"Splitter" modification for Belle and LaScala

[Jim E]

From what I have gathered, the splitter located in the rear of the Belle and LaScala cabinets facing the LF driver is about 13" in height by approximately 2 1/4" wide. The angle on the splitter face is 45 degrees centered on the woofer vertically. I believe the purpose of the splitter is to direct the sound out from the center of the cabinet to the outside of the folded horn. It would also help prevent reflected sound from re-entering the LF slot and improve high frequency output to some degree.

Question is, what determines the correct depth (thickness) of the splitter? As it comes from the factory, it is around 1/3 of the distance from the motorboard and rear panel. It occured to me if the splitter was formed closer to the driver it would improve overall performance. It would also make sense if the cabinets were angled on the interior rear corners reflecting the sound from the motorboard/ramp assembly to the front of the cabinet. Given the frequency plots I have seen, the low/mid crossover point and the problem of small compression horns loading down to 400 Hz. I would think improvement in this area would be of benefit.

As this is a 44 year old design, has anyone experimented with this?

[djk]

The LaScala is crossed over at 400hz. Figure out how long that wavelength is.

What do you think? (hint: peak signal amplitude is at 1/4 Wave, round trip paths add, so use 1/2W)

Do the same for the tweeter.

Now you know why Klipsch went to the flush mount Z-bracket on the newer Klipschorn and Belle models.

[WMcD]

Yeah, I can see the concern. I call it "the prism" and I've cogitated while building quite a few bass horns. All for nothing.

BTW, the same structure is used in the K-Horn.

The overall thought is that conditions at the throat of the horn (the small end) is critical to high frequency performance and close attention to the structure is needed.

Klipsch went to a small openning in the motor board. It is 3 x 13 rather than 6 x 13, as would be expected. Bruce Edgar wrote about this in SpeakerBuilder. He retracted his analysis in a following article. Essentially, he argued that it creates a small throat, and then realized it does not.

My thought was that PWK went to the small openning based on the science that a chamber in front of the driver can form a resonant element which boosts the high end. Barenek's (sp) book on Acoustics shows this with midrange horns. The problem is that the throat size in the Klipsch bass horns otherwise is so big (about 75% of the driver area) that there is no chamber formed with a normal 6 x 13 openning.

That was the subject of a post by me long ago. I think it got destroyed by a crash of the BBS a few years ago.

A year ago I did an analysis of a front chamber. It was based on an electrical analogy, per Don Keele, in a "SPICE" based circuit simulator. The result was that the effect of a front chamber is far less than Barenek's diagrams would lead you to believe. So much for pretty diagrams by Ph.D.s who have not built anything in the garage.

Evidently it does have SOME effect and PWK decided to keep it. Mr. Hunter of Klipsch sent Bruce some curves showing the effect and these were published (note there was a mislabling, so be careful to check "corrections" in the following issue of SB, if you go there). Perhaps I'll post those.

Bruce also has written about the "reflectors" at the corners of horns where the expansion takes a 90 degree turn. The thought is that without them, the area of expansion has a burp and thus this needs attention. You'll also see comments that the prism and the "reflectors" do indeed form a surface for directing HF waves.

- - -

Larry Clare, a distinguished member of our forum, reported that his K-Horns have a unique structure in that area which makes the geometry more precise. At least I think it was L.C.

= = = =

I do think that any analysis of the situation which envisions reflections, phase issues, or small errors in the cross section of the horn cross section are fundamentally flawed.

The reflector thing is wrong based upon wave mechanics. The rule of thumb in wave mechanics is that waves are not influnced by objects smaller than half a wavelenght. A 400 Hz frequency has a wavelenght of just over 24 inches. So there is not going to be any reflection by something measuring 3 inches.

The phase issue is wrong for the same reason. You'd like a precise, tight turn to keep the wave from getting delayed by following a long path. I.e. part of the wave makes a short turn at the apex of the corner, and the other part goes around a longer radius. The delay gets things out of phase.

