Audio Asylum Thread Printer
Get a view of an entire thread on one page
|For Sale Ads|
In Reply to: RE: Exit angle mismatch posted by Mr_Steady on February 02, 2017 at 16:11:54
Thanks for your answer, I Think I have a better understanding now.
Actually, I hade two things on my mind.
1. The eliminiation of the cavity-error caused by an 10 degree 2" driver exit meeting and a 4 degree 2" horn throat.
2. Since I am doing an extension of the horn to solve the first problem, Is there reason to extend the horn even further?
No, The Radian 750Pb is a 2" exit driver and the horn has a 2" throat.
But the driver exit-angle is 10.4 degrees and the horn-throat-ending is about 4 degrees. This make a smal (but not negligable) cavity inside the driver exit (an anomaly in the horn curvature).
By removing the dust protection cloth at the driver, I could extend the horn going into the driver, all the way down to the air-vents, which in my case is approx. 1,4" in diameter. This means that the problem with driver-exit-angle and horn-throat angle will dissapear.
Since, the preasure alteration in the horn throat is high, this extension must be mechanically firm and sturdy. Still it needs to be quite thin walled to go inside the driver exit bottom. The last inch lip will be coneshaped with a 10,4 degree outside angle, and a 3-4 degree inside angle. I have made some sloppy experiments using carbon fibre/epoxi for a bent extension, which would meet this requirements (se pix).
By making a 107mm extension of the 200Hz Tractrix horn curvature, the diameter goes from 2" to 1,4" (everything else the same). The horn is still a 200Hz cut off horn and the frequency bandweidth is unchanged.
Now, to clear my question (which I Think you have partly answerd).
Whould it of any benefit to make this extension even longer (than 107mm), going to a even flatter curvature than the 4 degree exit discussed above. Or is the tractrix curvature opitimized as it is?
Note, the pix showes one of my experiment with a bent extension pice which is 2" in one end, and 1,4" in the other. In this case, the length is extended from 107mm to 200mm (so itīs curvature does not follow the tractrix curvature by the book). But as I said, this was more of an experimnent using carbon fibre, than evaluating sonic gains.
By making the extension piece bent the extension could be extended arbitrarily ceeping the WAF-factor low ;o)
Finally, I am not sure if I should make this straight or bent when it came to the real extension. I just wanted to show this piece as an alternative.
Finally, I understand that the WE555 is a unique driver. And that I will never come close to its properites with modern PA-drivers. Still the questions must be asked if there is something to learn from those days in modern designs.
It sounds like you've been reading Geddes. It's been awhile since I've read his book, and his posts on diyAudio, but IIRC Geddes was concerned about matching the exit angle of the driver and his OSWG, but this had significate importance to him because of the characteristics and the theory of the OSWG itself. It didn't seem to be as critical to him on other profiles, although he would certainly recommend it even for the tractrix profile. For him it was critical for the OSWG.
I remember a thread where he and I guy I think who was in Sweden conversed about how the man had made his own OSWG, and had taken a metal band saw to his driver and cut off everything in front of the phase plug, which you called air vents. It's what you are talking about doing. Geddes approved of this very much, but once again this is critical to him because this was an OSWG. I would point out that the throat of driver, that is from the phase plug to exit (mouth) of the driver is a horn, so in effect the dust cover is the horn extension from 1.4" to 2" if you follow me. Yes it's probably a different expansion rate that the tractrix flare, but no biggie. What you really want is horn that has a throat entry 10.4 degrees. The HornsPl. company designed some of it's fiberglass horns from the outset with a certain exit angle for a specific driver. Check them out on the Autotech site and diyAudio. Or you could build a new paper horn for you specific driver. :)
If your are really interested about the exit angle matching, then just go to diyAudio on the "Geddes on Waveguides" thread and ask him yourself. I have an OSWG that is 12" in diameter and is good down to 1200hz. I can tell you it has almost no throat. It is almost an entry, then a fast flare with a mouth.
When thinking about this and your driver and xover point I keep thinking about that ancient saying, "You strain out a gnat, but you swallow a camel." I don't mean this in a bad way, but I think you might be breaking bigger horn rules than Geddes connection rule.
Everybody has their own theories about horns, and audio, so I will tell you mine since it seems to be different from most commercial horn systems I see. It is; I follow the rules. The rules of acoustic horns. I learn the natural rules governing horn as best I can, the I stick within them. Have I bent and broken them before? You bet. But I found that if I stick within the rules I get better sound. Simple as that. Certainly doesn't make me smart. All the guys who design and sell home horn speakers know way more about horns that I do, but they can never resist the temptation to cheat, because they have to make them just a tad more commercially acceptable. A little smaller, a two way. IMHO I modified a Klipsch Heritage fully horn loaded speaker, and made it sound better than Paul Klipsch did. Still PWK was a horn genius, and I am not, if you follow me.
Linked below is a recent testing review by John Atkinson of a German horn loudspeaker. He astutely caught that the mid horn was being run outside of it's proper operating range. $50,000.00 speakers. See, they can't resist bending or breaking the rules, not even for $50K speakers.
