DIY Flat Panel Speaker Love

[Kornbread]

Re: DIY Flat Panel Speaker Love

Placed one flat panel in a corner where the NHT's usually sit which is about 3' from the back wall and 4' from the side wall. The listening position was 10'5" from the front of panel. The listening position is in a huge bass void from around, IIRC, 25-50hz. This from past experience measuring the NHT's and eq'ing the 4 18" IB's in the same corner. The room is 16'x20' more or less as I have piles of stuff laying around in there.

Some surprises. Bass didn't seem to change much but the top end that was completely gone in the near field measurements is suddenly there??? The peaks and dips are very evident when slowly playing through the spectrum of sine-waves. This may be why the waterfall plots look as they do. Other than the obvious sawtooth response ... from about 200hz on up the response does look fairly reasonable.

[rmeinke]

Re: DIY Flat Panel Speaker Love

Hey Cornbread! This confirms what we thought might be going on with the high frequency extension with the solid panel mounting... there ought to be better extension as your measurements drops early and, more so, very quickly.

If you have time, grab some measurements with more smoothing (1/12 per octave would be best for DML's) as you can see the choppiness of a DML panel at 1/24 resolution... make these measurements a little more helpful without so may squiggles

[Kornbread]

Re: DIY Flat Panel Speaker Love

Measured from 1 meter in front of panel at about the midpoint of panel height with smoothing.



Notice the overlay of the nearfield and farfield measurements in post 437. The top end was there when measured at the listening position but not at the 1 meter position. Could it be the panel is beaming at high frequencies similar to other large panels? This could explain why the top is so very rolled off at close distance, the mike was in a treble null. The null fills in at distance given enough distance to merge. Could this also hold true for the rest of the frequency spectrum. Food for thought.

[rmeinke]

Re: DIY Flat Panel Speaker Love

Placed one flat panel in a corner where the NHT's usually sit which is about 3' from the back wall and 4' from the side wall. The listening position was 10'5" from the front of panel. The listening position is in a huge bass void from around, IIRC, 25-50hz. This from past experience measuring the NHT's and eq'ing the 4 18" IB's in the same corner. The room is 16'x20' more or less as I have piles of stuff laying around in there.

Some surprises. Bass didn't seem to change much but the top end that was completely gone in the near field measurements is suddenly there??? The peaks and dips are very evident when slowly playing through the spectrum of sine-waves. This may be why the waterfall plots look as they do. Other than the obvious sawtooth response ... from about 200hz on up the response does look fairly reasonable.

[ATTACH=CONFIG]61562[/ATTACH]

Wow... the HF response is very surprising actually At listening distances usually the HF trails off... I also assumed it was due to comb filtering but due to DML's random node generation across the panel, not sure how that impacts comb filtering.

3-4' away from walls is fine for bass. Put the panels a foot from the wall has really no impact on bass performance... they get squawky against a bare wall but that should be expected.

[rmeinke]

Re: DIY Flat Panel Speaker Love

Measured from 1 meter in front of panel at about the midpoint of panel height with smoothing.

[ATTACH=CONFIG]61573[/ATTACH]

Notice the overlay of the nearfield and farfield measurements in post 437. The top end was there when measured at the listening position but not at the 1 meter position. Could it be the panel is beaming at high frequencies similar to other large panels? This could explain why the top is so very rolled off at close distance, the mike was in a treble null. The null fills in at distance given enough distance to merge. Could this also hold true for the rest of the frequency spectrum. Food for thought.

No. beaming here, the opposite with DML. They are nearly omni-directional with only a very narrow band to the very side of the panel. Just pull them out into the room and walk between them... its amazing... the sound stage basically stays intact save for that narrow band.

Its very odd, your measurement are the opposite of what I get with both Omnimic and an RTA on my phone with a calibrated Dayton Audio iMM-6 mic. Other measurements over at AC also look like the measurements I get and are rolled off at the listening position including those taken with REW... really can't explain why you are getting the measured responses that you are.

Its been my experience that close miced measurements are closer to the response or balance you get at the listening position. This again is due to the near omni-directional output of the panels. The power response graphs that I have seen are tilted up about 6db(???) by 20Khz if I remember right. When EQed perfectly flat the balance is indeed very tilted up and are strident, thin and unlistenable.

[rmeinke]

Re: DIY Flat Panel Speaker Love

Ahhhh, heck yeah there is beaming... :-( Sorry...

After a quick test, directly in front of the exciter:


At the top right corner 4-5 inches into the panel:


I knew this sadly... when the length of the sound wave gets smaller than the voice coil the panel and exciter exhibits an oil can effect and then has higher output directly in front of the exciter. This seems to contradict the power response graphs but have to think more on this and how the on and off axis might sum... still hard to believe that the power response would be elevated all the way out to 20Khz now that I think this over more...

Was thrown off by your measurements and still can't explain it. What mic are you measuring with? I don't know much about different mics but wonder if that might be impacting the measurements?!?!?

[Kornbread]

Re: DIY Flat Panel Speaker Love

Dayton umm-6, calibration file is loaded to REW.

Forgot to mention there are substantial bass traps built and located in both corners targeting a large 120hz peak that I was getting with the NHT's and absorptive panels along the side walls targeted at a somewhat hot mid treble. Since these radiate differently than the NHT's I do not know what difference this made.

[rmeinke]

Re: DIY Flat Panel Speaker Love

The power response graph I saw is in the NXT Technology Review 1.0 paper. The power response is rises fairly steadily from a reference point of approx. 250Hz to 15.2kHZ for an overall increase of 7.27dB at that 15.2kHz peak.

So while not perfect extension to 20K, there isn't much music content above that 15.2K range and why I do not hear a change in balance as you move around the room. Again, no sweet spot for these kind sir. They sound very balanced even as you approach nearly between the speakers.

I remember being amazed when I pulled the panels all the way out into the room where you could walk freely between and behind them. The imaging was maintained as you walked from the front to the back and vs. versa. except for that very narrow band directly between the panel... that was flippin' amazing at the time...

Of course anyone could *could* EQ to taste.

[Gary.M]

Re: DIY Flat Panel Speaker Love

Hi All. I'm new here and have just read my way through all of the DML threads on the various forums. I'll be playing with this myself over the next few months as I get some time. I've enjoyed reading, its refreshing to find that here and on Audiocircle there is a friendly non-confrontational atmosphere. I have come from diyaudio, and its quite different over there!

[Gary.M]

Re: DIY Flat Panel Speaker Love

Thinking about the response curves that Kornbread has posted, perhaps the difference is the the full suspension on his construction? To date I think all the response curves I've seen posted have been from mostly unsupported panels.

[rmeinke]

Re: DIY Flat Panel Speaker Love

Hi All. I'm new here and have just read my way through all of the DML threads on the various forums. I'll be playing with this myself over the next few months as I get some time. I've enjoyed reading, its refreshing to find that here and on Audiocircle there is a friendly non-confrontational atmosphere. I have come from diyaudio, and its quite different over there!

Hello Gary and welcome to the TechTalk forum man! These panels are plain fun DIY with nearly immediate gratification. 1-2 hours is all that is needed to create a very nice sounding flat panel speaker.

Don't worry about how it looks... grab a pair of 2'x2' project panels and Home Depot and a pair of the Dayton Ultra's a few hours in the shop and then sit back and soak in their delicious spaciousness...

[rmeinke]

Re: DIY Flat Panel Speaker Love

Thinking about the response curves that Kornbread has posted, perhaps the difference is the the full suspension on his construction? To date I think all the response curves I've seen posted have been from mostly unsupported panels.

You would think its the solid mounting... but I find it hard to believe that high freq. can be that rolled off at 1 meter and then extended at the listening position... strange no doubt... :(

[rmeinke]

Re: DIY Flat Panel Speaker Love

Measured from 1 meter in front of panel at about the midpoint of panel height with smoothing.

[ATTACH=CONFIG]61573[/ATTACH]

Notice the overlay of the nearfield and farfield measurements in post 437. The top end was there when measured at the listening position but not at the 1 meter position. Could it be the panel is beaming at high frequencies similar to other large panels? This could explain why the top is so very rolled off at close distance, the mike was in a treble null. The null fills in at distance given enough distance to merge. Could this also hold true for the rest of the frequency spectrum. Food for thought.

Hey Kornbread... didn't ask yet... but how do they sound?!?

I really didn't read this very closely but now that I did... just curious... wonder what happens when you move the mic around the panel...

This seems like some type of measurement error, caused by what I'm not sure but these results make little sense to me given all the measures I've seen and taken.

Like I said, move the mic around the room and see what other measurement observations you can make...

[Kornbread]

Re: DIY Flat Panel Speaker Love

Plans are to take additional measurements @1 meter @ the same position then moving the mic position while watching the response on screen.

I like to get some kind of grasp on things before just going whole hog wild tweaking it. Hence, I have a baseline to compare mods with and if a measurement doesn't look right, and this one is odd for sure, the test can be repeated. This also gives others needed data to aid in their quest.

... of course this will happen only after I get a break from all this blasted work.

[Gary.M]

Re: DIY Flat Panel Speaker Love

Some notes on my plans...

I have purchased and received several of these exciters from Parts Express:

Dayton DAEX32Q-4

I will be using an extruded polystyrene branded Goldfoam 25mm thick.

I will dimension it according to NXT's guidlines (I'm using 1m high by 730mm wide, quite close to the Quad ESL panel actually)

I'll be running it with one exciter, foam shaped and treated according to best practice on this and audio circle forums (rounded corners, sanded off surfaces, diluted PVA treatment etc).

It will be some weeks before this gets underway as I am very busy with work at present.

I will report in due course.