A Study of DMLs as a Full Range Speaker

Founder of XSA-Labs
Joined 2012
Paid Member
Last edited:
Ok, getting back to "the study of dml's ..." on the subject of damping.

I did not realise that the water and glue treatment on EPS/XPS panels is actually a damping skin. I was under the impression it was a hard layer that improved high frequency propagation. Anyhow, the application of a thin latex film (Latex acrylic paint) seems very interesting since I have not read about any plywood treatment for damping purposes, other than varnish or paint for the looks. So, DMLBES's panels are damped along the edges (lightly clamped or sandwiched between foam strips), as well as a damping layer over the entire surface. It makes me think of OffGridKindaGuy's canvas framed panels - I am sure that the canvas and glue sort of dampens the plywood discs that he uses?

I am thinking of using Latex tubing along the panel edges i.s.o. foam strips for supporting and damping the panel, but still have to figure out how to glue everything together. I think the latex will provide more movement than the foam. In fact, I think one can use latex tubing all along the edges for damping, as there are various sizes and thicknesses available. Just cut a slit along the tube and slide it over the panel edge.

Still waiting for my exciters from Billionsound. The Wuhan virus situation is not helping things coming from China.

The glue water mixture is both used for damping as well as hardening. The reason why water is used is because glue will have to much damping if used without water to dilute it. Plus like I said before the percentage of he water glue mixture ratio is not set in stone you can change the ratio to your preference. For me I prefer 70% water and 30% glue so its less damped. With these panels its easy to over do it and under do it everything has to be JUST RIGHT.

There are basically two types of paint...Oil based and Water based.

For EPS I use Oil based high gloss enamel. Oil based paint is harder more durable then acrylic paint. Acrylic paint is softer and not as hard or durable and provides a little more damping. Why do different panels need different types of paint? Once again its all about BALANCE. If a panel is too stiff/hard its not ideal....If a panel is too soft its not ideal.....there needs to be a BALANCE as the panel cant be too hard yet it cant be to soft it needs to be just right.

Even after all that its still personal preference as some will like to over damp there panels as they prefer a darker sound and some will use less damping because they like a brighter more lively sound. One of the great things about these DML panels is you get to CUSTOMIZE it to your own taste.

If the latex tubing is hollow I wouldn't use it as the support sandwich as it will not be sturdy enough. If its not hollow then go for it.

Oh no make sure you sterilize the exciters and many factory workers in china touch that exciter lol Coronavirus comes animals. Dogs are the closest to humans. Chinese eat dog as delicacy but they don't want you to know about it. Virus jump from dog to human. In America most dogs are vaccinated with the coronavirus vaccine for dogs. In china most dogs are not vaccinated. I don't buy that it came from that ant eater looking creature. LOL
 
Also I wanted to add that if you do over damp your panels with to much paint or glue all you have to do is use a power sander with 220 grit sand paper and lightly sand the panels with 1-2 passes until you reach your desired damping goal.

Also one should always wear a mask and goggles when sanding as the fine dust can be detrimental to your lungs like the coronavirus. lol
 
I've been making REW measurements of my DML experiments for a few months now, and decided I should share some of my results.
The images below show how different suspension configurations affected the frequency response of my panels. In all cases, the panel was a 23" x 16" plywood panel (Revply underlayment, 0.2" thick) combined with the same exciter, at a location near (but not at) the center of the panel. In each case below, the panel was suspended by clamping it between two frames with EPDM foam strips in various arrangements on the two frames. Where there was no foam, the panel edge was unconstrained. Results for four cases below:

1) foam only in the corners
2) foam only along the short (16 inch) edges
3) foam only along the long (23 inch) edges
4) foam around the entire perimeter.

As you can see from the results below, generally the low frequency response improved as the foam suspension was added around more of the perimeter. The final image compares the extreme cases of (1) just the four corners and (4) the entire perimeter. Note the significant increase in low frequency output between about 60 and 160 Hz for the full surround case.

It was also interesting to note that the response above about 2000 Hz was virtually unaffected by the configuration of the foam surround.

Eric

4 corners by ekragness1, on Flickr

Short Edges by ekragness1, on Flickr

Long Edges by ekragness1, on Flickr

Full Perimeter by ekragness1, on Flickr

Corners vs. Full Perimeter by ekragness1, on Flickr
 
I've been making REW measurements of my DML experiments for a few months now, and decided I should share some of my results.
The images below show how different suspension configurations affected the frequency response of my panels. In all cases, the panel was a 23" x 16" plywood panel (Revply underlayment, 0.2" thick) combined with the same exciter, at a location near (but not at) the center of the panel. In each case below, the panel was suspended by clamping it between two frames with EPDM foam strips in various arrangements on the two frames. Where there was no foam, the panel edge was unconstrained. Results for four cases below:

1) foam only in the corners
2) foam only along the short (16 inch) edges
3) foam only along the long (23 inch) edges
4) foam around the entire perimeter.

As you can see from the results below, generally the low frequency response improved as the foam suspension was added around more of the perimeter. The final image compares the extreme cases of (1) just the four corners and (4) the entire perimeter. Note the significant increase in low frequency output between about 60 and 160 Hz for the full surround case.

It was also interesting to note that the response above about 2000 Hz was virtually unaffected by the configuration of the foam surround.

Eric

4 corners by ekragness1, on Flickr

Short Edges by ekragness1, on Flickr

Long Edges by ekragness1, on Flickr

Full Perimeter by ekragness1, on Flickr

Corners vs. Full Perimeter by ekragness1, on Flickr

Those measurements just confirm what I've been saying the entire time. When foam is adding all around the perimeter like my sandwich technique all the bass stays on the panel instead of escaping to the sides. I didn't need any measurements because my ears confirmed it for me.

Hopefully with this more people will see the importance of damping.
 
@DMLBES - Thank you for all the info on damping! Yeah, not sure how to sterilize the exciters - don't want to dunk them in disinfectant or cook them in the oven, lol.

@Veleric - Thank you for sharing. Most interesting and great to see actual measurements confirming certain things.

@BurntCoil - After Veleric's measurements - are you going back to the Tall Stories? I still like those panels the most and maybe the next step is to try them in a full frame suspended with foam along all the edges? That is what I am planning once I have disinfected exciters one day. Unfortunately I cannot measure yet.
 
@twocents

Yes Veleric’s data is most helpful and I am having a think about how this applies to the tall panels. It suggests some mods to the approach on damping. I was experimenting with measuring the bass frequency range last night and as Veleric suggested in another post the bass gets lumpier the lower you go. Currently I would roll off the tall panels from 60Hz and use a sub for the last octave and a half as the panels are smooth down to 60Hz then have a dip in level. You can’t hear this listening to them but it becomes obvious when you measure it. Eric provides good data!
 
Last edited:
Very interesting result Eric. The bass looks relatively smooth as well. It’s interesting that the mid to high frequencies were largely unaffected as well. Thank you for sharing .

You're welcome. I'm going to try to share more data, but it takes effort!

I should add that these plots are all applying 1/6 octave smoothing. I thought that would show up in the legend but it didn't. I use different smoothing levels depending on the data. More smoothing tends to hide flaws, but less smoothing is just overwhelming and hard to interpret. I find 1/6, 1/12, and "psychoacoustic" to be closest to middle ground in that respect.

Eric
 
Strings?

Okay, here's another comparison. This time, hanging from two strings vs. full foam surround. This panel is a 16" x 23" "SurePly" plywood underlayment (similar but not identical to Revply), with two DA Ultra exciters.
Neither produces a very flat response. In the "Full Surround" case I don't care for the broad dip between 100 and 200 Hz. But it seems to me that the "Strings" version, with generally weaker response below 300 H, except for the sharp spike at 130 Hz, is more objectionable. Both plots are at 1/24 smoothing, to reveal how sharp the spike in the "strings" case is compared to those with the "full surround".

49587194188_3b643552fb_z.jpg


49587931082_0ab5277042_z.jpg


49587932667_22cc1ca0a1_z.jpg
 
@Veleric

Eric I think the frequency response curves look good.

This is the frequency response of the Bowers & Wilkins 802 D3 Diamond, a 2019 Class A component according to Stereophile magazine.

The Bowers & Wilkins 802 D3 Diamond retails at $22,000 the pair.

Bowers & Wilkins 802 D3 Diamond loudspeaker Measurements | Stereophile.com

I would want to do some eq on this set of responses and look at where they are pitching their 0dB reference in this design. Speakers further down their range look far worse than the results you are getting
 

Attachments

  • B&W 802 Diamond.png
    B&W 802 Diamond.png
    487.3 KB · Views: 486
Last edited:
I am thinking of using Latex tubing along the panel edges i.s.o. foam strips for supporting and damping the panel, but still have to figure out how to glue everything together. I think the latex will provide more movement than the foam. In fact, I think one can use latex tubing all along the edges for damping, as there are various sizes and thicknesses available. Just cut a slit along the tube and slide it over the panel edge.


As far as using a damping material around the edges, this youtube vid shows an XPS (foamular 150) panel that he then put hot water insulation tubing around the edges , in stages to determine its effect...bottomline is that he believes it works somewhat but not as well as he would like...look around 11:42 for his first attempt, later on he tries more and more of the insulation until he covered the entire outside edges.


YouTube
 
Hi everyone, first post here.

We are designing our living room with my wife and I came across DMLs, which I think would be a fantastic addition to our wall.
I have read pretty much all pages from this thread trying to learn more about DMLs. While those designs can be very creative from an engineers side, they lack basic research and modelling from a physicists side I think. If you try to search modelling information for subwoofer housings for example, you will find every detail that you want. But there are no tools or simulation software for speakers like these, except those which NXT might had created. So there will be a lot trial and error and rabbit holes for an engineer.

From what I have read here, I am thinking of making plywood tall thin rectangular speakers, and as a dampening material I think I will try the canvas design. I think that this design had better/lower bass response not because there was an extra kind of vibration to the speaker, but the canvas between the frame and the DML circle (mine will be rect) acted as a dampening material. Just a thought though.
 
Welcome LiliumJSN. I suspect you are correct about the canvas surrounds. I also considered starting with framed canvas panels (using rectangular plywood discs) but decided to try simple suspended plywood panels first as a reference. There are not many framed canvas examples out there yet, so keep us posted.
 
If you try to search modelling information for subwoofer housings for example, you will find every detail that you want.

True. I think that's because bass enclosures are easy to model. However, the bass modeling that matters isn't easy: modal bass behavior in small rooms.

Earl Geddes followed by Harman International developed tools and methods to improve bass performance in small rooms but they aren't as widely used by DIYers as simple bass enclosure software.

YouTube