Danley DTS-10 "Super Spud" DIY kit - Page 77

The only output that counts, IMO....What good is more output if it sounds like crap? SPL comps?

I also think that CEDIA standards, while stricter than any other that I know of, are still pretty lenient...I don't allow more than 5% THD in my setup. I test with -3.0dBFS sinewave tones....it is easy to hear the difference between 1% and 5% THD down low (below ~30-40Hz) this way, esp. if the 3rd or higher harmonics are the major contributors. Higher up in freq, I find it harder to hear distortion, as our ear's response levels out more.

JSS

"There has got to be some left in the tank at least above 30hz. I'm telling you there is no way that any of my other subs can hit the levels the DTS-10 can in that range."

ricci, you are hardcore. how would you respond if somebody called b.s. on your claims? are you saying a dts-10 beats two sealed re xxx's from ~40hz up and/or are you saying a dts-10 beats your super custom ported re xxx from ~40hz up? if so, what data have you to back such claims? are you amp limited on the relatively inefficient xxx's or is there something else? is it a measured effect or a subjective effect? at what frequency does one system win/tie/lose to the other?

Yes, 2m groundplane numbers seem most sensible given that it is equal to anechoic 1m results. Ilkka's numbers are 2m, but I think the CEA-2010 standard is 1m groundplane, so let's keep that in mind when comparing to other CEA-2010 results if/when such results are posted by manufacturers or individuals.

One thing that popped into my mind is the form factor of the DTS-10. The baffle on which the horn mouth is located is very large and the baffle step effect will consequently start at a low frequency. At some frequencies the radiation is to quarter-space or in between half/quarter-space. I already considered this in my own measurements of a compact 15" sub with only 0.5m x 0.5m baffle where I expected at the highest a 2-3 dB difference in mid-upper bass distortion figures compared to a point source (or the power response of the sub).

jss, i'm in the same ballpark with test tones. however, with music, distortion gets masked to a much higher degree to me. you?

I have owned horns in the past so I had certain expectations

I would think the eD's would have the advantage in frequency response since I could spread them out for better coverage. I love horn midbass so it would be close. It would still be subjective and some prefer the eD's and others the Danleys. If I could find a perfect spot for the Danley to give me a good frequency response I would go with the Danley. I am biased to the midbass punch but I would not want one without the other. The DTS-10 does it better than anything I have heard. I am selling my sono's for cheap($200 each) and 7 of those I would take over the 1 Danley.

Nothing sexier than a good looking woman actually building. Marion, you better hang on to her. Be sure and show them in place.

You bet, same here. I figure if I keep distortion to a low-to-no audibility level with test tones, It means I will always have clean sound for program material...

What first got me curious about distortion and overexcursion was the THX cavalcade trailer that would sound significantly different when the volume was turned up on my old system....

JSS

for xmas all i want is danley to release the actual plans with measurements so i can build one from scratch. pretty please danley with chocalate and a strawberry on top.

"I gotta rock "


dbl

The problem with that is getting a driver to work.

I'm just hoping for an announcement that the kit's selling so well they've decided to offer it indefinitely. I was going to try and upgrade my HT receiver next year with Emotiva separates, but honestly such an announcement would change my priorities real fast.

I've done similar THD tests using REW before and I find that indoors it leaves much to be desired, so much so that I don't bother with it anymore. It's influenced way to much by the room to be useful for anyone but the owner. You must have a very good room and ear to hear those small differences in bass distortion between 1 and 5%. I find it hard to discern it like that through the random rattles and buzzes of things throughout the entire house vibrating, even with sine waves. There's also the fact that one level (5% THD) is much louder than the other (1%THD), not to mention (10%THD). This is before we even get into microphone distortion at low frequencies and other issues. Interestingly I think it's easier for me to hear higher freq mid range distortion. With typical material it has to be quite loud to get the distortion to be noticeable in the bass for me.

Good points.

In short you are suggesting that the very large baffle of the DTS-10 may be goosing the distortion harmonics level up to some extent?

