Big BenD Bass Horn: Belts and Braces Part 4
In this final article on bracing, we will look at the mouth section braces, and also compare the measurements of the final horn with some of the original, unbraced measurements. This article will be quite long, as there is a lot of ground to cover, and many experiments. As before, I can't show all the measurements, as there is just too much data too look at. But I will include measurements at strategic positions, which I used to evaluate the performance during application.
At the bottom of the article is a photo of the braces with letters to identify them (the letter R indicates the mouth without bracing). These letters correspond with the letters in the curve legends.
The bottom wall was fairly well braced from the start, as the braces were used as a jig. But additional braces were added later. We will take a look at the bottom wall when looking at the final results.
Mouth Side Wall
The first set of measurements show the vibration at the unbraced side wall at all the measurement positions. While quite messy, we see some strong modes at 112, 185 and 350Hz, the lowest is the bell mode of the mouth and is very hard to get completely rid of.
Adding the centre horizontal brace on the side wall, S, shifts the bell mode down somewhat, and cleans up another weaker mode around 58Hz. The range above the bell mode is also cleaner, but the upper modes are not much affected. Adding the T braces (upper and lower horizontal braces) helps in the range above 300Hz.
A vertical side brace, U, does not do much with the bell mode by itself, but it cleans up the upper modes.
At this stage I was starting to think I may have to apply some more drastic measures to the mouth to kill the bell mode. So I got some more angle iron, and applied around the mouth. Below are the measurements of these reinforcements by themselves, V indicates side angle iron only, W indicates side and top angle iron. It's actually not much of an improvement by itself, apart from at the 58Hz mode. Well, at least that indicates that something is right...
Some slight changes by adding the vertical side brace:
We'll revisit the side when more bracing is in place.
Without bracing, two things are clear: The absolute level of the peaks is lower (for the most part) than for the side, and the actual bell mode is the mode at 58Hz.
Adding the angle iron (W) reduces the bell mode by 5dB, and adding the front lateral brace (WB) reduces it by another 5dB. But these braces do not do much for the rest of the modes.
Adding a rear lateral brace (WC) and two longitudinal braces (WD) helps somewhat, adding another four longitudinal braces helps even more.
Mouth Cross Brace
There was no way around it: The mouth needed a cross brace. A cross brace is of course very efficient in killing the bell mode, but it's not as good for killing the other higher modes of the panels. So both are needed. Adding the cross brace certainly reduces the panel vibration over a wide range, the bell mode is now down by 25dB, and the rest of the range up to 300Hz is much cleaner. Adding Mutestrip between the two braces, where they intersect, doesn't do much, but I left it there since very little of the material is used anyway.
But of course the mouth braces themselves will vibrate too! The horizontal brace has two distinct peaks at 67 and 70 Hz and a strong one around 300Hz, while the vertical brace has several peaks above 300Hz.
In order to combat the resonances of the horizontal brace, which were the most troublesome since they were quite strong in-band, I tried various things like stiffening with steel L-profiles. But the most efficient turned out to be mass loading the arms of the horizontal brace with pieces of flat steel. The result of this is shown below, it took the level of the peaks down by over 10dB.
The braces added to the outside so far are shown in the photo below:
Order of Application
When braces are glued onto the side walls, it's not easy to experiment with the order of application. But since I'm building two horns, I decided to test out the order of application at the mouth section.
For the first horn, here are a couple of measurements with no bracing (R), all outer braces (WF) and finally outer braces pluss cross brace (WH).
If the cross brace is added first (includes also the rim braces, I haven't measured without those), there isn't actually too much of a difference when the outer braces are added.
But what really killed the vibration, especially the lower modes (at this position) was the addition of a second layer of bracing at the mouth, that can be seen at the photo below. This was an idea suggested by my friend Torbjørn at the very end of the experiment, so it is not included in the previous mouth bracing experiments. It basically turns the rim braces and horn walls into 'C' beams.
For the top wall, similar tests were done. Again, R is with no bracing, WF is with all outer braces, and WH is all braces including cross brace.
RA indicates the rim braces and cross brace, but now the outer braces (RD) makes some difference, especially around 100-200Hz.
The measurement of the efficiency of the bracing has been quite extensive. Over 300 measurements have been saved, but probably the same number have been taken in the process of testing things, finding good measurement points, and so on. The articles have presented only a select few of these. For those interested in a summary of "before" and "after" measurements for all measurement points, curves wiht a few comments are presented here.
For the most part, the bracing has been successful. Vibration has been reduced by at least 10dB most places, and in many cases a lot more. Several places the modes have been pushed out of band, other places they have been usefully attenuated. In a few cases, coupling has introduced vibrations in surfaces that weren't there to begin with.
Bracing and vibration are definitely complex topics, and this study has taught me a lot about practical bracing. The main points:
- The braces must be stiffer than the panel it is meant to brace, and considerably so. Otherwise they will just add mass and lower the resonance frequency.
- Raising the resonance frequency by partitioning the panel works, if the braces are stiff enough.
- The bell mode is hard to kill by external bracing, and typically need cross bracing, although turning the rim braces into 'C' beams helps considerably.
- Cross bracing can transmit vibration from one panel to another, and excite vibrations in the panels that were quiet. They are not perfect solutions and should be used with care.
- The larger the panels, the more difficult it seemed to raise the mode frequencies by bracing. My initial goal of moving all modes out of band were not met for the larger parts of the horn, probably for a large part because the braces weren't stiff enough.
But all in all, the horn is a lot deader than in its unbraced state, and although it's hard to quantify the amount of vibration contributed by the horn walls, I think it's probably not much compared to the direct radiation from the horn mouth, as there are no definite dips or peaks in the response that could come from spurious resonances or vibration.