Measurements attached for rubberized undercoating on an XPS panel. This worked quite well and will need a lot more testing to find how much and on which kind of panel yields the best synergy. I tested a couple other panels too and this brought very nice improvements there as well. For panels that are too wild and uncontrolled, this coating helps tame them.
I used an $8 can of Dupli-Color Rubberized Undercoat for 3 panels and I ran out, so I ordered some 3M and Rustoleum equivalents from Amazon so I can continue testing and comparing products. A medium-weight coat was very good and the second heavy coat was better in some ways, but not in all ways. I'm thinking something between medium and very-heavy weight coating will be best. The coating will smudge if pressed too hard, and any tape stuck to it will pull some off when removed. I put a light layer of acrylic lacquer on at the end and this makes the coating feel much more durable. It may have given a slight reduction of performance but it is probably worth it. I might try diluted PVA instead of lacquer next time, or just skip it. This lacquered undercoat feels similar to pleather.
Both of my 'temporary' panels in my living room that I use with TV are now rubberized and lacquered. This is definitely a nice step up from how this system sounded last week. The panels are a deep black color that looks much nicer than purple too.
Unfortunately I forgot to weigh all the panels I coated before each coating, but the one I attached measurements of I got the weights after each coat. No starting weight though, oops. Also, for each test the driver and panel had to be remounted. The frame and mic was always in the same place but any time you remount a driver and/or panel in a test rig like mine there is certainly going to be a difference in results because of alignment of the parts. Even just a 1/4" ( maybe 5mm) difference in driver location can cause significant variances in FR and distortion, which then cascade into impulse and GD and others. I think I did a pretty good job at keeping the alignment very close, but keep in mind that it is not perfect.
The measurements that I think show the biggest impact are the Early Decay Time on Filtered IR tab, along with all those other values. Group Delay also shows a big change before and after the first coat. After coating the distortion is smoothed out some and is generally lower.
Listening to the coated panels does sound better - they are a bit sharper and more dynamic. It was significant enough that I noticed it on the subjective listening test of the first coat before using REW. It reminds me of OLED where the blacks are so black. The panel is also literally extremely black
I used an $8 can of Dupli-Color Rubberized Undercoat for 3 panels and I ran out, so I ordered some 3M and Rustoleum equivalents from Amazon so I can continue testing and comparing products. A medium-weight coat was very good and the second heavy coat was better in some ways, but not in all ways. I'm thinking something between medium and very-heavy weight coating will be best. The coating will smudge if pressed too hard, and any tape stuck to it will pull some off when removed. I put a light layer of acrylic lacquer on at the end and this makes the coating feel much more durable. It may have given a slight reduction of performance but it is probably worth it. I might try diluted PVA instead of lacquer next time, or just skip it. This lacquered undercoat feels similar to pleather.
Both of my 'temporary' panels in my living room that I use with TV are now rubberized and lacquered. This is definitely a nice step up from how this system sounded last week. The panels are a deep black color that looks much nicer than purple too.
Unfortunately I forgot to weigh all the panels I coated before each coating, but the one I attached measurements of I got the weights after each coat. No starting weight though, oops. Also, for each test the driver and panel had to be remounted. The frame and mic was always in the same place but any time you remount a driver and/or panel in a test rig like mine there is certainly going to be a difference in results because of alignment of the parts. Even just a 1/4" ( maybe 5mm) difference in driver location can cause significant variances in FR and distortion, which then cascade into impulse and GD and others. I think I did a pretty good job at keeping the alignment very close, but keep in mind that it is not perfect.
The measurements that I think show the biggest impact are the Early Decay Time on Filtered IR tab, along with all those other values. Group Delay also shows a big change before and after the first coat. After coating the distortion is smoothed out some and is generally lower.
Listening to the coated panels does sound better - they are a bit sharper and more dynamic. It was significant enough that I noticed it on the subjective listening test of the first coat before using REW. It reminds me of OLED where the blacks are so black. The panel is also literally extremely black
Moray James.
https://www.amazon.co.uk/dp/B0C919M..._csd=d2lkZ2V0TmFtZT1zcF9ocXBfcGhvbmVfc2hhcmVk
I was looking on amazon and noticed some veneer products, which seemed reasonably priced and a good size.
This seems to be about the maximum width.
This could be interesting for a future panel.
Steve.
https://www.amazon.co.uk/dp/B0C919M..._csd=d2lkZ2V0TmFtZT1zcF9ocXBfcGhvbmVfc2hhcmVk
I was looking on amazon and noticed some veneer products, which seemed reasonably priced and a good size.
This seems to be about the maximum width.
This could be interesting for a future panel.
Steve.
Hi Eric
Yep, embarrassingly, I misinterpreted your symmetry reference, and of course you are on the money given the disparity between transverse and longitudinal shrinkage/swelling.
Additionally, these panels are in test mode and have yet to be coated in any way, AND, I now live in SE Qld which is quite humid in summer ( not the case ATM). Quite the test bed I have set up for this material eh ? 🤭
Nevertheless, I remain optimistic. If it was poplar I wouldn't be so cavalier, but this material is like chalk and cheese in comparison, and the veneer thicknesses are only 0.5mm each, half that of the 3 mil 3 ply poplar.
In the end, on their hoped for survival, I'll probably finish them with clear auto acrylic spray which is a good lightweight and moisture proof coating.
👌🤞🤞
PS .. Did you look for/find any self adhesive felt to test as a comparative mounting material??
Eucy
Over the last year I’ve been using basswood ply in 3/32” thickness from Wicks Aircraft Supply, it’s the best material for me so far (it’s only a short drive for me, luckily!). this is a 3-ply that is probably similar to a thicker veneer, & very light. I tend to agree with Spedge’s assessment of card stock/thick paper applications. But I also “dope” the wood with multiple stains of diluted graphite powder w/ a bit of wood glue, which stiffens it quite a bit without adding much weight. I’ve also mixed in various acrylic paints & copper powder. I have not gotten around to recording & measuring yet, but I am very happy with even my 12”x12” panels. They get quite loud & reproduce the piano music I play through them very well. Indeed, they are much like piano soundboards imo.
I am going larger soon (love my 24”x24” panels), but the ply is a bit expensive admittedly, though not prohibitively so. The 3/32” ply is also the exact thickness of a saw blade, so I simply run a slit into a chunk of hardwood for bases (heavy, like ironwood). I create more trapezoidal bases so that they are slotted in more at a point, which keeps the panel as free as possible (cantilever?). Haven’t been using braces or frames & been quite pleased. Going more in direction of rounded panels, perhaps a Genelec-ish shape, as well as odd number polygons (eg 11-sided), out of curiosity about Manger diaphragm shapes.
I am going larger soon (love my 24”x24” panels), but the ply is a bit expensive admittedly, though not prohibitively so. The 3/32” ply is also the exact thickness of a saw blade, so I simply run a slit into a chunk of hardwood for bases (heavy, like ironwood). I create more trapezoidal bases so that they are slotted in more at a point, which keeps the panel as free as possible (cantilever?). Haven’t been using braces or frames & been quite pleased. Going more in direction of rounded panels, perhaps a Genelec-ish shape, as well as odd number polygons (eg 11-sided), out of curiosity about Manger diaphragm shapes.
Attachments
The only way to guarantee a solid centre image is to ensure you're getting identical frequency responses from both panels at your listening position.
Check your spectrograms too for symmetry, as well as group delay.
Let's hope for the best!
My wife was at the craft store today so I asked her to look for it but alas they didn't have it.
Eric
SS,
What were you using for the clamping material this time?
If you do any more PVA coating of fresh XPS panels, I'd be interested to see close mic measurements before and after PVA coating.
While I often look at the IR graphs, I actually never looked at the numbers before. But I did look at the EDT numbers and didn't see much change (in the number) with the various coatings, even though it did look like the IR decayed faster with the rubber coating (to my eye). What EDT numbers did you find, maybe I'm not looking at them right.
I also never really look at group delay, as I don't know what I'm really looking for. What would good look like?
Eric
"Center image" is weird to describe. I think maybe I mean sound stage more than center image, not sure. It's probably more timing related than anything else. With my box speakers the center sounds very precise and focused, but with these panels it sounds larger and more diffuse. I've been trying to think of some numbers I could measure to test this more empirically but I don't know how to go about doing that. Getting equal distance and FR and spectrogram from L and R is pretty obvious, but there's a whole 'nuther dimension to it. I think maybe it has a lot to do with reflections and the time arrival of them all, combined with their loudness.
@Veleric I used the same frame as before but with butyl removed and some weatherstripping running full top and bottom (short sides). Clamped quite firmly.