Strangely enough, I first started testing with REW when I was trying to figure out why my old sub distorted and sounded bad at higher volumes with some movie scenes. I would slowly turn the volume with sinewaves up dB by dB, and it would become obvious when it would start to go bad...and once it starts going bad, just a few dB more can put you over 10% THD quickly...

I agree about the house rattles possibly messing up results indoors, but indoors is where we listen, so it is of paramount importance to me that it measures as distortion-free as it can indoors. As a plus, sinewaves are also great for diagnosing and fixing house-induced rattles and distortions (picture frames, cabinet doors, etc). Once you fix those, you are hearing distortion from the speaker itself, if there is negligible THD in your electronics and no clipping.....

JSS

I don't know. Probably shrug my shoulders and allow them to present the definitive proof that it is otherwise, if they have any.

If you are asking if I think that it trumps any of my other subs in efficiency, and maximum output >40hz I'd say yes. The only proof that I have is what is presented in this thread and my subjective impressions along with those of others who have a DTS10. Obviously all of the subwoofers I have are amp limited above 40hz short term. Long term gets into fried drivers with as much power as I use. How much more power can you put on a driver after you get to the 3KW range? Each driver in the DTS10 has a nominal 1000-1200w on it BTW, so it's not like I'm running it off of a modest little amp.

I agree totally that the room is where you listen and so it is of utmost importance to have low distortion there, but what I meant is if you get 110db at 20hz with 5% THD or whatever it happens to be, that is great for you, but it really doesn't tell anyone else how it will perform in their room, or anything about the sub itself unless you have some sort of control to gauge with.

About the noises of things vibrating...I didn't mean that it would mess up your THD results too bad, but that it's what gives me difficulty in hearing a big difference in the sound going from 1% THD to 5% in the low bass. That and the additional volume level. What can and does affect your THD readings indoors is room modes and nulls. Your sub could be relatively clean at 20hz outdoors, but if there's a null at 20 and a spike at 60hz...Bam! There goes the 3rd harmonic THD skyrocketing, or vice versa.

I called and orderd one today. Jeff said they will build more when they get a shipment of drivers next week (12/18). He mentioned there were 5 others in front of me. I'll be picking mine up at DSL next week or the 21st. I can also spend some time in their audition room. Might be dangerous for me though...must stop the upgrades!

BTW-I will be replacing my IB (2 FI18's) with one DTS-10.

Have you guys noticed that most people here are replacing multiple 18's for 1 Danley DTS-10, that should say something about the design.

I too noticed that.
MK-could you draw out a quick diagram or take a picture of your final placement? I realize every room is different, but why not start where you stopped.

I will later. It is not finished yet as I wait for the 3732's. Basically they are in the front corners on the side of the screen. If you look at them from the front(screen) they measure 60 inches tall, 16 inches wide, and 44 inches deep. The horn is located on the inside(towards the speakers), next to the floor, towards the front wall. So looking at the left sub the horn is on the bottom, towards the back, on the right. Confused yet?

Ok. I've went back to looking at Illka's tests again for the billionth time and I noticed something which may, or may not be of interest. The top 4 subs that he tested for CEA2010 output 20-80hz average were the 2 different LMS 18 based diy, the dual SDX15 diy, and I believe the SVS PB13 Ultra. Here is what struck me as curious. If you look at the maximum output graphs before serious compression sets in it is significantly lower than the CEA2010 THD limited output using a 6.5 cycle burst for all 4 of the top systems and I'm sure many others too. The maximum output is using a very long 30sec sine wave sweep, while the distortion tests are a short burst. The distortion limited output is significantly higher than the long term maximum output. Note that he also has distortion plotted for the maximum output sweeps and it exceeds the CEA2010 allowable levels in many cases at the much lower volumes. See where I'm going here?

It seems like you could make a case that a short duration burst can indeed net quite a few extra DB before distortion gets out of hand as compared to a long duration sine wave.

