinda say thanks again, to Peter, Don, the forum itself, everyone... for the ongoing education that’s let me play around and try something like this…
Best, mark
 

Attachments

  • horns R.jpg
    horns R.jpg
    302.6 KB · Views: 56
Hi Mark,

That’s a great solution for a simple ground stack rig. I have been looking at something similar using some Turbosound dendritic horns which were designed to array a bit like L-Accoustic’s Arc’s.

Having the drivers recessed like you have will give you some gain. I would however be amazed if the efficiency of the DIY and your box are that similar. On the SIM the DIY has about 5 dB more … however after you apply the EQ to flatten things they should appear similar in terms of the input signal to the processor and measured output from the speaker (assuming the same crossover gain).

The trick with the DIY is that the EQ where you have cut some of the peaks out is the extra drive you can give the speaker before you exceed its actual power handeling.
 
Hi Mark,
Lovely!
Your gain per driver should indeed be astonishing in a tight pack configuration like yours. How you have come to observe Peter Morris mid high-SPL figures I'll leave for you and Peter to investigate.
I'm curious as to your nice selection of horns and which ones those top two ones are. XR1464? If so, how do you like them compared to the XT1464?
Best regards, Fred
 
Hi Mark,

That’s a great solution for a simple ground stack rig. I have been looking at something similar using some Turbosound dendritic horns which were designed to array a bit like L-Accoustic’s Arc’s.

Having the drivers recessed like you have will give you some gain. I would however be amazed if the efficiency of the DIY and your box are that similar. On the SIM the DIY has about 5 dB more … however after you apply the EQ to flatten things they should appear similar in terms of the input signal to the processor and measured output from the speaker (assuming the same crossover gain).

The trick with the DIY is that the EQ where you have cut some of the peaks out is the extra drive you can give the speaker before you exceed its actual power handeling.

Hi Peter, thanks for the kind words.

Regarding efficiency, .....yes, as you surmised, I was using the term 'efficiency' in a post crossover, post EQ, context.
I should have stated that, sorry.

I use the same crossovers and individually flatten response for each box. It's pretty easy with the FIR to get measured traces that look near identical. Then I run pink noise through that processing, and take 1-2 min SPL LEQ's, along with simultaneous RMS voltage averages at the speakers.

A little math and I feel I have a valid real world comparison that reflects intended usage.....kinda a measured "average net efficiency"
Does this hold water? See any pitfalls? It's really helped me compare subs...this is the first time I tried it higher up in freq.

Major edit: I just realized in trying to retest efficiency outdoors, that the update to smaart has changed the way LEQ triggers.
Will post new meas when confident....I apologize

OK, chasing my tail... although SPL choices/toggles in smaart have changed, it seems I was lucky when i made prior indoor "efficiency" meas.
Today's new outdoor measurements agree with prior ones.....
 
Last edited:
Hi Mark,
Lovely!
Your gain per driver should indeed be astonishing in a tight pack configuration like yours. How you have come to observe Peter Morris mid high-SPL figures I'll leave for you and Peter to investigate.
I'm curious as to your nice selection of horns and which ones those top two ones are. XR1464? If so, how do you like them compared to the XT1464?
Best regards, Fred

Hi Fred, thank you.

The top two horns are from a JTR 3TX, one is 60 degree and the other 90. I don't know who makes them.

I wanted to be able to directly compare them against the rcf950 and xti-1464.
Also wanted to compare bms 4594HE against the 4593 used in the 3TX....
I haven't had time to make fair comparisons yet.

That's been a big part of this project...how can I play with and learn from stuff I already own, with no more cost than making sawdust :)
 
Last edited:
  • Like
Reactions: Jason Joseph
Hi Mark,
Absolutely inspiring bit of experimentation! Can't wait to hear how these perform in the wild.
BTW Your wife must be incredibly understanding of your "hobby" ;-)
Looking forward to your test results.