Regarding Impulse Response, I hardly know anything about this so someone please correct anything wrong with what I'm saying, but here's my interpretation:
Looking at the IR graphs can tell a lot just by comparing amplitudes from two tests across the timeline. When comparing you need both sources to be playing at the same dB level though because a quieter one will naturally respond faster. Generally the faster the onset and decay the better. On the Filtered IR tab there's some values calculated in a little window that are useful for comparing:
We can see that the EDT is quite a bit lower with the rubberized coating, and clarity is better, and all the other impulse related metrics are quite a bit faster. This is taken with very close mic so the room has almost no effect here, hopefully. When comparing these at 3ft / 1m the numbers are much different which I believe is because of the room, though it could be more of the panel registering on the mic rather than the room, I'm not sure.
When looking at the impulse graph we can see how the (blue) rubber ramps up in amplitude faster than the (red) rubber-free panel, and then it decays much faster too:
We can see how with the red one much more of the sound is left behind after the amplitude spike and it kinda keeps ringing for a while. That's not good. This red one is much better than a lot of other panels I've made, but this rubber coating really steps it up here with the damping. The faster the impulse response the better the dynamics of the speaker and the clearer it sounds. This is one of my main goals when working to improve panel response. Good off-axis response is the other main goal, which DML seems to deliver without any extra attention needed.
I might be wrong with my understanding of this IR stuff but this graph above aligns well with what I hear subjectively. There is definitely an improvement in clarity and dynamics when this rubber coating is applied to this panel.
Very interesting treatment of your panels and pretty unique. I assume you have a very fine ear for testing your panels, listening to piano music and probably being a pianist yourself.
Could you explain in what way this treatment enhances / improves the sound of your panels, apart from just stiffening them?
I tried adding very fine graphite powder to epoxy once and the epoxy never cured hard. I never retried it so maybe I just didn't add enough hardener or something, but that was disappointing. I also tested a couple wood glues (on paper samples instead of a panel) and decided that plain white PVA was better. Same with mod podge matte and glossy - plain white PVA seemed better to me on paper samples. I never tested them on panels though.
I started preparing for the next round of rubber undercoating tests tonight and the start was a plain 'virgin' XPS panel. This one is about 0.6" thick just because I have a big sheet of it at that thickness from previous cuts. I put it in my frame and attached a newly unboxed XT32-4 driver and tested it very close and at 3ft. Then repeated with heavy PVA'd XPS with 2x rubber coat:
The yellow virgin panel is 4dB louder in the 3ft measurement, on average, but still - the blue heavy rubberized panel has much faster decay. It does have slower rise time though. In the close mic test the FR of virgin panel is more flat but at 3ft the rubberized panel is significantly less volatile. Distortion is far higher for the virgin panel.
My next test is going to be rubber coating directly on the virgin panel, medium-light coat, and no lacquer. I think anything hard on the top of the rubber counteracts its purpose.
I got some 3M and Rustoleum rubberized undercoating today and compared to Dupli-Color, and all three are different. 3M goes on the smoothest, Rustoleum had tons of little bubbles that formed after spraying and then they popped leaving a thick mesh behind with lots of holes in it. The Dupli-Color was right in between them but tended towards stringing with tiny bubbles, not bigger ones. I'll shake them more tomorrow and retry to see if their patterns are consistent. I'll start with a 3M test because I think it looks the best of all 3.
The yellow virgin panel is 4dB louder in the 3ft measurement, on average, but still - the blue heavy rubberized panel has much faster decay. It does have slower rise time though. In the close mic test the FR of virgin panel is more flat but at 3ft the rubberized panel is significantly less volatile. Distortion is far higher for the virgin panel.
My next test is going to be rubber coating directly on the virgin panel, medium-light coat, and no lacquer. I think anything hard on the top of the rubber counteracts its purpose.
I got some 3M and Rustoleum rubberized undercoating today and compared to Dupli-Color, and all three are different. 3M goes on the smoothest, Rustoleum had tons of little bubbles that formed after spraying and then they popped leaving a thick mesh behind with lots of holes in it. The Dupli-Color was right in between them but tended towards stringing with tiny bubbles, not bigger ones. I'll shake them more tomorrow and retry to see if their patterns are consistent. I'll start with a 3M test because I think it looks the best of all 3.
I found a vast improvement in the imaging and soundstage by semi closing the back with a panel made of fabric and insulation (this dropped the rear output from approximately 800Hz up by 10+db). Before in my lounge there was too much rear sound bouncing off the walls the sound was very diffused and un centred.
