|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
50.184.121.207
In Reply to: RE: More horn ruminations: Tractrix? Or La Cleac'h? posted by Presto on February 09, 2015 at 13:38:58
I'd go with the Le Cleac'h. If you have a horn midrange and super tweeter above it, you're likely already screwed in regards to nice vertical polar response. Making the midrange horn a bit bigger probably won't make much difference. I'm working on a system right now using a LC midrange horn. I'm crossing a 380Hz horn, 1.5" throat and the full rollback to a super tweeter at ~9k that's above the horn. The vertical polar map has all sorts of crap in it around 9k because of the large spacing, but I don't notice it when I'm listening. The vertical -6dB angle is actually pretty consistent from the midrange to the super tweeter. It just gets choppy around crossover. If I go from sitting to standing, it sounds fine though. I'm still working on the crossover, but it looks like if I tweak it a bit and go with a very non-textbook alignment, I can get the on-axis, listening window response and power response to be pretty decent - that is, without a big dip at the crossover. Same for my midbass crossover which is at 800hz and is probably the more important one. I still get a bit of a bump in the DI there, but it's not as bad as if I just do a typical Linkwitz-Riley acoustic crossover. I'm finding that good measurements and good simulation capabilities are really required to sort all this out.
Follow Ups:
Thanks John.
Yes, I worry about edge diffraction off of a thin-walled petal-horn tractrix. One goes through all that trouble for a tractrix curve and then the mouth truncates suddenly, or worse, some arbitrary curve shape or round-over is used. I believe the tractrix terminates at 90 degrees, but whether the curve extending past that point is a mirror image of the curve shape or some other mathematical function, I will need to calculate - the calculators I have only go to 90 degrees, some not even that far.
I appreciate the input on the "far away" super tweeter. I am going to try this with the JBL 2380A / 2441 horns I have and a 2402 or 2405H super-tweeter... or any tweeter properly level matched.
Sometimes I wonder about using a co-axial horn-mounted driver... hmmm... horn within a horn? Probably a bad idea. Or some suggest to use titanium diaphragms to get the most upper end response out of the 2441 and EQ it up, while others don't suggest pushing 4" diaphragm 2" throat compression drivers beyond their upper frequency roll-off point.
What I really really want to try is those 2441's doing midrange say 500 to 5K powered by my 300B SET monoblocks! I can only imagine how good vocals could sound through a properly done horn!! :)
Thanks for your input and food for thought.
Oh... any preference of yours for supertweeter? I have JBL 2402 and 2404 and want to try and convert a pair of 2402 into 2405 with the requisite parts. (I think only two parts would be needed).
Cheers,
Presto
On my horn (1.5" throat JBL 2435 - beryllium diaphragm), I find the polar pattern goes screwy above 9k (c/d) and the horn / driver starts to store energy, so on your 1.9" throat, I'd try to stay below 7k as a start. So your idea sounds good to me.
I use a JBL 2402 tweeter on a custom Le Cleac'h horn / phase plug I designed. It gets me the DI I want to match up to my midrange horn plus low diffraction. Response is crap (some very sharp notches), but I don't find that to be a problem once the average response is EQ'd to my target. That's just in the design of the driver as far as I can tell - haven't investigated it too much more. I know one notch is a circumferential mode in the plug, but there are others that don't go away when I fix that. I just know my FEA model of the original driver had the same response issues, but the DI was way worse. I'm not sure it matters at those frequencies though - the notches are very narrow. I currently have it crossed at 2nd LR at 9k acoustic, but some modeling is showing that it might be better crossing it 1st order acoustic at 9k with another 2nd order highpass at 2.5k. I measured it as being good for distortion at the levels needed to do 115dB/1m in the passband with this acoustic roll off, so it should be okay with this for my purposes. For this I would roll my mid off at 6k 2nd order butterworth acoustic, then another 1st order butterworth acoustic at 15k. Then playing with the delay gets it working right - it needs to be within about a 7mm window to not completely screw up the response at crossover. The actual acoustic crossover is about 7k, but it's at around -5dB - kind of an oddball. This gets me on axis, listening window, and power response without a dip at crossover. The downside is that there is kind of a funky interference pattern vertically, but I think it's dense enough that it's not perceptually audible. I found you can do a crossover without this interference, but it has other downsides that I thought were bigger risks. Still have to try listening to this though.
