Separate SMAART Measurements on mids and peizo arrays

[hclague]

Hello All

I have just finished converting on of my DR200’s to be able to use it as either passive (original design), or active 2-way. In the active 2-way mode, the load for the piezo array is provided by one of these:

http://www.partsexpress.com/pe/showdetl ... er=019-020

I have wired in a switch to select between passive and active.

This has allowed me to take some SMAART measurements on the mid horn section and piezo array separately.

First the mid horn section, here is a screen shot with no corrective EQ filters and only a high pass filter at 125 hz. (linkwitz/riley 24db/oct):

http://www.frappr.com/?a=viewphoto&id=3 ... myphotos=1

This picture pretty well illustrates why people say that the DR speakers need EQ to make them sound “good”.The blue trace is the frequency response. The red trace at the top is the coherency. As you can see the coherency is good to about 4 khz meaning the measurement should be valid to that frequency. The peaks between ~ 300 hz and 900 hz as well as the dip at ~ 1 khz are “Equalizable events” and can be corrected with parametric filters.

Here is a shot of the same box’s mid-horn section after corrective EQ filters have been applied:

http://www.frappr.com/?a=viewphoto&id=3 ... myphotos=1

There is also a low pass crossover filter applied at 2 khz.

Next we have the piezo array’s. I took measurements on both an 8 element cut and glued 1016 straight array and a 10 element ( 5 and 5) crossfiring 1016 array.

First the 8 element straight array:

http://www.frappr.com/?a=viewphoto&id=3 ... myphotos=1

The peaks between ~ 4 khz and 6 khz and the ~ 9 khz spike are “Equalizable events” and can be corrected with parametric filters. True to claims, the piezo’s roll off dramatically below 2 khz.

And now the 10 element crossfiring array:

http://www.frappr.com/?a=viewphoto&id=3 ... myphotos=1

with this array we are seeing some phase or comb filtering effects caused by the crossfiring. These are illustrated by the spike at ~2 khz and the dip at ~ 4 khz. I believe the 2 khz spike can be tamed with a parametric filter, but the 4 khz null is probably “Un Equalizable” because it’s a phase problem. Bill does mention a null in the upper frequencies in the DR200 plans when using the crossfiring array.My null was lower than Bills stated 6 khz probably because my 2 banks of 5 were a little farther apart form each other.

I was also able to due an A/B comparison between passive and active since I was using the Behringer DCX2496 crossover and I had installed the switch on the DR200. When in active 2-way mode, there was definitely a difference in “smoothness” through the upper mid range. Audibly better. This makes sense since I was able to time align high to mids through the crossover region using SMAART.

My personal conclusions:

1. DR200 as designed ( passive/acoustical crossover) is certainly a usable box after corrective EQ is applied. Better sounding than most MI level speakers. Certainly more efficient. However, Biamping did improve the sound in the 2 khz to 4 khz range and since I have the capability I will be converting the rest of my boxes to be able to go active.

2. The difference in the straight vs crossfiring array is very audible to me. Although the crossfiring array certainly has wider dispersion, the straight sounded much smoother and still appeared to have good dispersion to at least 90 degrees. This is fine for 90 % of the gigs I do. I will be converting the remaining 3 crossfiring DR’s to straight.

3. I question how well the two different types of tweeter array’s couple with each other when stacked vertically. Two of the same type sounded better than one of each in earlier listening tests.

I will be doing some further testing before I post exact settings for the DCX2496. I want to try different crossover types ( Butterworth, Bessel etc…) to see what difference they might make. Once I’m through testing, I will try to post the settings. Then I will start to try and establish settings for the Crown Xti amps, specifically the Xti2000

DISCLAIMER! These measurements and any future settings posted are for the DR200 specifically. Please do not expect them to work for other DR’s. I would expect exact Frequency response to vary between models.

Hal

[Strapping Young Stu]

This is great I think this is exactly the sort of thorough testing that could really make the difference for Bills designs, if you could get this data along with waterfall and phase plots that would be good data to be able to pass on to those who require it

Stu

[Bill Fitzmaurice]

Do plots of the straight array and the cross-fire array at 45 degrees off-axis. I think the results will be most interesting. If all I'd gone by was the on-axis response I'd have never done the cross-firing array to begin with. As far as the combining of flat and cross-fired arrays goes, with only a couple of cabs that's not going to work well. You really have to run at least a three cab high array to make it work, preferably more, and at that point the upper and lower cab HF sections are operating pretty much independantly to different sections of the audience. You'd even want to EQ the straight and cross-fired cabs independantly, as is done, for instance, with an array of EAW KF760/761 cabs.

[Dave Non-Zero]

This is great I think this is exactly the sort of thorough testing that could really make the difference for Bills designs, if you could get this data along with waterfall and phase plots that would be good data to be able to pass on to those who require it

Stu

Great post! Just dont start on 'waterfall' and 'phase plots' till ive got my head around 'coherence traces' etc! haha

[jeffbabcock]

Good stuff! This is a very useful post. Interesting that you noted such an improvement in biamping..... certainly much discussion should stem from here. As Bill mentions it would be interesting to see off axis response. But regardless for those users with smaller stacks they may want to carefully evaluate the crossfiring vs no crossfiring options based on the test outcomes.

[gdougherty]

As was recently suggested to me on my question and Bill reiterates here. Ideally a small stack (mine are 2 pairs of OT12's) would be all the same. Trying to cram large array physics and practices into very small arrays just doesn't work.

[hclague]

I will try to do some measurements on crossfired vs straight arrays at 45 degrees when time permits.

Hal