OK – so after 600+ pages on this thread, I still have fundamental questions, the answers to which may be out there so here goes:
1/ Are we able to establish any type of rules/guidelines for the minimum size of a panel of a given material to allow full development of mid range frequencies (say from 150Hz), accepting that above coincidence, the importance of panel size reduces and hf is highly governed by panel material and the driver.
2/ For a given exciter and material, is there information on output vs size for a range of frequencies..eg say for 500Hz, 1000, etc. ie: what is the response curve for a typical exciter as the panel size varies.
3/ Panel Material: We have Christian’s work on panel material selection, and I have been prompted to crawl through this again as a result of my most recent Blackwood panel.
I first tested this material yonks ago at 2.5mm thick against an identically sized 3mm Poplar panel, and although the response was more even that the Poplar, the output level was depressed compared to the Poplar, so I put it away. Having reduced the thickness, I find that the 2mm Blackwood panel using a DAEX30HESF-4 now equals the same size 3mm poplar panel (DAEX25SHF-4) in output level with an exciter rated 10-15dB lower across the whole frequency range.
Compare now the Blackwood panel (900x300) with a 1200x400x4mm Cedar panel driven by a DAEX25SHF-4, and it exceeds it at moderate amp levels, with the Cedar only catching up as the power is increased.
All this despite the following theoretical figures, which are after all, aimed at efficiency and output levels:
Note that the 2mm Blackwood panel is theoretically a very poor candidate material (Refer to Christian’s charts for details) (PS: Christian – I can’t follow why figures 3 and 4 show fc <5kHz target when Section 7.1 states targets of 6500Hz < fc < 6500Hz)
With respect to the Blackwood/Cedar comparison, I can understand the results from an intuitive standpoint, but the theoretical performance figures of the 4mm Cedar far exceed those of the 2mm Blackwood. Hence my question 2.
Also note the Acrylic data in comparison to the reportedly strong high frequency performance gained from this material. (I don't know what the rest of the range is like with material, but there a a number of tests and written reports on the hf end.)
In summary, unless I’ve messed something up here, it seems to me that the theoretical formulae are not a strong guide to actual material results.
I've a very strong suspicion that one material characteristic which is absent from the plate theory plays a big part in explaining at least some of the apparent discrepancies, and that is hardness !
Feedback would be appreciated... please put me out of my misery, this stuff is driving me nuts! .
Cheers
Eucy
1/ Are we able to establish any type of rules/guidelines for the minimum size of a panel of a given material to allow full development of mid range frequencies (say from 150Hz), accepting that above coincidence, the importance of panel size reduces and hf is highly governed by panel material and the driver.
2/ For a given exciter and material, is there information on output vs size for a range of frequencies..eg say for 500Hz, 1000, etc. ie: what is the response curve for a typical exciter as the panel size varies.
3/ Panel Material: We have Christian’s work on panel material selection, and I have been prompted to crawl through this again as a result of my most recent Blackwood panel.
I first tested this material yonks ago at 2.5mm thick against an identically sized 3mm Poplar panel, and although the response was more even that the Poplar, the output level was depressed compared to the Poplar, so I put it away. Having reduced the thickness, I find that the 2mm Blackwood panel using a DAEX30HESF-4 now equals the same size 3mm poplar panel (DAEX25SHF-4) in output level with an exciter rated 10-15dB lower across the whole frequency range.
Compare now the Blackwood panel (900x300) with a 1200x400x4mm Cedar panel driven by a DAEX25SHF-4, and it exceeds it at moderate amp levels, with the Cedar only catching up as the power is increased.
All this despite the following theoretical figures, which are after all, aimed at efficiency and output levels:
| ρ (kg/m3) | E (Gpa) | h (mm) | μ | B | fc | Zm | fhigh | T=B/μ3 | 'B/μ | A.f0 | Eff (dB) |
| | | | | | | | | | | | |
Poplar | 440 | 8.9 | 3 | 1.32 | 20.025 | 4835.46 | 41.13 | | 8.71 | 15.17 | 12.24 | 87.7 |
Blackwood | 640 | 14.82 | 2 | 1.28 | 9.88 | 6778.97 | 28.45 | | 4.71 | 7.72 | 8.73 | 86.4 |
NXT | 550 | | | 0.55 | 6.6 | 5436.84 | 15.24 | | 39.67 | 12.00 | 10.88 | 91.0 |
Cardboard | 140 | 1 | 3 | 0.42 | 2.25 | 8137.11 | 7.78 | | 30.37 | 5.36 | 7.27 | 90.4 |
Cedar 4mm | 370 | 7.66 | 4 | 1.48 | 40.85 | 3584.71 | 62.21 | | 12.60 | 27.60 | 16.51 | 88.5 |
Acrylic | | | | 3.8 | 8.3 | 12743.53 | 44.93 | | 0.15 | 2.18 | 4.64 | 78.9 |
Note that the 2mm Blackwood panel is theoretically a very poor candidate material (Refer to Christian’s charts for details) (PS: Christian – I can’t follow why figures 3 and 4 show fc <5kHz target when Section 7.1 states targets of 6500Hz < fc < 6500Hz)
With respect to the Blackwood/Cedar comparison, I can understand the results from an intuitive standpoint, but the theoretical performance figures of the 4mm Cedar far exceed those of the 2mm Blackwood. Hence my question 2.