That is good science, but bad arithmetic in this setting. The path difference is just not so bad going around the bend, particularly if you look at the central path. It might go plus and minus 3 inches, but that is not much for a 24 inch wavelength.

Any error in cross section is also not enough in this setting either. At best they exist over 2 or 4 inches. Given that an LS-Belle has a total lenght of 36 inches, the overall expansion will dominate.

- - -

PWK had stated that you can't argue with Physics. I think that is the situation here. Physics dictates that relatively small structures don't have influence. It seem to me he tried to challenge his own argument, and found he was correct in the first place.

The bottom line is that there is not much to do to improve PWK's designs. He was the master.

Be happy with PWK's design. He didn't miss anything.

Best,

Gil

[Jim E]

Thank you Gil. All excellent answers. I do agree that achieving any fine control in this design would be a fruitless endeavor. What I'm looking for (hopefully) is to improve the upper cutoff of the low frequency bin. Even a modest upward tilt in level above 3oo~400 Hz. would help.

I am in the planning stages of building a pair of Belles from scratch so modifications are no problem. The external dimensions will remain stock because they look great just as designed. Stock cabinets from the factory would be nearly impossible to modify.

So the question still stands science be damned (with all due respect). Has anyone actually done modifications to the splitter or "prizm" and/or added reflectors?

[djk]

"So the question still stands science be damned (with all due respect). Has anyone actually done modifications to the splitter or "prizm" and/or added reflectors?"

Yes, I have. It is a waste of time. Both Gil and I explained why.

The mass corner of the K33 is the limit on HF response.

If you use a high Q low pass filter at 400hz like the newer models that will boost the output at the crossover point.

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=/netahtml/srchnum.htm&r=1&f=G&l=50&s1=4,237,340.WKU.&OS=PN/4,237,340&RS=PN/4,237,340

Basically add 100µF~125µF across the woofer terminals. Klipsch used film caps, not electrolytics.

[WMcD]

I'm gonna stick to my guns here and go with classic wave mechanics.

I believe the LS or Belle does not have a problem with the high end. If anything, it works too well.

One thing to look at is the mouth size. It is about a two foot square. This means it does beam and there is a narrowing of the output with some gain. Note that this is consistent with the previous post. The mouth size is a wavelenght at 400 Hz.

Gil

[Jim E]

----------------

On 12/5/2004 7:15:56 AM djk wrote:

"So the question still stands science be damned (with all due respect). Has anyone actually done modifications to the splitter or "prizm" and/or added reflectors?"

Yes, I have. It is a waste of time. Both Gil and I explained why.

The mass corner of the K33 is the limit on HF response.

If you use a high Q low pass filter at 400hz like the newer models that will boost the output at the crossover point.

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=/netahtml/srchnum.htm&r=1&f=G&l=50&s1=4,237,340.WKU.&OS=PN/4,237,340&RS=PN/4,237,340

Basically add 100µF~125µF across the woofer terminals. Klipsch used film caps, not electrolytics.

----------------

DJK,

I hear what you and Gil are saying. Since you have tried the modifications what was the end result? You stated "It was a waste of time". What qualifies this statement? Please understand, I respect your opinion and appreciate the information on the subject. I am not challenging anyone, just looking for an honest unbiased answer. Like many of the Forum Members, I am on the same quest for better audio and I'm sure it will go on 'till I lose my hearing or kick the bucket. The interest in Klipsch "Heritage" speakers is kind of new to me. I have been a "horn" addict for thirty years though. As you fellows have been involved with Klipsch speakers so long, your information is of great value. I have gained a lot of information on this site over the last few months. There is much passion here.

The K33 woofer should be capable of response out to at least 1 kHz. as a direct radiator. The foldings of the bass bin will limit (rolloff) the upper end. I will not be using passive crossovers as I prefer active DSP based electronics.

[LarryC]

----------------

On 12/4/2004 1:25:21 PM Jim E wrote:

From what I have gathered...the angle on the splitter face is 45 degrees centered on the woofer vertically. Question is, what determines the correct depth (thickness) of the splitter? As it comes from the factory, it is around 1/3 of the distance from the motorboard and rear panel. It occured to me if the splitter was formed closer to the driver it would improve overall performance.

