Early folded horn questions

[WMcD]

[WMcD]

[WMcD]

Couple of things:

DJK has a good point about Figure 5.11. As the horn gets shorter (throat bigger) the resistance is rolling off more quickly as we go down to Fc.

OTOH, please note that in each successive set of data, the throat diameter is doubled. Comparing any one to the next does not show too much difference. Increasing the diameter by a factor of 2 increases the area by a factor of four.

In our orginal point of discussion, we were only increasing area by 2 or diameter by 1.414. Therefore, the effecive resistance is not going to change too much.

I'll point out that in Figure 5.10.D has a mouth diamter of 40". One rule of thumb in the old days was that the mouth should have a circumfrence equal to the wavelength of the Fc. The wavelength of 100 Hz is 135 inches. The circumfrence of a 40 inch circle is 125.6 inches. So it is close.

Gil

[D-MAN]

Marvel,

I have read that a waveform suffers virtually no ill effects when going through a column or a pipe of unchanging geometry and especially of short lengths compared to the wavelengths envolved, such as that occuring at the 90 degree corners between the two expanding channels in the La Scala. As far as I can figure, it's use is to keep the outside (and back chamber) dimensions within the desired values. It has no additive effect on the length of the exponential horn(s) themselves.

Anybody check me on this?

DM

[WMcD]

You may be concerned over how to measure the lenght of the horn when there is a hard turn. Do you consider the inside of the turn, which adds no length, or the outside of the turn, which may add three or four inches.

It looks to me that some designers are taking the middle path. On the other hand, in the LS, this is only 10% of the overall lenght. If you look at Olson's calculations of throat impedance and compare one to the other, this is not so much to worry about. I.e. if there is a difference in the Fc because of lenght, it ain't much.

One thing that PWK and others were worried about was the problem that the wave going the outer path (long, say 4 inches) and the inner (short say zero) might get out of phase because of path delay.

This is probably why PWK designed these speakers (KH, CW, Belle, and MCM) with two paths. The bends with a 3 inch width don't have as much outer radius (delay) as an equivalent, single, 6 inch width.

On the other hand, the Valerie type uses a single path. But there are no hard turns.

There is also the issue of whether the turns (of any radius) should be, or not be, make up of angular pieces of wood which are not curved at all.

Some designers take another "path" to honor a real curve. There is a fellow publishing in AudioExpress who uses plastic pipe on the bends. Also the early movie theater speakers look very much like French horns in construction.

Gil

[JohnA]

Hey Gil,

What's a "Valerie Type" (horn)?

[D-MAN]

Gil and all,

I ran across a reference that explained the benefit of bifurcating the horn as having to do with a limit on the location of bends in a folded horn as having to occur before the expansion reaches a certain ratio to the frequency (waveform) being transmitted. More or less (I haven't got it handy) the foldings need to occur close to the throat and should not occur after that (i.e., where the horn dimensions meet or exceed the value). Bifurcating the horn at the throat keeps the ratio from being reached compared to a non-bifurcated horn in that the respective horn dimensions are reduced in physical size in relation to the waveform which, of course, doesn't change.

I've also read somewhere that most designers take the "center of the channel" approach to horn path length calculations. As I stated in my previous post I understand that a straight conduit does not harm the waveform - however the 90 degree turn in the LS has no expansion qualities that I can see but also that no attempt was made to accelerate the waveform around the bend using a triangular outside brace as is the usual practice (like the Khorn) and others, I am wondering why PWK decided that it wasn't worth the effort... perhaps just a manufacturing shortcut?

DM

[WMcD]

I think I said what D-Man said. He said it better.

The Valerie is the bass horn with a unitary sound path.

It may be that the LS doesn't have any issues at the top end of the bass range and therefore the extra pieces are not required.

The LS may be beaming and therefore getting some treble gain on axis. If you look at the paper on the Jubilee, you'll see that getting the two mouths close together and facing forward avoids losses in the treble end of the bass. Naturally that descibes the two mouths of the LS.

Gil

[HDBRbuilder]

What Gil refers to as the "Valerie" horn is actually called the LB-76. LB stands for "little bastard", and 76 was the year it was designed/built. It has a monopath basshorn which fires rearward(downward at an angle), then curves downward and forward...pushed by a 12" driver(kinda like the university classic on its side with its mouth on the floor...but with fewer bends to the "nautilus-shell" horn lens). The top end uses a K77 tweeter mounted vertically at one end of the horizontally-mounted midrange horn which was, I believe, a Cornwall mid-horn, fired by a K55V driver. It was intended to be a lower cost replacement for a LaScala in applications where a smaller room was to be used...it also was intended to be easier to man-handle around due to its smaller size. Another novel innovation was the addition of two panels, one to each side, hinged along the left and right front edges of the cabinet...these were to "tie-into" the walls when the speaker was used in room corners...and had something to do with elimination of possible standing waves and rear reflections of the speaker cabinet. When PWK showed it to Mrs. Valerie, and told her he designed it for her, he mentioned that the builders of the prototypes called it "a little bastard to build"...and he asked her what she wanted to name it...she responded: "Why not just call it The Little Bastard?" So the nomenclature originally given by its builders stuck. There were maybe 5 or so pairs of these built. It was deemed that the production cost of them would not allow for them to be priced significantly lower than the LaScala, so they never went into regular production. While I worked there, we had at least three pairs set aside for employee usage as "loaners"...along with some LaScalas and Industrial Heresys. The PWK biography has the drawings for the LB-76 in it, but no dimensions are given.

[D-MAN]

Good thread, you guys...

Again, you have brought up ample things to think about that I have not run across before...

When I built my DIY horns, I took the approach that every obstruction in the exponential channel(s) was to be avoided. I've even read that some designers/builders even recomend painting the interior of the horn with high-gloss enamel for that reason (which I did NOT do, but again, it SEEMS REASONABLE and SCIENTIFICALLY SOUND).

The occurence of a 90 degree "naked" channel corner is unusual from that mindset... seems like an cause of turbulence to me. It may be pointless, but I am afraid that I would want to "fix" it if I were building them...

The question that I am leading to is this:

Is there anything to this point of view or is it just a waste of time?

DM

[Tony Reed]

HDBRbuilder,

Do you remember what crossover network was in the LB-76? Also, did it use the same 12" woofer as the Heresy? I wonder what happened to all those LB's.

Thanks,

Tony

[WMcD]

I think the patent document showed a K-33. What was used in production I wouldn't guess at.

Re the LS and the corner piece: It could be that PWK had two built, on with and one without. If he determined, after testing, there was no significant difference, he may have just decided to not use the corner piece. I doubt it is hurting anything.

It is just an engineering mentality saying there is no reason to make something more complicated than needed. It adds a bit of cost and of course he had to watch the bottom line.

As home builders, we're in a different situation. We've got a lot of time involved and our labor is cheap. Also, we can't test prototypes with expensive equipment and be certain. It is only prudent to put in the piece. That way, we don't have to wonder in hindsight, "I wonder if it would have sounded better if I built it with the extra piece?"

I wonder if I'm arguing against my general position against superwire, super connectors, and other minimal tweeks. Maybe. OTOH, it is possible to reverse those installations and perhaps learn they are minimal tweeks. My thought is that the extra pieces are very cheap compared to the superwire and connectors. So why not.

Gil

[HDBRbuilder]

The LB-76's I saw had their own crossover networks...not like any of the Heritage models...matter of fact, of the two pairs I borrowed at different times, each pair had different crossover networks in them....one pair had so much crossover network it took up two boards in each cabinet!...that pair was VERY "busy" under the hood in the H/F section! A 15" woofer MAY have been used in them, but they were narrower than LaScalas, and that 15" woofer would have been at least as snug a fit as it is in the LaScala doghouse, if it was used. Because these never went into standard production, we paid little attention to them when I worked there, other than the fact that they were fairly popular "loaners"...but they were great sounding little speakers...I often took them to the Little Missouri River swimming hole north of Blevins for BBQ/parties on the gravel bar there in the summer months...I just removed the covers on the door speakers of the old 1975 Datsun pick-up, pulled the speakers, and alligator-clipped some speaker wire to the door speakers' wires and ran it out to the LB-76's, and then let the Blaupunkt in-dash cassette deck rip with its meager ten watts max output and those LB-76's would fill the riverbed with great sounds! It normally never took a volume setting even halfway up to get the output up to party levels!

[WMcD]

You'll see the patent document(s) shows the crossover with values.

Sorry I don't have the numbers here.

Actually, there was one patent application. It claimed both the throat structure and the use of an autotransformer to increase the output of the tweeter. In the vaguries of our patent law, two or more "related" inventions can be claimed in one patent.

Evidently the patent office felt there were two distinct, unrelated, inventions. This makes sense in that the tweeter crossover had nothing to do with the woofer structure. One was divided out to make a second patent award. This is why the patents eventually awarded look so similar, except for the claims.

If you read the specification section of the patents, the crossover components (values) were selected to have some peaks where the response was falling off.

Best,

Gil

[mungkiman]

I believe this is the bastard of which you speak. Patent # 4,138,594. It mentions a natural cut-off of close to 67 Hz, using a 15" K-33-E. Also mentions crossing over to a K-55-V at 400 Hz., and to a K-77 at 6000 Hz. PWK goes on to mention 108 dB SPL output at 1 meter with 1 watt, and 200 peak watts input capacity.

[WMcD]

<img src=http://forums.klipsch.com/idealbb/files/valerie1.gif>

[WMcD]

[WMcD]

[WMcD]

Ahh, finally got that working.

Anyone interested in horn design should go to the USPTO site, get the special GIF plug in, and download the full patent document.

Just about all you see in the patent (any patent) is provided by the applicant (or applicant's attorney) as the application. I suspect PWK wrote most of it himself.

The result is that the patent is a primer by PWK himself on horn design. The text can be a bit challenging to read in a few places. In many ways it is more straightforward than much of his other technical writing.

Only a few dimensions are given. They are the mouth height and width. We know the woofer has a basket of about 15 1/8 inches. That is certainly enough to allow someone to build one with a little scaling from the drawings.

Best,

Gil

[Tony Reed]

I can't seem to scale those drawings. If you go by any one of the three know dimensions for your scale reference, the other two dimensions don't come out right. Am I doing something wrong?

I'm intrigued by this little speaker. Any info would be appreciated.

Tony