drboar

Agree about the walls. The measurements was on Klipsch clones and the annotation is the other way around. In a proper corner with both walls concrete the horn kept going to 40-50 Hz. With the outer wall still concrete but the backwall in flimsy plaster the response below 80 Hz or so lost 5 dB. Mesurements were made 30 years ago using a Technics SH-8000.

The curves below are from "Kuben" a horn similar to the La Scala horn but in this case using 12" drivers. The measurements was with the horn on a table and on axis on about one meters distance. Upper trace is Beyma PA driver Fr about 60 Hz and Qes of 0.2 a typical good horn driver. The lower trace is a Bulgarian Gamma Gamma 1231 driver, light cone very soft suspension and quite high Q (25 Hz 0.42 Qts and 340 L Vas) so the mass corner frequency is about 100 Hz. For PA use the beyma is far better more output by far for the same voltage and greater powerhandling as well. For domestic use a trade off between relative bass output and sensitivity might be scewed in the oposite direction. I am not claiming that frequency responses are everything, they have to sound good as well.

I have been playing around witha LS type bass horn. The throat extends into the dog house and the driver is a 12" But it is still a short bass horn with the same mouth area as a LS. In my horn at least the basic response is set by horn lenght and mouth area. Different flares and chamber sizes does not make a big difference. These things can be used to change the knee above the high pass drop off.

That is cool looking horns, I have some plans to build curved La Scala type horns, I do not know if the sound will differ but as modeled in HornResponse, the actual horn shape (within reason) have minor effects on the modeled response. So with a set lengths mouth size and throat then changing the curvature and chamber size mostly change the roundness of the knee at the high pass frequeny. BTW, do you have any ideas why 4 x 8" sounds cleaner than one 15"?

Cleaning up the seals and also add some crossbraces to the side walls as per Volti http://www.klipschupgrades.com/lascalaupgrades2.shtml That horn slam is addictive

Regarding corner reflectors. I did some experiments on vintage horn "Kuben" very similar to but not inspired by the LaScala. I tested a slew of drivers and adding corner reflectors (attached by double sided tape) was beneficial to some driver, harmful for some and, did nothing for others. Above 300Hz cone breakup causes different effective pathdifferences for different drivers (I assume) and the complex bouncing around can either add up or substract with or without those reflectors. An optical reflector like a mirror is easy to grasp but accoustic reflectors does not behave in quite the same manner. Vibrations in the sidewalls is a problem If I was to build them with no braces I would work with laminating two different materials rather than doing brute force things like 30mm MDF.

Have you seen this crossover update? http://www.troelsgravesen.dk/JBL_L100.htm

Calculate 1.the chamber volume 2. Horn lengh 3. Mouth area 4. Throat area Play around in horn response and you will discover that flare does little to change the bass. Parabolic section, conical, exponential or hyperbolic change some few dB and the roundness of the high pass knee. The chamber volume changes that roundness to. The low pass frequency is determined mainly by lenght of the horn and the unevenness/"peakyness" of the response is set by mouth area and radiation angle. Curvature and implementations of slots and bends will affect the real life low pass response but will do little in the bass response.

Bruce Edgar studied the "rubber throat" of the Klipsch and concluded that the throat constriction was changing path lengths for the driver-horn start and that this benefical for the output in the upper range of the horn. If that still holds true and how driver dependent that is I have no idea.

Braces like these http://www.voltiaudio.com/newsletter/NL8.shtml

I tried out corner reflectors (Kuben in diyaudio forum) and the result varied with different drivers. I expeced the reflectors to allways improve things but it did not. My guess is that there was different breakup pattern in different drivers so different phase shift was introduced already on the way into the horn. The shift in phase beaviour with the added reflectors could then be benefial or harmful interference.

The Jericho 08 IHR is a 50mm hole into a chamber. No vent tube or anyhing fancy just a drilled hole so something simlar in LC seems feasable

I have seen other horns (Jericho 08) that use a small internal helmholtz resonance chamber to flatten the horn response. Outside the first flare section between the dog house and the back wall there are two unused closed spaces. Have anyone tried to use them as IHR and reduce the LS peak at about 150 Hz? In a PA setting it would make more sense to EQ it and keep the radiation resistance but in a domestic setting, why not trying to tame the peak?

I have been playing around with various flares of a LaScala type horn. That is a horn about 1 m long and a mouth area close to 0.36 m2 Different flare shapes like exponential, hyperbolic and of different flare rates as well as chamber volumes from 30 to 80 liters. At least to HornResponse the differences seems small. The overalll horn response is set by mouth area and length not the exact flare (within reason) The chamber volume can be reduced to produce a sharper knee at the low pass and also reduce cone motion below the bandpass of the horn. I have my own ideas about making a La Scala light version with a curves and a twist added to it. LaScala does not have bass below 50 Hz but there is something about the slam of a large mouth horn...

The sidewalls of the LS tend to vibrate and slur the bass. Crossbraces are a good thing. If I had pair I would also replace some caps as suggested by others.

If you look at vibration isolation of either sensitive equipment or vibrating machines, nowhere outside the HiFi world there metal spikes is used as way of insulation. That claimed diode like action of spikes suggest an exeption to Newtons third law of motion...

I had my DIY Klipsch Corner Horns in solid concrete corners and had good bass down to about 40 Hz. In my next home the side walls was still was concrete but the wall behind the horns were drywall. I lost about 8 db below 100 Hz, I was thinking about bracing the wall but then other factors came into play and I had to scrap the horns. If you corner load the K-horn with flimsy walls you will get nowere near the bass the horns offer.