----------------

Jim,

FWIW, I measured my K-horn splitter, and as best I could measure, the splitter is 1 1/2" high and reaches about half-way across the 3" front-to-back distance of the first stage of the bass horn. While none of that may apply to your Belle, I thought you might be interested in the following cross-section drawing of the splitter, horn throat area, and 3-inch wide cavity opening to the woofer. You are looking from the side; the 3" high cavity opening to the woofer is to the left, up is "up" and down is "down." The drawing is not perfect, as I think the splitter triangular cross-section is not as flat as shown, but more nearly an equilateral, suggested by the 3" base and 2.5" sides:

Regarding your thoughts about bringing the splitter closer to the woofer, I'd definitely be careful about that. The drawing above shows some factory wooden inserts (in pink), which I then had removed. Expert conclusions were that they were put in to raise the upper range of the bass horn output, probably for the original 1962 drivers and crossovers.

They had a very adverse effect on my K-horns' frequency and phase response until they were taken out. With current drivers and X-overs, they boosted the midrange and greatly weakened the deep bass, and apparently (as I understand it) caused a non-linear woofer action that might also have affected mid-range output. Details of the adventure can be seen in my thread, Restoring the standard bass horn throat in a '62 pair of Klipschorns.

The inserts acted by narrowing the entrance to, and increasing resistance at, the bass horn throat. You might not want to cause a similar effect. Perhaps D-man might chime in here.

Larry

[D-MAN]

The splitting wedge on the Khorn is quite a bit larger than the LS/BELLE splitters and PLEASE NOTE: the throats and throat cavity openings in all 3 are EXACTLY the SAME SIZE.

So much for physics. The original point of the Khorn wedge was to "follow" the same angle as the cone itself as described in some early patent material (not Klipsch's, if I remember correctly). This is somewhat strange in that the cone itself is conical, and the wedge, of course, does not follow a conical form at all. However, you will note that the angle does "follow" the cone's angle when seen in cross-section.

The LS/BELLE splitting wedges are quite a bit smaller and seem to be there as almost an afterthought. The purpose is to bifurcate the waveform and turn each respective channel 90 degrees. The angle prevents reflections back through the throat cavity opening.

I conclude that the main purpose is to bifurcate the throat into 2 equal channels. Second to promote a turn with the least turbulence possible, third to prevent reflections related to turbulence.

PWK noted in his patents that there were "better" ways to turn waveforms, but that straight angles were used for material and ease-of-manufacture purposes.

The horn throat is the area of highest resistance and is the area where the highest distortion occurs. Naturally, the throat is also the area were turbulence could best be avoided as much as possible.

I would suspect that the LS/BELLE would operate just as well if not better with a Khorn splitting wedge.

DM

[D-MAN]

----------------

On 12/5/2004 12:02:51 AM Jim E wrote:

Thank you Gil. All excellent answers. I do agree that achieving any fine control in this design would be a fruitless endeavor. What I'm looking for (hopefully) is to improve the upper cutoff of the low frequency bin. Even a modest upward tilt in level above 3oo~400 Hz. would help.

I am in the planning stages of building a pair of Belles from scratch so modifications are no problem. The external dimensions will remain stock because they look great just as designed. Stock cabinets from the factory would be nearly impossible to modify.

So the question still stands science be damned (with all due respect). Has anyone actually done modifications to the splitter or "prizm" and/or added reflectors?

----------------

If you want MORE UPPER-BASS (above 400HZ) then use a larger throat cavity opening than the 3x13 slot in current use. The horn throat itself is 78 sq. in., so you can go all the way to 6x13" (exactly as the original Khorn!). The smaller opening (as we see it today) is to REDUCE the passage of upper bass frequencies through the horn.

The throat cavity opening is a physical upper-band-pass limiter. When sized 1:1 with the throat area, you will have no band pass limiting other than that effected by the horn itself and the driver used. Bear in mind that upper bass frequencies do not take to turning though folds very nicely, so there may be audible ramifications, such as peaks and valleys in the response.

That's why PWK changed the size of the cavity opening in the 60's to deal with the K33 (from the CTS to the Eminence version, I beleive) frequenciy response for the Khorn. He evidently kept the size for the LS/BELLE too. He must have tested it out and wanted it that way, although it need not be a "hard and fast" rule.

Also, if you want to use an 8 Ohm driver vs. the K33E, then a 6x13" slot would be in order. The slot opening tends to cause high resistance to the driver (aka: horn loading) and therefore, the larger, the less resistance "seen" by the driver, which in the case of an 8 Ohm driver, is already higher than optimum. The driver VC impedance and the slot size work in concert when loaded into the horn. When choosing other suitable 8 Ohm drivers, this consequence is not particularily apparent from T/S parameters, as the reactance at the throat, the slot size, the volume of air in the horn, the length of the horn, the mouth size and all need to be taken into consideration.

DM

[Jim E]

Larry and DM,

Thank you. This is exactly the type of info needed. I read Larry's post on the K-Horns. Excellent post and very well written. It is interesting that just changing the spacing of the woofer mounting to the motorboard affected the performance.

It looks like I will have to construct a re-configurable cabinet and run some tests. Are there any area's on the Belle design you guys feel could be optimized beside the splitter and LF aperture?

Thaks again,

Jim

[djk]

[D-MAN]

Jim, I would add corner "braces" to aid turning the waveform at the corner(s). PWK left these out, but they are used in the Khorn.

45 degrees ought to do it. Try to leave a full 3" clearance in the channel.

DM

[D-MAN]

Jim, I'm using a "convertable" throat opening and motor board. Here's a pic of the back chamber of my corner horn.

The motor board mounts to the carraige bolts seen in the pic (motor board removed). The slot in the motor board is 3x13" for the K33E, and another motor board is slotted at 6x13" which matches the 6x13" baffle throat cutout seen in the pic. Convertable! Works great, I recommend it highly.

This should look somewhat like the insides of a Belle or LS, except that it is much deeper, of course.

DM

[swells]

I am experimenting with a scaled version of a Khorn ( ~81.25%) using the drivers from a Heresy. what kind of problems will I get into regarding a cross over. If the orig 3x13 cut freq. ~ 400 hz how would I reach the 700hz the Heresy was crossed at. or should I look at trying to cross at the 400 range.

As I mentioned, this is an experiment mostly because; first, I have the drivers at hand.

second, the size is more suitable for my room and finally, I get the chance to see if I am capable of building a khorn cabenet.

[D-MAN]

The cross-sectional area of the horn throat of 78 sq. in. at 81% size, the calculator puts it at around 63 sq. inches.

This correlates to a throat cavity opening as large as you can make it, 6x10" or so. I would not use a smaller opening than that. The crossover could be upwards of 600Hz, I would think, possibly as high as 800Hz, but there would expectedly be some "coloration" issues contending with the folds. But maybe not noticable, who knows?

DM

[D-MAN]

Jim E, one more thing, ironically, the corner "braces" can be the exact same size as the splitter. This diagram is from the La Scala plans posted by Andy some time ago.

Just use extra long wood when cutting the splitter so that there is enough for the 2 corners, too.

The dimensions and angles work perfectly.

DM

[Jim E]

DM,

Thank you for the info. Here is a quick drawing of the basic mods I am considering. The "blue hatched" areas indicate the modifications.

Jim

[Dylanl]

DM, I noticed you use fiberglass in the woofer chamber. What effects does this have on your bass output?

[D-MAN]

----------------

On 12/12/2004 9:09:28 PM Dylanl wrote:

DM, I noticed you use fiberglass in the woofer chamber. What effects does this have on your bass output?

----------------

It cleans up the bass. There is one more piece not shown in the pic that goes right behind the woofer basket.

This is to lessen the amount of reflections coming back through the cone (and into the horn) out of phase.

DM