About your bent tube. Remember, when the size of the soundwave is smaller that the neck of your horn, the horn ceases to exist. Paul made a good point about what happens in the bend when the soundwave is contacting the sides of the horn, but when the soundwave is smaller that the tube the soundwaves just start bouncing around inside the neck. This causes internal reflections, High Order Modes as Geddes would say. This creates distortion. A 10Khz wave is smaller that 1.4". In your case that means all the signal above 10Khz is going to become distorted.
Last thing. I will share my thoughts with you on the highpass for a midhorn. I once ran a 350hz tractrix horn with an EV K55 driver using a first order slope. It went into gross distortion constantly. In fact sometimes the diaphragm unloaded totally, and you could hear the diaphragm flopping and buzzing around, probably banging against the phase plug. Not good. It's not the K55's fault. It can go very low, because it uses a phenolic (polymer) diaphragm. Your Radian uses titanium or beryllium? Neither are made to go low.
I take a different view about all this than is probably commonly accepted. Paul talked about letting the horn's theoretical cufoff be part of your acoustic crossover. PWK did that. I don't agree, and I try not to do this. Several reasons.
If your driver is putting out signal close or below the Fc then the horn can't pass it, and the soundwave collapses at the mouth of the horn, then falls back down the horn back to the throat, and becomes standing waves in the throat. This creates distortion, and helps to overload the throat, which you may be overloading anyway by sending a lot of high end signal out as well. Like using it as a two way, trying to send one and a half decades of bandwidth through one horn. At least this is how I understand it.
Then on top of that you are sending substantial electrical signal to the driver below the horn's Fc. So the diaphragm is operating unloaded, which probably means it is vibrating violently, which is not conducive to good sound.
Myself I like to have the signal attenuated down -12 to -15db at the horn's Fc. Less distortion. Better sound. YMMV.
Now if you get lucky, and the driver just happens to roll off where you need it to, then you can always incorporated that into your acoustic crossover.
Whew! Man, I ain't even proofreading this one.
Big speakers and little amps blew my mind!
Thank you very much for an exhaustive and personally experienced reply. After all, there are no "truth" in this. Just different point of views, and a choice on whom to learn from.
First, to answer your question. No the Radian 750Pb has an aluminum -membrane.
Second, what you write about "You strain out a gnat, but you swallow a camel" is really what I needed to hear to get this horn extension ide a rest. I often get stuck in one specific detail, when I should discuss the whole system. I will later write down my complete system asking for advise where my weakest points are, and initially focus on them.
Btw, calculating with the horn cutoff (As PK did) as an acoustic filter sounds really like jeopardizing with the devil. A flapping membrane will surely put hard strain on the membrane surrounding even if not hitting into the driver dome. I prefer to handle this electrically with the filter. As I mentioned, I had settled on a 1īst order, 350Hz HP-filter for the horns. But surely there are more room for experiment here. Looking around, most people with similar setup uses higher order filters, which make me wonder if I turned back to my initially solution to easy. I have previously tried higher order passive, and active filters and also gone digital by using a miniDSP device. But after listening for months I always return to the simple solutions.
My experience of commercially driven gear is (as you described) that they are often far from optimized for their task, and that's why I seldom visit highend shows or hifi stores anymore. I prefer to learn and do it by myself. Ones avail I ask myself what is the goal of this hobby. For me itīs the craftsmanship, finding new ways to make things and understanding how (why) they work, more than the music itself. The feeling of turning (often cheap scrap) into good locking, well working gear gives me a good feeling.
You're welcome pix. I was just trying to point out some pitfalls that you have to remember are there. I think the truth lies in the fact there are a lot of right answers, because everybody has a slightly different need. For instance I don't know how loud you like or are able to play because of your living arrangement and neighbors. You might listen at a rather low volume, so your speakers are less stressed, and distortion or compression isn't so much a problem. You can get away with more. I was writing from my view point, which is I like music, and I sometimes like it loud. Usually in fact. I can get away with playing it loud, so I need my speakers to sound good at loud levels with low distortion. That's harder to do. I have to be a little more conservative. I am happy to say that I think I have a pair of speakers that play as loud as any home speaker you could wish, and it has incredibly low distortion. I really love the way they sound, and with only three and half watts! I think if any of the bad things I described before were happening with your speakers you would hear it.
Since you like the technical side and you have the miniDSP, then you should buy the calibrated mic for it. As DIY as you are you will need to test your horn and woofer. You may be using the old tried and true passive xover calculations. Try finding a program that helps design an xover with a flat impedance. ALK Engineering sells xover design software. It really makes a difference with SET amps. Trust me on this one. If you don't mind active xovers then try one of the Marchand active xovers. I can't recommend them highly enough. They really are great, and they have little pc cards to change the slope and type of crossover. It makes it very easy to try different arrangements. Symmetrical, asymmetrical, anything you want.
Lastly a conservative recommendation for your set up might be a second order slope at 500Hz, if your woofer goes that high. Third order slopes are good for horns, and at least they actually do something, unlike first order slopes. First order slopes are good for direct radiators generally, or direct radiator only speakers I should say.
I hope you get to a point where your technical side and the musical side come together for you.
Big speakers and little amps blew my mind!
Post a Followup:
Post a Message!
This post is made possible by the generous support of people like you and our sponsors: