DIY Flat Panel Speaker Love

[Gary.M]

I don't think the exciter location is supposed to be equidistant from two edges. Check the information on this again...

Wow that does look smooth. You've done some really great work there!

I tried messing around with REW 5.1, and I'm getting a lumpy mess.. even with 1/6 octave smoothing. It could be the panels; it could be the room which is narrow and doesn't have anything soft in it.. or is could be the way I'm trying to measure. I have the UMM6 and calibration file loaded in to REW, SPL test is around -40db .. there are a lot of parameters that I'm uncertain of and left default. Below is a pic of the 'RTA' screen following a measurement.

I started out with two 20x28.25x1 pink XPS panels, rounded edges, rounded corners, sanded skin, two coats of white glue/water on each side and using a single DAEX32U-4 Ultra placed 2/5 from top and side. It's hanging from duct tape off the edge of a shelf about 20" from the back wall. It sounds ~ok~, but the vocals are shouty and the bass has a pretty narrow sweet spot

[jp7189]

Re: DIY Flat Panel Speaker Love

To put it simply it seems like my panels lack a tweeter compared to my bertagnis. The bertagnis seem to have a higher frequency extension. Not sure if its the bertagnis exciter alone or because they use a passive crossover to one of the exciters making it appear to sound like it has a higher frequency extension or both. My next goal is to order the DAEX25TP and a passive crossover to see if that helps. I already solved my bass and noise panel issues, I just need to solve this high extension clarity issue to match them to my bertagni speakers then I will be satisfied.

UBS, I think you're on the right track with this idea. It's seems like multiple exciters tend to cancel out the high frequencies. It sounds like the right move to have only one exciter making the highs and keep the one in the middle to the mids/lows.

This video was posted earlier in the thread, but it bears repeating. I wonder what your panels would look like if you set them horizontally and spread some salt, sand, rice, etc on them to see what patterns popup for sine waves in the upper range.. and how it changes with 1 and 2 exciters, and with high/low pass filters. would be awesome to *see* how a second exciter interacts with the first.

[jp7189]

I don't think the exciter location is supposed to be equidistant from two edges. Check the information on this again...

maybe I misunderstood the placement guide.. seems most people say 3/5's .. but I figure 2/5 is the same when measured from the other side ... so in my panel of 20x28.25 I placed the exciter at 8x11.3 (2/5s of 20 is 8 and 2/5's of 28.25 is 11.3).. is this not the correct way to figure placement?

[Gary.M]

This is a useful post with everything explained...

http://www.audiocircle.com/index.php?topic=70541.msg1248765;topicseen#msg1248765

maybe I misunderstood the placement guide.. seems most people say 3/5's .. but I figure 2/5 is the same when measured from the other side ... so in my panel of 20x28.25 I placed the exciter at 8x11.3 (2/5s of 20 is 8 and 2/5's of 28.25 is 11.3).. is this not the correct way to figure placement?

[Unbiasedsound]

I hooked up a exciter to my 50 watt polk sub amp just to see how it sounded as a sub. Through this experiment it just reinforced the theory of having some sort of surround to hold the outer edges off the panel using a frame, the same way my bertagni speakers utilize this method.

[[email protected]]

I could only find one company that is currently selling an affordable nxt type panel speaker. Onsia.com sells these Art Panel speakers which are essentially a stretched canvas frame with a panel and two very cheap exciters mounted inside the frame. They also include a bluetooth mono amp to drive the speakers and they cannot be hooked to a nomarl audio system which rules out stereo pairs or use in a surround sound application. The real issue though is that they sound like crap. virtually no bass, no high end and no volume. The sound a bit like listening to an AM radio. They definitely have a novelty factor and the bluetooth connection is intriguing, but I am hoping to get a full refund.

[[email protected]]

They did end up giving me a refund. This product of theirs could actually be good if they spent an extra $10.00 on better exciters and panel material as I also noticed when I took it apart that they are using cheap foam core as the panel material, which is a good reason there is no high frequency response.

[rmeinke]

I have been following this thread along with many other articles and sites about NXT and DML speaker technology. I also happen to live about 10 minutes from Parts Express's warehouse which is a big bonus. It seems that panel material is everything and after trying numerous materials discussed here and on other sites, and thanks to some friends at Wright Patt, I have been able to find a material 3mm thick, but lighter and much stiffer than gatorboard, xps, or basic foam core. With this I have been able to make speakers as small as 11x14 but that still provide good volume, and this panel material has the high frequency response that the foam core boards did not have. Even though they do sound pretty good for full range speakers, (especially considering the material cost for each is about $20.00 including the frame), they don't sound like the $1200.00 Boston Acoustics towers sitting next to them in the pic, but with a subwoofer they sound great as full range speakers, and as the rear channels of a home theater system they have real wife appeal.

These look fantastic jmanes!! Would the high-tech panel material you used to be similar to a more common carbon fiber panel or is it a honeycomb material?

Also, I too have a thin foam core board to be used for a wall mounted panel. I've treated it with the typical PVA:water treatment and the skin has much improved qualities over the untreated and heavy kraft paper skin. I'm surprised by the improvement but have to find a way to prevent the panel from warping following treatment. I've had so little free time lately I've just not had a chance to see how we can make this inexpensive foam core board "straight" following treatment. Seems to have potential and given the availability, it is something worth troubleshooting. Anyone want to step up?

Keep us posted as you progress!!! Looking good man...

[rmeinke]

Yes, they are prints glued directly to the panels using Scotch Photo Mount.

Curious jmanes, not being familiar with Scotch Photo Mount, does it dry "hard" or does it remain pliable?

[[email protected]]

The panel is a solid material, not a paper backed core type board. i tried the foam core with the various treatments, but I think it is the foam core itself that is absorbing some high frequency response. You may have seen the post I just left concerning the Onsia speakers, they used a foam core panel and their high end was so bad that I literally could not hear cymbals and acoustic guitar was almost inaudible. I have never tried to remove any of the pics I have glued down so I am not sure it the spray glue dries hard or not.

[rmeinke]

Hey everybody!

Read Rich's upbeat article on DMLs in a recent PE magazine, and spent the last few weeks working my way through this thread. It's like I've been on a fast forward ride through the last few years with you all.

I especially love the art vs. science undercurrent that's represented here, and totally dig both sides of it. I love the points Rich makes about experiencing the sound, and I totally feel Kornbread's scientific measurement approach. Personally, I'm more tone deaf than golden eared, which is great 'cause everything sounds good to me, and which is bad cause.. ya know.

Hate to admit it, but another IT guy here, and being such, am a data driven dude. I'm already leaning towards the think, build, and measure.. rinse and repeat methodology. To that end, I picked up a UMM6 mic and have been trying to figure out REW.. any pointers to getting usual, repeatable measurements would be much appreciated.

I have to say I've had tons of ideas while reading through this thread.. "I've got an idea" was usually followed by someone trying just what I was thinking... sometimes with good results and sometimes with bad.. but always I learned a lot. I could totally feel the disappointment in some of the posts, and to that I would say that it's all been a real inspiration to me! ..and thank you for sharing.

Hey jp7189... welcome to the flat panel thread my fellow IT guy!

I think I have to clarify the "art" vs. "scientific" approaches.

As an IT guy, I am typically and solidly in the scientific approach to assessing nearly everything I involve myself with. Have to say I do own iPhone based high-resolution RTA, OmniMic and minDSP/REW setups that have guided me, at various points, through the discovery phase with these DML's panels. My personal use of measurements has been to troubleshoot and general learning but more so to EQ the finished panels. Think its been pretty well known on the various forums that I prefer to EQ the panels to get a more balanced finished panel... considering both distortion and freq.resp. for improved sound quality.

But the "art" side of creating DML's is arguably more important IMHO, especially given our current state and the reason for some past posts here on PETT. Put simply...a freq. resp. graph will NOT tell us how good the speaker will sound overall. Case in point, my treated cardboard panels are the smoothest measured panel I have and even a simple passive crossover could have been used to create a quite flat panel (with a somewhat rolled off HF response notwithstanding). Measured distortion was also good. So if we are concerned about only measured response, we have succeeded in producing a smooth panel and all is well. In this case, the measured response would be terribly misleading as the panels measured well, but did not sound good - this particular set of panels were over damped and lacking in transparency compared to any of the normal poly panels. Bottom line, my point was why measure something we feel does not sound good to our ears. Use to troubleshoot, to learn, to make improvements... but a measurement by itself is not very useful if sound quality is not considered... am I right?!

Anyway... hope its a little clearer. Think I stated this as well... IMHO, we still have LOTS of materials, panel sizes/thicknesses and exciters, etc. to listen too before we can begin to settle on a smaller number of "best" sounding panel implementations... have many more conclusions to draw fellas!

At some point it would be very useful to summarize what we have collectively learned to date... it is all scattered though years of posts with many twists and turns... think that is what our good friend Kornbread was searching for in measurements and answers to all that has been experienced over the last 8-9 years. Is this right Kornbread?

Later bros...

[rmeinke]

Wow that does look smooth. You've done some really great work there!

I tried messing around with REW 5.1, and I'm getting a lumpy mess.. even with 1/6 octave smoothing. It could be the panels; it could be the room which is narrow and doesn't have anything soft in it.. or is could be the way I'm trying to measure. I have the UMM6 and calibration file loaded in to REW, SPL test is around -40db .. there are a lot of parameters that I'm uncertain of and left default. Below is a pic of the 'RTA' screen following a measurement.

I started out with two 20x28.25x1 pink XPS panels, rounded edges, rounded corners, sanded skin, two coats of white glue/water on each side and using a single DAEX32U-4 Ultra placed 2/5 from top and side. It's hanging from duct tape off the edge of a shelf about 20" from the back wall. It sounds ~ok~, but the vocals are shouty and the bass has a pretty narrow sweet spot

Hey jp7189! The smooth pics above are mainly due to the very low resolution of the RTS used (full octave). The measurements you took are pretty normal for DML... possibly a bit more uneven than normal but DMLs *are* not smooth. The roll off of high freq. is a little earlier than what I measure but they do begin to roll off around 10K with the larger exciters. The smaller 25mm voice coil exciters roll off closer to 12K due to the smaller VC.

I've never listened to panels completely free mounted like yours, but have to think there are going to be some panel resonances that need damping. Try them on an open back chair resting the panels a towel and the top of the panels leaning against some other damping material. They can't touch any hard surface or you will have many nasty's. Not even the wires can touch anything hard.

For your REW measurements, I wonder if you are measuring with your laptops built-in mic? Sometimes the settings revert back to the default settings and do not use the external mic. I've had various problems to with mic/soundcard settings reverting back to defaults at some point or another.

[rmeinke]

maybe I misunderstood the placement guide.. seems most people say 3/5's .. but I figure 2/5 is the same when measured from the other side ... so in my panel of 20x28.25 I placed the exciter at 8x11.3 (2/5s of 20 is 8 and 2/5's of 28.25 is 11.3).. is this not the correct way to figure placement?

Your placement is perfectly fine JP. Monacor place for exciter 1 is very close to the PE ratio... the difference would not be audible IMHO.

[Unbiasedsound]

I found another vid of some Bertagni speakers, these I think are the top of there line with the piezo electric tweeter in it. https://www.youtube.com/watch?v=nShwW6T8Q30.......Also I am going to order the DAEX-25TP to see if the highs are better then the exciters I am currently using.....If all else fails I will just buy another pair of Bertangni speakers lol

[Crapflinger]

These look fantastic jmanes!! Would the high-tech panel material you used to be similar to a more common carbon fiber panel or is it a honeycomb material?

Also, I too have a thin foam core board to be used for a wall mounted panel. I've treated it with the typical PVA:water treatment and the skin has much improved qualities over the untreated and heavy kraft paper skin. I'm surprised by the improvement but have to find a way to prevent the panel from warping following treatment. I've had so little free time lately I've just not had a chance to see how we can make this inexpensive foam core board "straight" following treatment. Seems to have potential and given the availability, it is something worth troubleshooting. Anyone want to step up?

Keep us posted as you progress!!! Looking good man...

I've done a lot of craft work with foam core and....glue warp is a yuge problem. it's caused by the paper absorbing the glue while the foam doesn't, so as it dries it contracts the paper with it. it does it with PVA, paint, spray glue...just about anything that wets the paper. rubber cement seems to be a bit less of a bear as it doesn't soak in as much.

I've had SOME success in the past with affixing the "back side" (i.e. whatever side you're not gluing at the time) to a bit of a frame made out of wood or something stiff with double sided tape. also, if it's even remotely possible (pro-tip it usually isn't) you can mitigate a lot of the warping by painting the glue on both sides at the same time in exactly the same way (the one time I did this was for something specific project. had to suspend the panel from a stand I made out of bent up coat hangers and me and a friend painted each side at the same time, in the same place...like I brushed on the right side of my side and the friend brushed on the left side of her side....it was a MASSIVE PITA).

maybe if the side were treated with a very very thin coat of primer of some kind, it would prevent the glue from absorbing as much?


also, fellow IT guy