The corners of the canvas are limited in movement by the wooden frame itself. The shaped Nidaplast's 'fingers' extend into those corners, and are glued down quite solidly. Therefore, the tips of those fingers are fixed points that don't move at all. The gradually-increasing width of those fingers means that the panel is progressively damped closer to the tips. If the panel was not mounted to the canvas, and was not mounted by its corners, then the corners would have to be rounded off to control the undamped vibration.
I used the canvas specifically as a high compliance "surround" to allow the panel to move freely, but yet not in an uncontrolled manner, in order to improve the bass response. In my Twinwall gig panels I had the panels clamped down all around, and this reduces the bass response significantly.
I'll do further testing and try to optimize the whole canvas setup to mimic a high-Qts open-baffle driver—a driver such as the Eminence Alpha-15a. If I can get the cheap Alpha-15a driver to deliver an almost flat response down to 45Hz in a completely open-baffle configuration, then there's no reason I can't get a panel, with a larger surface area, to do the same or better.
The only thing that I'd have to address is the inevitable suck-out at the first break-up mode.
do you mean that rather than the corners going to a long point that they wold have a small semicircular notch at the end which would result in two smaller points at each corner? I hope this makes sense.
The Nidaplast fingers are stuck into the corners horizontally between the wood and the canvas, and then glued solid. The tips of those fingers can be sharp or truncated or scalloped, it doesn't matter what the shape is because they're fixed points encased in glue.
Just a plain open baffle. No slots. No loading. No wings. A single 20mm thick plank just wide enough to accommodate the driver and high enough to screw a tweeter into.
Quote "To me driver placement does not seem essential to what modes can be excited. " please explain what you mean? the impact of the motor in whatever location you place it will have varying effects on the diaphragm which will vary with frequency as well as level and there will also be panel related effects which will also be variable. For every different driver location you have a different situation then adding multiple motors and it only becomes more and more complex. The motor and the panel combination is a veritable can of worms of back and forth interactions. It makes me dizzy to think about it.
Moray .
I just took a few pictures of some old test panels, which were lying around.
The first on the left is the back of my cascamite canvas panel with the enlarged crate ply attached.
I would have tried a larger crate ply but that was all I had.
Minimising the amount of loose canvas to stop flapping or buzzing.
I do not remember any problems with the corners?
The second picture is the front of this panel.
I had forgotten that I cut away the canvas from the panel to minimise the damping on the panel 🤗
Which was handy for the cling film , giving a mm or so separation .
3rd and 4th are of a crate ply with weights for this panel only.
Two points were used for mounting, plus some film for testing.
The last is an old 5watt (I think) exciter on a low density EPS panel.
Steve.
I just took a few pictures of some old test panels, which were lying around.
The first on the left is the back of my cascamite canvas panel with the enlarged crate ply attached.
I would have tried a larger crate ply but that was all I had.
Minimising the amount of loose canvas to stop flapping or buzzing.
I do not remember any problems with the corners?
The second picture is the front of this panel.
I had forgotten that I cut away the canvas from the panel to minimise the damping on the panel 🤗
Which was handy for the cling film , giving a mm or so separation .
3rd and 4th are of a crate ply with weights for this panel only.
Two points were used for mounting, plus some film for testing.
The last is an old 5watt (I think) exciter on a low density EPS panel.
Steve.
Attachments
Thanks I will have a look at these and give them some thought. I appreciate you posting them. do you remember how many card panels you made? I am still having a bit of a hard time adjusting to panel size and level with DML panels. one of your card panels was 6x9 inches so how much larger would a panel need to be to play appreciably louder than that one would (I guess I should ask hoow loud that one willl play with subs) assuming that the bass roll off of the larger panel was adjusted to match the smaller panel? I cannot seem to understand how small or how large I need to build. I would be happy with the ability to play average level of 95 db and peak capability of over 100 db db. I don't often play this loud but some music just calls for those levels to feel right. Is this expecting too much from a small panel? thanks Steve.
with a drill press you could lose a lot of panel weight and see what that sounds like?