Also note the Acrylic data in comparison to the reportedly strong high frequency performance gained from this material. (I don't know what the rest of the range is like with material, but there a a number of tests and written reports on the hf end.)
In summary, unless I’ve messed something up here, it seems to me that the theoretical formulae are not a strong guide to actual material results.
I've a very strong suspicion that one material characteristic which is absent from the plate theory plays a big part in explaining at least some of the apparent discrepancies, and that is hardness !
Feedback would be appreciated... please put me out of my misery, this stuff is driving me nuts!
.Cheers
Eucy
@JoskaNZ - Yeah I'll have to experiment with insulation more in the future. That was my original plan last year but in my most recent tests of hanging heavy jackets behind panels that gave it a much more 'boxed' sound. At the time I strongly preferred the open sound but I'll revisit again later. There's a lot of talk out there about absorption and reflection for room treatment and I think that might bring big improvement without losing the openness. Not sure, and idk how to measure it.
Out of interest, the 3M product contains asphalt (I suspect they mean bitumen as asphalt is a bitumen/aggregate mix (in this case maybe talc)), and the Rust-Oleum comes in clear, white and black with lots of nasty solvents (3M also), with the colours being attained by adding titanium white and carbon black respectively. Duplicolor also contains carbon black. So they're not the same products as the 3M item, probably explaining the difference when applying them.
I must admit to being surprised that this doesn't absolutely kill the panel response, given how we've been concentrating on minimising coating thicknesses - I happen to have some of the Rust-Oleum product, so I am tempted to at least try it (on some scrap first). If it pans out it could be useful on the woofer project .
Eucy
Last edited:
Oh no, I just checked on my 3M and Rustoleum samples and they both dissolved some XPS. It happened pretty slowly. Usually spray-on things that are going to react will eat it pretty fast, but these were both pretty slow. I've watched a bunch of things eat into XPS so I know this rate was far slower than normal. The reaction made the Rustoleum kinda hard, while the 3M is still soft-ish. I sprayed new samples on paper and will check tomorrow. Maybe some PVA or shellac will be needed to protect the foam.
I guess that reaction explains the holes in the Rustoleum product. It was reacting with XPS, it just didn't eat down very far while I was watching.
Dupli-Color Rubberized Undercoating definitely did not react on bare XPS though, it is safe.
I guess that reaction explains the holes in the Rustoleum product. It was reacting with XPS, it just didn't eat down very far while I was watching.
Dupli-Color Rubberized Undercoating definitely did not react on bare XPS though, it is safe.
Hmm, yes, I see these 3 below which would be great to test but is pretty expensive for just a small test.
Second Skin Spectrum $100 for a gallon
Lizard Skins Sound Barrier $120 for a gallon
Silent Running SR1000 $50 for a quart plus shipping, $155 for a gallon plus shipping
If someone wants to donate to the cause I can test them
My paper samples for 3M and Rustoleum show that 3M does apply more smoothly and is significantly more dense than Rustoleum and Dupli-Color.
I'm going to need to pick up another can of Dupli-Color I think so I can test it on a bare panel. I want to keep the weight down as much as possible so skipping the PVA preparation for 3M and Rustoleum would be good.
Second Skin Spectrum $100 for a gallon
Lizard Skins Sound Barrier $120 for a gallon
Silent Running SR1000 $50 for a quart plus shipping, $155 for a gallon plus shipping
If someone wants to donate to the cause I can test them
My paper samples for 3M and Rustoleum show that 3M does apply more smoothly and is significantly more dense than Rustoleum and Dupli-Color.
I'm going to need to pick up another can of Dupli-Color I think so I can test it on a bare panel. I want to keep the weight down as much as possible so skipping the PVA preparation for 3M and Rustoleum would be good.