There is no rise in effiecency. Lowering the impedance below the bass resonance increase the power input to the driver and hence input. The idea was that the LS is a closed box below the working range of the horn. I will go back to the "true" closed box and try other driver to see how sensitive this the bass extension, peak flattening is to driver variations and system Qs.

In a high Q closed box the bass can be extended and the peak tamed with addition of a serial cap The cap increase the impedance above the resonanse frequency (reducing ht ehigh Q hump) and lowers impedanse below resonanse (increasing output) This is a KEF B200 with a Q of 1.0 to 1.1 in this box A horn similar to the La Scala also is a closed box with a high Q. Adding caps do increase and decrease the impedance as they should using a Eminence PA driver And then the proof of the pudding What the heck? Is it the higher Q or is it something intrinsic in hornloaded boxes that precludes the use of serial caps to extend the bass?

I had the impression that the lower the Q the better for a horn driver I have "Kuben" something that looks simlar to a La Scala but with in my case 12" drivers. The drivers were measured using Omnimic at 50 cm distance with the horn on a chair in the middle of the room. Same voltage in all measurements Black trace is a Beyma G320. A real horn driver with a Q of below 0.25 and Fr of about 60 Hz Red trace is a Eminence driver Q of 0.5 and Fr about 50 Hz Brown trace is vintage Gamma LA1232. Fr 25-30 Hz Q 0.4 and Vas = 300 liter it should be unsuitable as a horn driver and have a mass rollof above 150 Hz or so. In a PA setting the Beyma is vastly superior to the Gamma with 10 dB higher sensitivity and better power handling as well. Hower, if we compare the -10 dB point relative to the averege 200-1000 Hz output the Beyma reach -10 dB at 90 Hz while tha Gamma has this point at 55 Hz. It does appear that suitable Q in horn is a matter of application and that mass roll off and bass horns not allways match up

I agree with the previous post. A timedomain EQ can not fix everything in a passive loudspeaker. In the ranges were only one speaker contribute or at least much more than the other ones say 20 dB it is possible to time EQ. For example 10-20 kHz were only the tweeter is working.The problem is in the crossover region were two drivers are working perhaps around 5 kHz were both midrange and tweeter is contrributing. what every delay you will have at 5 kHz will only fit one of them. I have built speakers with passive crossovers since the late 70s and recently bought some second hand active crossovers. It such a joy to be able to adjust xover and level by turning a knob. I might use som passive crossovers but my guess in going forward to miniDSP is more likely than back to passive.

Real corners are important. I had DIY Klipsch corner horns in solid concrete corners and then moved and then had the side walls still in concrete but the back wall fof the horns in plaster board I lost something like 6 dB below 80 Hz, so true corners and sturdy corners is a must for decent bass.

I could measure the effect of a filter trap using such simple tools as a XTZ Room Analyzer II both on a Goodman Axiette Fullrange driver and a dome tweeter, so qith any decent setup this should show up accoustically not only electrically.

I realised if I do the talk I better walk as well[H] This is the setup with the horn set against a back wall. XTZ Room Analyser II connected to my macbook (with WinXP in Bootcamp). Microphone is center line about 70 cm in front of horn opening. The driver in the back is the Beyma G320 (Fr= 52 Hz and Qes= 0.2) built to be a horn driver. The other is a Fane Crescendo 80W Silmilar Fr but a Qes of about 0.5 Starting with the Beyma horn driver With no reflector(green) and then adding the bisector in front of the driver (blue),not much happens below 1 kHz Adding corner reflectos add some dB jsute below 1 kHz and a huge bump around 2 kHz Adding both the bisecting reflector and the corner reflectors, brings us back at square one! Here we have the impedance plot Dayton Woofer Tester 3. System resonance at 77 Hz and medium Q value of 0.8 Note the gently sloping response below 200 Hz Turning to the Fane driver with a higher Q adding a bisector. Note the hump in the bass just above the HP. Here there is about 4 dB added both at 400 and 900 Hz by the reflectors Then adding both (I missed measuring just the corner reflector) Either some few dB added 200-800 Hz by te reflectors (blue) or back to square one Impedance plot show a Fr 10 Hz lower and higher Q of 1.4 Tossing in a Vintage driver with a Fr of 30 Hz Qes of 0.4 and Vas of 300 Liters! The setup is different the green line is tha Beyma and the green is the vintage driver that is a theoretical mass rollof of 150 Hz but a practial one is this case of 1200 Hz The images are cropped for some reason. The system with this driver has a Fr of 67 Hz and a Q of 1.4 suggesting having a low Fr higher Q driver might be better than adding ports and stuff. I also note that Klipsh original driver is far from the classical horn driver of say 50 Hz an a Qes of 0.2

Modeling vs measurement. I have a horn similar to the LS and to my surprise the low pass is way higher than calculated byt the mass rolloff. See my blogg at "http://drboar.blogspot.se/" Do note that a classical horn driver like the Beyma falls below 100 Hz using "worse" drivers with lower Fr and higher Q gives more bass output but also a hump at the high pass band Corner reflectors: I added both corner reflectors as well as a reflector in front of the driver. Attached by double sided tape I tried no 1. No reflector 2. Just in front of driver 3. Just corner reflectors 4. Both infront and corner I expected the effect to be additative 1 worst 2&3 better and 4. Best. Not so, it varied with the driver what combinations that worked best!