Here's a pic of an open baffle Alpha15a similar to the one I built with my partner a while ago.
Attachments
I'm not saying it has no effect, but that it doesn't seem essential. I cannot claim to understand exactly how and why, I just found that moving around the exciters does have an effect on the FR, but it is much more marginal than I would have expected. The properties of the plate and how it is suspended seems to pretty much decide the FR, and then you can do minor adjustments with for example exciter placement and weights.
The driver will vibrate the whole panel and make sound regardless of being placed on a node or an antinode. I think this is why the placement of the exciter is more forgiving than the suspension of the panel.
If it made sound only from the resonance modes of the panel, the frequency response would look really bad.
If it made sound only from the resonance modes of the panel, the frequency response would look really bad.
It's true that DML panels make sounds at all driven frequencies, not just at the resonances. But they do tend to make sound more readily at resonances, and especially at some particular resonances that are especially good at radiating sound. And not only can they radiate really well (i.e. loudly) at those frequencies, they can also ring (i.e. resonate long after the originating signal is gone) at those frequencies. And for those reasons, sometimes the frequency response and spectrograms of DML's truly do look really bad! This is especially true at low frequencies, where modal density (resonances/octave) is low.
In my experience, there are basically two ways deal with this basic issue. One is simply to use a DML only above the frequency at which all the lumpiness and ringing disappears. Depending on your choices, this could be in the range of 300 to 1000 Hz.
The second way is to use good design principles to (a) distribute the natural frequencies regularly based on the aspect ratio, suspension points and elastic moduli of the panel (b) judiciously position the exciter to equally excite those resonances, and (c) add sufficient damping to spread/widen the resonance frequencies from individual sharp spikes, to broad humps that overlap. With this method, it's easily possible to obtain good frequency response down to 100 Hz and perhaps lower.
Eric
Sapphire,
Wow, big LF bump with the baffles. Nice to see data! Can you share any pics of the panels and baffles? I've never tried baffles but have always wondered about whether simply making the panel bigger (self baffling) would achieve the same or better result than adding the baffle. Any thoughts?
Eric
Hey Moray,
I could be wrong but I'm not sure that you can really do much with localized application of CLD. Frankly, I'm skeptical of any treatments attempting to address individual modes, and even more skeptical of treatments for individual antinodes. I prefer a more wholistic approach like I described in an earlier post tonight.
One possible exception would be the use of local weights for the purpose more evenly distributing natural frequencies along the frequency spectrum. I have not tried this myself, but for me this would be similar to using shape, or aspect ratio, or elastic properties to tailor the distribution of natural frequencies. But it would take modeling capabilities (like LISA or better) to do decently.
Eric
I used some scrap chipboard to make an L that I mounted to the back of the frame. It's about 15 inches deep and runs a bit further than the length of those 2 sides. I used duct tape around the entire panel-to-frame gap for a seal but I don't think that worked very well so I'll have to figure something else out. The frame is made of 2x6 so it's already got at least 6" of baffle built in all the way around. This was my first test and I haven't had time yet to experiment with baffle sizes. Building a driver brace is a higher priority right now, as is my day job...
Bracing the two short ends of the panel made an improvement by itself but it was a bit lumpy. The baffles added more and smoothed it out.
I never really thought about baffles in the past because the panel itself looks like a baffle (self-baffling), but there's a lot of interaction between the front and back near the edges. The smaller the panel the more intense that is. You can feel it in the air with your hands near the surface and hear the weirdness with your ears if you inspect the edge of an unsecured panel while it is playing. I figure that baffles should dilute that front-back interaction resulting in higher efficiency and less distortion.
He is a man of mystery.
We seek him here we seek him there, we seek him every ****** where 🤣
https://poets.org/poem/macavity-mystery-cat
Steve.