Two drivers loading the same horn.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi Tyimo,

Well, since nobody has answered, yet, and speaking strictly as a newbie, I'll hazard a wild guess. Most likely all of what I'm about to say is wrong, wrong, wrong.

In adding a second driver, presumably the chamber size would double.

The thoat, mouth, length and flare are all mathematically related to each other. If you know three, you can derive the fourth. So you can't just double the throat while keeping the rest the same. (But you could cut a new throat somwhere further up the horn, but that changes the length as well).

Having said all that, when calculating a horn, you tend to size the throat as a ratio to Sd, with the old rule of thumb being throat = .3 to .7 of Sd. Adding the second driver doubles Sd, so if you were designing from scratch, presumably you'd double the chamber, double the throat, keep the mouth the same, and change either the length or the flare (and then solve for either length or flare, which ever is unknown). Don't put any faith in what I'm writing, as I'm merely repeating what I recall reading.

I have learned that designing a BLH is like trying to hit a bullet with a bullet -- there are so many mathematically "valid" variations but finding one that will sound good will take months of an expert's time to design, and even then, the final result's sonic quality still has so many external variables (driver tweaks, amp, room etc.)

Tough stuff, way over my head! Tyimo, did you ever build a Replikon-style horn as you were planning?
 

GM

Member
Joined 2003
Greets!

It's true that to design the horn proper it doesn't use any driver specs per se, IOW all you need to design it is a low and high corner frequency (FL6, FH6) which defines its throat and mouth area with its expansion rate defining its axial length. Finding the right (FL6, FH6) for a given driver for a given app is where the math and to some extent the 'art' of horn design gets complicated, so while a driver's Sd isn't a design criteria per se, it becomes a big deal where compression ratios and/or desired HF BW and/or sheer size is concerned.

Bottom line, design a horn for one driver and add a second one results in the same kind of reduced performance that occurs when you do it in a typical sealed or vented alignment. Also, WRT horns, cram multiple drivers into the throat of a horn designed for an audibly different HF BW than they require and you can wind up with a poorly performing system and why so many mix n' match systems that litter the 'net are such poor performers that folks like me respond to queries of '......will XXX driver work in YYY horn?' with 'only if measured driver specs are reasonably close'.

GM
 
Hi GM, may I ask a question? If one doubled the CSA, the throat and mouth would double, while the length and flare stay the same. Is it correct to say that:

(a) the double-sized mouth would now be capable of going down another octave, but
(b) the design would be limited by the the horn's length (and thus wouldn't necessarily go down any lower)?

In other words, if the original Am was tuned to 50hz, and the length was set to a quarter wavelength of 50Hz, doubling the mouth would basically have no effect on how low the horn could go? You'd have to double the length as well? Sorry for the newbie nature of the question.

@Tyimo, did you ever build a Replikon-style horn?
 

GM

Member
Joined 2003
Greets!

?? Of course you can, I can always decline if I prefer not to.

Right, the only thing that changes is the bigger throat's cut-off is lower. If you double the axial length also, then output continues only a half octave lower due to doubling mouth area:

frequency = ~13560"/((SQRT(area/Pi))*2*Pi)

or

area = (((~13560"/frequency)/Pi)/2)^2*Pi

GM
 
oooh good some one who knows how to do the maths.

quick question, am i right in thinking that in a ML TQP, if i mount two of the same drivers parallel conection, verticaly along the length of the pipe, do i calculate the offset length as the mid point between the two drivers??

and what effect will it have if i increase the effective height of the driver by trying four drivers spread verticaly, with series parallel conection??

has anyone tried this and what effect does it have/ how does it sound.

cheers, steve.
 
Correctomundo!

The line just 'feels' an oblong driver with it acoustic center at whatever the offset is along its length and of course for every doubling of drivers you double the line's cross section area (CSA) to maintain the alignment of a single driver.

GM
 
yes both / or all four, inline down the front baffle.

was wondering if i could get the wide not high dispersion of a tl array, with the extended bass of a ml tqp (c. h.) ??

just thinking aloud realy, but i got use of an 8ft x 4ft cnc router at weekends and a few spare drivers.

cheers for the help, steve.
 
Well, the longer the driver line WRT the pipe's path-length, the further down it is acoustically, ergo the less average pipe loading and at some percentage of line length it starts to roll off the pipe's LF gain. At some point then, it's better to divide the line in two acoustically to keep gain high.

This is the way I did a proof-of-concept line array with (8) car audio 6.5" co-ax drivers, though ultimately I had to curve it to get the desired coherence, but then it had a 'head-in-vice' 'sweet spot', so didn't pursue it further.

GM
 
Tyimo,

it depends on how good your skills are at designing cross-overs is, although that is done for you with a plan or published design, in my oppinion, not worth the effort or expense. you can get as good if not better with single or multiple same drivers.


GM, the sweet spot isnt a problem, nor curving to maintain time alignment as the listening distance is between 28 and 45 foot!!

straight line length works to around 1878mm but i got it down to about 1280mm playing with the taper and open / closed end ratios and mass loading / tunning

with my 4 inch full range drivers that works to about 400mm so starting at 80mm down add 200mm (half the effective length) that makes 280mm down. 1280 divide by 280 equals 4.57 almost half way between a fifth and a quarter down the pipe, but as you say, the lowest will be 200mm lower.

cheers for the help, steve.
 
psychosteve said:
you can get as good if not better with single or multiple same drivers.

Tell that to people who like to hear uncompressed bass transients at ~'live' SPLs. You won't get far with an FR driver, short of perhaps Olson's 15in jobs from 50 years ago & a couple of others. Multiples would be better but that's still like trying to use a cork from a bottle of cabernet to fix the burst hull-plates of Titanic, plus, there's the inevitable issue of output combing if you're planning on running all FR on the front baffle. If you do, you're begging for trouble, as the HF will vanish & you'll get severe combing in the frequency domain through the midband -how low depending on how far apart the centres of the drivers furthest away from each other are. You can Eq the HF up, but that does nothing to sort out the midband issues. You can roll some off above their mass-corner, but then you'll run into sensitivity matching issues if there are more than two drivers. A curved array is about the only way around it, but GM's 'head in a vice' remark was not said in jest.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.