I also noted a few other things like the fact that the PR LMS system's low end actually expands during the compression sweeps. You put 5 db in and get 6db out. I don't doubt that he measured that as I've seen something similar myself before. I never really bothered to think about the cause though. Anybody?

Amen. I have found one +$300 driver that comes close, but you lose ~4dB of sensitivity.

The over the counter LAB12 will not work. My posted graphs were only an approximation of the DTS-10s abilities. To give people an idea of what to expect. If you model the real cab and Lab12s you get a totally different output, and it ain't pretty.

The best bet is to design your own. That is where the fun is.

This is an estimate based on Tom's post and the sine wave numbers vs. CEA numbers in the HTS tests ... It's likely a pretty safe bet to call the HTS subs somewhere between the sine wave SPL and the burst SPL when comparing to the DTS-10.

If one "eyeballs" the PB-13 Ultra in 15 Hz tune (for the sine wave test), it looks like this at compression ...

12 Hz ... 93 dB
16 Hz ... 103 dB
20 Hz ... 107 dB
25 Hz ... 109 dB
31.5 Hz .. 108 dB
40 Hz .... 110 dB
50 Hz .... 108 dB
63 Hz .... 108 dB
80 Hz .... 108 dB

For review, here are Tom's numbers:

So, anyway as of now, using a sine wave, one DTS-10 in half space, at two meters, reaches the following levels;
CEA 2010 Max RMS Output @ 2mtrs.
12Hz 102dB
16Hz 108dB
20Hz 109dB
25Hz 111dB
31.5 114dB
40Hz 114dB
50Hz 117dB
63Hz 115dB
80Hz 127 dB
Ultra low AVG 20-31.5Hz
111.3dB
Low Bass 40-63Hz
114.3
20Hz-80Hz AVG
115dB

I hope I am not irritating anyone by participating in this conversation ... this DTS-10 is just too much fun.

Hi all

Ricci asks and others wonder why aren't the CEA2010 numbers higher?

Well I don't know off hand, these numbers should be set by different things depending on frequency.
When time permits we will try to establish a test which uses a tone burst like the criteria calls for.
Sitting there for minutes at a time with a high level sine wave getting a measurement is probably a bit to harsh as the dark brown voice coils on the first two drivers, the first time around suggest.

Thinking about it now, the 5-6 cycle tone burst signal would already be about 1/3 octave wide in bandwidth so it may be that the readings on the detection side would be different with such a short excitation. I think we will visit this again as the upper frequency numbers ought to be higher even with a sine wave.
Alternately, I am curious to find out what aspect of it is the limiting factor up high.

Also, while I have no reason to suspect anything in the electronics chain, the test criteria is reached which the sum of the flaws in the system reach the levels the standard calls for.
For example, amp clipping, mic, limiting, mic pre clipping, input clipping, any of these could cause the threshold to be reached prematurely.
Lastly, if a speaker made enough air noise, it could exceed the criteria at the high end where the test is most sensitive.
This is sort of why distortion measurements are questionable when taken in a room as anything buzzing or vibrating can be picked up as well.

Like I said, I think we will visit this rating thing again and I think we will try to establish numbers at 8 and 10Hz for fun.

What can you hear
Well, a numeric THD figure is essentially useless in defining what you can hear.
The reason is due to something called masking.
You hear pitch not in a continuous span but in small segments or critical bandwidths called a Bark. Two pure tones of equal amplitude within the same Bark, sound like one note. Two equal loudness tones in separate Barks are heard separately.

A nice over view of all this can be found here;
http://is.rice.edu/~welsh/elec431/psychoAcoustic.html

Take a little bit and read this and follow it to about half way down to where it talks about Auditory Masking.
The little graph has an example of he masking curve for what looks like a 250Hz tone.
The dark curve is the masked threshold that the 250Hz tone produces.
Now, in the presence of the 250Hz signal, ANY tone below the black curve is inaudible, masked by the 250Hz tone.
The reason for the CEA 2010 (it would seem to me) is that it weighs each harmonic more like the shape of the masking curve for the frequency under test.
This is partly why one can have a huge second harmonic and not hear anything, it is the most strongly masked harmonic AND it is a musical second which is part of a chord.
The third harmonic threshold is lower stepping down to the highest (10) harmonics which are at -40dB

For those extraordinarily interested in single digit bass, examine the shape of that masking curve, notice that there is downward masking also. While there are no sub-harmonics generated by loudspeakers like on the high side, that downward masking also means that to hear the single digit sounds, they have to be the loudest thing spectrally or they can be masked by a louder low bass signal higher up.
Remember at low frequencies, keep the shape of the hearing sensitivity curve in mind.

Now, all this has applied to tones (steady state signals).
An old speaker guy acquaintance, Earl Geddes, has concluded and proposed the rather controversial conclusion that with music, the low order distortions most loudspeakers exhibit (as opposed to high order distortion like electronics tend to produce) is (up to a significant level) inaudible and had backed it up with some blind testing.

Anyway I am rambling and have to get ready to go out.
So, at some point we will re-visit this measurement, possibly with a tone burst..
As far as the upper frequencies, above 30Hz particularly, I do not know what the limiting factor in the measurements was here, the levels that were measured certainly are not the maximum output but that is something I will look into.
Have fun all,
Tom Danley

With some of the comparisons between sweep tests and CEA-2010 burst tests we may be seeing different effects at play limiting the output. For example in the comments for PB13 Ultra it was noted that a fuse blew during a higher sweep and thus the real limits at mid/upper bass weren't reached. Also the 5 dB step between sweeps is quite large while the burst test may get closer to the limits.

Thermal compression is taking its toll at the highest level sweeps while distortion remains much lower than CEA-2010 limits. Notice the CEA-2010 allows ca. 32% 2nd harmonic, 18% 3rd harmonic and so on. It looks like these were reached at only the lowest frequencies with the highest sweeps. When the distortion limits were reached earlier than compression (dual SDX15 @ 20 Hz) the result was also reflected in the burst tests, the 20 Hz result is much lower than the next higher frequency and quite well predicted by the sweep test too.

I would imagine that thermal compression at the tested levels would be much less of an issue with a horn than normal subs when tested with continuous tones. I wouldn't expect much more clean SPL when tested with bursts.

To a great extent Earl's right, since all that harmonic distortion does is to add harmonics not present in the original signal. Unless you're comparing speaker A that's totally clean with speaker B that's distorted you have no way of knowing which is the 'correct' sound, and you may in fact prefer the latter. That's the reasoning behind 'Aural Exciters', which add additional high frequency harmonics to the original signal to 'sweeten' the sound.

Where Earl misses the mark IMO, but not by much, is with regard to subs. That's because the added harmonics result in above bandwidth output that shouldn't be there. If the subs and mains are located close to each other even that's not noticeable, but if they aren't and you can directionally locate your sub the reason may be harmonics its producing at significant levels above its passband.
Horn loaded subs in general and folded horns in particular are mechanical low pass filters, with far less above bandwidth harmonic output than direct radiators, so they run far cleaner. That makes them more amenable to Earl's theory, but Earl doesn't like horn loaded subs. Go figure.

I think Dr. Geddes is more a fan of bandpass subwoofers and has often pointed out the low pass function of those and would probably agree with your comment about lessened localization issues.

It would be very instructive to see multitone testing with the DTS-10 similar/same to Ilkka's tests. I predict we will see good non-compressed SPL from that test with a horn, less worries about amp clipping than normal subs (the test has around 11-12 dB crest factor) and good insight about distortion and out of band noise with a more realistic spectrum of signal.

"the PR LMS system's low end actually expands during the compression sweeps. "

That could be explained by the hot VC having higher Re/lower Qe, which would give higher output at Fb but and overhang.

thanks ricci.

(by the way, i think you have done an excellent job in this thread, all things considered )

what type of glue are you guys using on these cabinets? Titebond, liquid nails, ???

thanks