Hi Don, many thanks!
It really came about when I started to make another pair of PM60's to hang from stands like you were showing me what you had done.
But as I kept thinking through the gigs where I wanted two per side, I couldn't think of any where a stand would really work that well....as compared to just stacking on subs. So I said, what the heck....sawdust time...

I have a beautiful ex-wife that I still love dearly...especially when I get to play with gear all over the living room now !!!

As soon as I get a good set of off-axis I'll post them.
 
Hi Peter, thanks for the kind words.

Regarding efficiency, .....yes, as you surmised, I was using the term 'efficiency' in a post crossover, post EQ, context.
I should have stated that, sorry.

I use the same crossovers and individually flatten response for each box. It's pretty easy with the FIR to get measured traces that look near identical. Then I run pink noise through that processing, and take 1-2 min SPL LEQ's, along with simultaneous RMS voltage averages at the speakers.

A little math and I feel I have a valid real world comparison that reflects intended usage.....kinda a measured "average net efficiency"
Does this hold water? See any pitfalls? It's really helped me compare subs...this is the first time I tried it higher up in freq.

Major edit: I just realized in trying to retest efficiency outdoors, that the update to smaart has changed the way LEQ triggers.
Will post new meas when confident....I apologize

OK, chasing my tail... although SPL choices/toggles in smaart have changed, it seems I was lucky when i made prior indoor "efficiency" meas.
Today's new outdoor measurements agree with prior ones.....
Hi Peter, thanks for the kind words.

Regarding efficiency, .....yes, as you surmised, I was using the term 'efficiency' in a post crossover, post EQ, context.
I should have stated that, sorry.

I use the same crossovers and individually flatten response for each box. It's pretty easy with the FIR to get measured traces that look near identical. Then I run pink noise through that processing, and take 1-2 min SPL LEQ's, along with simultaneous RMS voltage averages at the speakers.

A little math and I feel I have a valid real world comparison that reflects intended usage.....kinda a measured "average net efficiency"
Does this hold water? See any pitfalls? It's really helped me compare subs...this is the first time I tried it higher up in freq.

Major edit: I just realized in trying to retest efficiency outdoors, that the update to smaart has changed the way LEQ triggers.
Will post new meas when confident....I apologize

OK, chasing my tail... although SPL choices/toggles in smaart have changed, it seems I was lucky when i made prior indoor "efficiency" meas.
Today's new outdoor measurements agree with prior ones.....

Here are some rough SIM's for comparison - These are power response plots not SPL, they do not allow for the increase in efficient that comes directivity. I have also shown what happens when you apply EQ ... it looks like you have lost efficiency but you haven't, the box is still capable of producing the maximum SPL based on the non- EQ response.
 

Attachments

  • doble 12 compared.jpg
    doble 12 compared.jpg
    141.6 KB · Views: 72
Here are some rough SIM's for comparison - These are power response plots not SPL, they do not allow for the increase in efficient that comes directivity. I have also shown what happens when you apply EQ ... it looks like you have lost efficiency but you haven't, the box is still capable of producing the maximum SPL based on the non- EQ response.

Thanks Peter, I'm OK with power response vs SPL, but where I struggle is with an idea of "net usable efficiency" for lack of a better term.

And I get that a box is still capable of producing max SPL based on the non-EQ response (much as a power amp can still produce max power even as gain is turned down).

But it seems the only way the max power could be realized in a freq range where I need to cut to flatten response, is most likely to have other frequency's running out of gas, particularly where there are non-EQ response dips.
So if I want to maintain level response throughout, don't I need flattening in place?
IOW, I don't get the value in response peaks in considering "efficiency"? Likewise, I also think "efficiency" should be knocked down for any EQ needed to bump up dips.....

Anyway, don't mean to belabor my own views...just trying to see if I'm making a mistake in my reasoning.

Just because it's so easy and I take so damn many measurements.....
I ran the SPL LEQ / average voltage test on both boxes on the floor this morning just using x-overs, no EQ. LR8 @100, LR8@650.
The PM90 measured +0.8 dB like this (vs +0.3 with EQ flattening)

I wonder if my new box could be getting any gain from behind the drivers....it's not really ported, it's more like a transmission line that gets progressively deeper from top to port. I dunno what to call it really...it's a effort of trial and error..built via lot's of measurements cause I haven't yet learned hornResp.....
Any thoughts? Thanks as always !
2 x 351 empty box R.jpg
 
Hi Mark,

I modelled your box as a reflex enclosure with a very short horn (3") ... The PM90/60 is horn loaded above about 150Hz and more or less a reflex loaded below that so that it had useful response to 100Hz.

If your box is a transmission line with a short horn in front you indeed may get some more efficiency. My best attempt to model this seem to indicate an efficiency just below the PM 90/60.

Its been a long since I have heard a TL speaker but I can remember just loving the way they sounded.

This video may be of interest
 
Thanks a bunch Peter, for continuing to delve into my box and for trying to model it.
I've simply got to stop and put some real effort into learning HornResp I think...

It does sound very nice, either singly or arrayed...better than I hoped really. Tonality and transients are excellent.
Doesn't image as quiet as well as your boxes up close, given the greater vertical distancing, but still quite nice.

While i was working out the dual mb351's, I also tried mb405's.
They are waay too heavy to put two into a box, as you said sometime back I think, when talking about your DIY driver choices.
The single box I made below, to go with the modular concept, weighs 40 lbs, only 9 less than the dual 351's ! It's a shame it's so heavy, it really is a great sounding driver.

Thanks for that link...it's awesome so many innovative designs are showing up...wow, a back-loaded transmission-line cardioid-midbass !

405 r500.jpg
 
The weather let me test outside a little today, so I started with the biggest concern,
......off axis vertical. Started with a single stack.

First plot is on-axis, then down 10, 20, and 30 degrees.
Second is on-axis, but with up 10, and 20. Can't see ever needing up 30...
Showing two plots cause all together it's too messy to see anything IMO. 1/3 oct trying to get big picture first

Trying to sort out some junk around mid to high x-over at 650Hz, but measurements are at only 9ft which may be too close not to have junk.
Dunno...step at a time...but happy so far..

Down axis
mod vertical down axis.JPG

Up axis
mod vertical up axis.JPG
 
Last edited:
  • Like
Reactions: Klaus Zimmermann
That actually looks great for a non-symmetrical double 12" + ,horn. There is no perfect solution just the best compromise. The trick is to get the two 12" drivers to play together without phase cancellation (vertically between the two 12's) and then integrate the HF horn without major off axis issues at the crossover point.

Have a look at these two boxes below ... note the improvement in the newer box's vertical directivity with the lower HF crossover frequency.

If you go too low with the crossover point you will get the horn's pattern control collapsing and becoming too wide and you will also put too much strain on the HF driver ...

I suspect with 2 or 3 boxes arrayed you may (don't know just a guess) run out of HF power. If you could raise the crossover a bit, say 700Hz it may still work ???? and you will have a little more HF horse power.

http://www.rcf.it/c/document_library/get_file?p_l_id=249233&folderId=22490&name=DLFE-3110.pdf

http://www.rcf.it/c/document_library/get_file?p_l_id=16359553&folderId=22490&name=DLFE-29707.pdf
 
Thanks so much for the ongoing coaching, Peter.

I see how the newer TT-22a's lower HF crossover frequency improved directivity....like in your DIYs....
Which is one reason I've been keeping close to crossover at around 650Hz crossover as in your boxes..trying to take advantage of the horn directivity as far down as possible.

Yes, I'm kinda expecting to probably run out of HF when arraying 2-3, at least down to where the horns are holding directivity.
And I am a little concerned with overdriving the HF, particularly given the crossover type I'm using.

I've been wanting to ask you, given the steep knee HP I'm using, what you think the safe freq should be for the 4594.

I realize the steep knee puts more loading on the driver. Does 7000Hz still work for more power, or higher?
The steep knee is pictured below vs a LR 24.

I'm using such a steep xover to minimize the freq summation range between mid and HF, figuring that helps minimize off-axis variations.
It seems to help alot, especially since my box doesn't enjoy the much better acoustical focus of the PM 90/60's with their dipole mids surrounding the horn.

xover ellip mid to high.JPG


As always, please feel free to tell me where I'm screwing up !!
 
Last edited:
Been working on a modular, easily array-able, mid-high setup, designed for ground stack use.

Two 12” mid drivers are in one box, and a BMS CD is in a second box that fits on top of the mid box.
The drivers are the same as in the PM90 / 60 …. rcf MB12N351 and BMS 4594HE.
The two boxes together look like a large JTR 3TX.
Basically, I simply took things I’ve learned from studying Peter’s DIYs and Jeff’s 3TX, and put it into a form to meet some specific goals.
Goals being: getting a ground stacked horn as high as possible, separating the horn from the mid to allow easy swapping of horns and compression drivers, easy arraying, ...and keeping anything I have to lift up high below 50 lbs.

The boxes are 40 degrees (as like pie wedges), which came from best splay angle I was able to measure for the xti-1464 .
The size of the high boxes was designed to barely hold the xti-1464 with the 4594HE. The other horns and CD’s I have are all smaller, and can easily be interchanged.
A pict with different horns is below.
View attachment 208565

The mid box matches the dimensions of the high box, other than height of course.
Put together, its 47”H x 16.5” W front x 14.5” D (119 x 42 x 37) 6” wide at rear
Mid box w/o grill 47lb. High box 23lb.

The high box has front rubber feet that nest into routings on the mid box, along with a rear jack foot on the high box to vary downward tilt..

On top of my labsubs or orbitshifters, the horn ends up about 7 ½ feet high at center.
I can gain 2 more feet of height pretty easily, by adding some diy 18” subs into the middle of stack.
(The diy subs sum almost perfectly with the hornloaded subs.)

A pict of 3 of the modular stacks is below. I figure first stack (with xti1464) is 60 degrees, and each incremental stack adds 40 degrees.
View attachment 208564
So I can go 60, 100, 140 degrees H pretty easy.

I haven’t yet measured multiple stacks outside. So far, indoors is looking like really smooth H coverage. And it's crazy powerful....I mean it's almost like 3 DIY60s arrayed....I measured efficiency, modular mid box vs the DIY60 mid, and was pleasantly surprised to find the modular was down only 0.3dB....100-650Hz.
I know there’s no way the vertical coverage/directivity can match the DIY 90 /60, but that’s a tradeoff I decided I could live with…..
Can't wait to test/listen outdoors.

Anyway, I realize this is a project that’s unlikely to generate a lot of interest, given its unusual nature…but I wanted to post it on this forum now that the new bulletin is running, and kinda say thanks again, to Peter, Don, the forum itself, everyone... for the ongoing education that’s let me play around and try something like this…
Best, mark

Did you mention what kind of 12's you're using? - if so sorry I missed it. I've been doing a similar ground stack for a while and I like it a lot. I'll try to post a driveway pic from FB.


https://www.facebook.com/photo.php?fbid=3217687595513&l=0c425ca4f3
 
Wow Gene !
Those are 12"s ? With a 2" CD ? What drivers? Who made them?

How tall, and deep are your boxes? They look like they're over 5ft easy.
How do you use them? Stacked on subs? If so, how do you keep them in place?
Do you use different settings for 1, 2, or 3 together?
I've already found from just a little indoor listening/measuring that I will need to have different presets for multiples....
Sorry for all the ??

Oh yeah, my 12"s are the rcf mb12n351, sorry that wasn't clearer....