I would avoid the slot horn - never played with it, but the polar is going to be very wide in one direction which won't match to anything you're likely to cross it to. This will color the in room response.
Every time I've looked at a coax-mounted tweeter (not in the driver itself, but just hung in the horn mouth), it hasn't looked good.
On the rollback, a long time ago I did some simulations and found you don't really need the full rollback to get most of the benefits. You do want it to go beyond 90 degrees to on axis though. You could just add a large radius to the end of a tractrix that looks about like a Le Cleac'h and it would probably be fine, but I've always liked the sound and simulated polar responses of the Le Cleac'h horns better than tractrix.
Right on John!
Thanks for all that fantastic info. Yeah, I don't think I would want to just add a terminating radius to a tractrix... if I'm going to wrap around I'd want to stick with the La Cleac'h formula. Thing is, a guy can build a set of tractrix petal horns. La Cleac'h petal horns?! Don't think so!
So I took apart the 2441 drivers last night for inspection and cleanup prior to starting the project. Terrible.
One is original 2441 aluminum diaphragm with good throat, good concentric pole piece and top plate (gap looks good) and the bakelite throat is not cracked and there is no scraping or signs of contact on the coil former or the windings themselves. Yay! The second one. Sigh. (a) Cracked bake-lite throat which is a symptom of (b) shifted magnet/pole piece assembly which was easily visible - the gap was not concentric at all. This explains the 2445J diaphragm in there, put in when the previous owner tried to fix it himself (without repairing the shifted magnet!!) This resulted in rubbing of the voice coil that resulted in shorted winding and a DCR of 5.1 ohms instead of 8.5 or so.
The throat is actually repairable with some epoxy and careful sanding... or I can have some machined like those guys did in Japan... make from brass. Tempting. The internal taper from 1.5" to just under 2" would not be machining for the faint of heart... In any case... Now, I plan to build a two-part jig that allows me to center the magnet, the pole piece, and the top plate all relative to each-other AND the cast front housing *all at once*) - which will also have provision to control the speed of the magnet engaging as to not slam it to the housing, it being AlNiCo. What I would do to be able to do this work in a JBL shop with the right jigs and tools! ;)
Might be worth making a youtube video. I am quite confident that a well planned jig can get this to work...
Wish me luck. Failing this, I will probably consider other compression drivers than the 2441, ones with latest technology and parts more available.
If I can't repair the bad 2441 driver in a cost effective way and get a reasonably matched pair, I won't build custom horns around that particular driver and cutoff and throat dimensions...
Have a good weekend!!
Cheers,
Presto
> Thing is, a guy can build a set of tractrix petal horns. La Cleac'h petal horns?! Don't think so!
Sure you can. Why not?
Btw, it is 'Le Cleac'h', not 'La Cleac'h'.
John:The projects that I was seeing online for Tractrix petal horns were usually not "wrapped around", and most I saw were terminating at 90 degrees at best. I don't think I want to turn horns, and I am tempted to spend the time working OT and just save for horns from a company that is set up to custom fabricate. (I have a bad habit of putting 100's of hours into DIY projects, when sometimes I probably would just prefer to get the end result! For sure, sometimes I spend more doing it myself when I count tools and finishing materials, so I have to be sure it's going to bring more enjoyment than frustration.)
I believe I have a graph here of a tractrix equation that is not truncated but a complete solution set, save for the throat opening. The calculators I have for tractrix stop at 90 degrees, and I would not want to go through the trouble of a tractrix only to add an arbitrary radius at the end.
Turns out I have one "good" 2441 with original aluminum and one which now has a repaired throat and need to be reassembled. I was overestimating the difficulty. The thing practically centers itself on reassembly - except I am not sure if they intended for there to be a space between the pole piece and the shorting ring of the top plate. I will have to look into that further.
Cheers,
Presto
Edits: 02/17/15
Presto JMLC did an excel spreedsheet for petal horn calculation . I have it in French. If you want it send me your email. Or maybe you can search the WEB it is called pavillon_JMLC.xls
Chris
...I checked my "horn calculator collection" (yikes, I have one of those... lol).
And there it was!
Thanks for pointing me (back) in the right direction! :)
Cheers,
Presto
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: