|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: On Dinsdale on Horns. (long) posted by Paul Eizik on April 16, 2003 at 13:44:02:
Well, Paul, you just got sucked in by Dinsdale's mishmash of right and wrong horn theory. I was also impressed until I started to take apart his design philosopy and found some very wrong assumptions.The first wrong assumption is the use of any old driver in a horn. He used KEF drivers which are mass loaded. If you place these drivers on a horn, you lose bandwith. Notice that he does not present any frequency response plots.
The second wrong assumption is the design of a tractrix bass horn. It's an contridiction in terms. You can't get bass out of a tractrix design.
The third wrong assumption is bad things happen at the flare cutoff so Dinsdale says to back off by 20% the flare frequency. Some software out there uses this assuption. You plug in a frequency but you end up with a horn of lower flare frequency (20% below) that is longer than you need. But even worse, the mouth size is calculated for the input frequency, not the lower value. So the end result is a lower flare rate and a longer horn than you need with a too small a mouth. I alway do my own horn flare calculation.
I would say that about 5% of the horn literature out there is good, the rest is a mishmash of garbage and good stuff. Caveat Emptor!
Follow Ups:
BruceI have'nt gone over to the dark side, honest! Back in 1974 (at the time of Dinsdale's articles) my Dad had dismantled his horn rig, as my Mother had gotten sick of the sight of big raw plywood boxes in the living room for 10 years (she was very patient). I started cobling my horns together as a P.A. system about 1977, the same year "Suitability of Low-Frequency Drivers for Horn-Loudspeaker Systems" by R.H. Small was presented to the A.E.S. I did'nt read Small's paper untill last year as I previously thought you had to be a member of the AES to get the papers. It looks like Dinsdale just picked some good drivers that were handy just as I did, but I lucked out and picked some EV 15B's which were on sale due to my intended P.A. use. It seems that you had your great horn revelation about 1978, but I did'nt get my horns up and running untill late '79 during a very cold spell here (-25 below). I put one together and was astounded that it totally overpowerd one of my Altec direct radiator Model 9's, I could'nt even hear that the Altec was playing! The other horn was quickly assembled, and the resulting rig crowded out the Altecs as my main system. The whole thing was a bit on the rough side (to put it very mildly from todays perspective, but it had that "big sound" that my Dad's rig did) and I could'nt think of what to do about this untill I orderd some back issues of Audio Amateur and Speaker Builder and became one of the horn hungry few who eagerly awaitd your once a year articles and occasional answers to letters. Any practical horn info was hard to come by in those days, resulting in a lot of guess work. I have to look at the Dinsdale articles in this context.
A while ago there was a link here to a European horn site supplied by Bill Geiger (Bill, we miss you!) which featured a ceramic brick wall horn. The designer outlined his design protocol and settled on the tractrix contour for the bass horn, rejecting the hyperbolic for what he felt was it's high distortion. I immediately thought "Dinsdale", and "this guy has'nt read Edgar obviously". Anyway it's good to see your critique of Dinsdale here (I've heard it in your talks), as I would hate to have a multi ton brick tractrix bass horn in my basement. I hope our European readers see this, as the wrong horn costs and weighs just as much as the right one does.
Paul, Bruce:I too have read Dinsdale. Moreover, in developing the Horn Kit I used one piece of information from Dinsdale. That's this notion. If one is not trying for maxiimum efficiency, then one may chose to use one of those horn expansions which do not result in as high a pressure proximate the throat where the driver mounts. In doing so, one avoids, or perhaps more correctly, inceases the safty of margin, distortion due to "over" compression of the air column near the horn throat. That's why I chose a parabolic horn.
On a somewhat related note: Bruce have seen and if so, how do you feel about the discrete element model for horn expansions developed by Jean-Michel LeCleac'h?
See:
http://perso.club-internet.fr/ndaviden/pavillon/lecleach.htm#ancre68820
Hi LarryThe only hi-fi variation on the low pressure side of the horn family which I am familiar with is Tom Danley's Unity Horn, which uses a conical horn with multiple drivers in the horn walls and one in the throat. I've also heard str8arrow's D.I.Y. version. The overall sound of these horns is more toward the sound of the direct radiator, in other words not as dynamic and focused as a compression driver, but also quite smooth generally. In a horn the pressure factor only becomes a problem in a P.A. application where you are pushing 120 dB continous, and is not much to worry about in the home. If you look at Dinsdale's graphs, most of the pressure distortion in a horn is 2nd harmonic (the octave), and this is the same distortion that the solid state delegation and the scope jockeys had warned us was so evil in S.E.T. amplifiers. For me, the ultimate midrange is the tractrix horn loaded compression driver, and the entire rest of the audio system should be reverse engineered from this. Just my opinion you understand. Most of us are familiar with the parabolic horn in a re-entrant configuration, typically in an industrial paging system, but if the conical horn can sound good, the parabolic must have hi-fi potential too. How did your horn work out?
Are you sure that it's a parabolic? I think you meant hyperbolic. Bruce
hey Bruce - horn lore sounds like a hot-dog - haven't read Dinsdale since new - completely forgot about KEF - Percy Wilson's windup acoutic player may have beat thosetake good care,
Fred
This post is made possible by the generous support of people like you and our sponsors: