New Home for the BG Ribbons - Part 27

[JonMarsh]

Another point I wasn't thinking about clearly is that this is a line array- (doh- smack self on forehead) - you'll have an integration issue anyway, re the minimum distance- this applies to the midrange or HF line source you use like the RD drivers.

Bohlender Grabebener's recommendations for minimum listening distance for the line array, to avoid perception (and measurement!) of decreased output in the high frequencies is as follows:

RD75: 15 feet
RD50: 10 feet
RD40: 8 feet
RD28.1 6 feet

This will probably impact things more in the HF range, but it's obviously going to have an impact on the lower frequency range- it may explain your notches in response trying to do nearfield.

How big a room are you going to use them in? Combining both line array and dipole might make things a bit trickier.

BTW, when you were measuring "nearfield", which driver? Just curious- it might be interesting to experiment.


Just for the record, I use Praxis and Fuzzmeasure Pro 3, though lately, mostly the latter. I also have a CLIO system and Just MLS came with my LspCAD, but I much prefer Praxis or Fuzzmeasure Pro; both use programmable sine chirps, and have robust S/N good options for windowing.

[chasw98]

How big a room are you going to use them in? Combining both line array and dipole might make things a bit trickier.

One of two roms, either 12 X 15 X 8 or 24 X 14 X 8.

BTW, when you were measuring "nearfield", which driver? Just curious- it might be interesting to experiment.

The Peerless 830875. Right at the very center between driver 6 & 7 (44"), or above at driver 7 (54"), or below at driver 5 (41.5").

Just for the record, I use Praxis and Fuzzmeasure Pro 3, though lately, mostly the latter. I also have a CLIO system and Just MLS came with my LspCAD, but I much prefer Praxis or Fuzzmeasure Pro; both use programmable sine chirps, and have robust S/N good options for windowing.

I have also pondered the fact that the wavefront requires distance to develop and therefore measure, also that the rooms I have to use it in are not ideal. When the rain lets up (and I get my next shipment of drivers) outdoor measurement is definitely in the cards. Now where did I put the crane to hoist it up in the air? :E :B

[Amphiprion]

The 'moving dips' phenomena was well known back in the Bass list's RD75 group buy days, wasn't it? That's why I bet my $1.

[JonMarsh]

The 'moving dips' phenomena was well known back in the Bass list's RD75 group buy days, wasn't it? That's why I bet my $1.

I'm sure you're right, Mark. I'm working on work-work stuff today and didn't see your post earlier.

Another thought or suggestion, when I go back and look at some of my notes and studies in preparation for the Saint-Saens project some years ago, one conclusion I leaned toward was that you'd mostly have to pick a target distance for the line array design; especially if you don't use a tapered array that's effectively floor to ceiling; that should be your planned listening distance, and the integration issues such as encountered with the RD drivers suggest a minimum distance. It's a bit tricky overall, too, because usually folks want to be "nearfield" in the line array to avoid the issues with fall off in the behavior from a true line source to something else, that doesn't behave that way.

One way to ameliorate the issue, of course, is to use a floor to ceiling line, so that you have virtual arrays added to the effective height, by virtue of the boundary reflection. Otherwise, what happens is that too far from the array, the bottom end doesn't fall off at the lower rate, but at a faster rate, and the overall balance becomes somewhat "shouty", favoring the midrange, which is still working in effective long throw line array mode.

I heard this behavior in Mark K's family room a few years back when he hosted a Northern CA DIY event in his home, with a set of line arrays constructed like Jim Griffin's original paper and design. At a certain distance and front axis, they achieved their optimal presentation; at longer distances, well under what one might listen in Mark's room, they were getting rather on the shouty side.

So, there are a number of variables to juggle, and in general unless you do a floor to ceiling array, I think you find you have to design for a specific listening distance.

[Dennis H]

Bohlender Grabebener's recommendations for minimum listening distance for the line array, to avoid perception (and measurement!) of decreased output in the high frequencies is as follows:

RD75: 15 feet
RD50: 10 feet
RD40: 8 feet
RD28.1 6 feet

Yup, here are measurements of the pro version of the RD75 from one of their white papers -- bottom to top, 1m to 4m. There's a huge null at 2m where mic distance equals line height and you really need to get out to 2x line height before things flatten out.

[Amphiprion]

I'm working on work-work stuff today and didn't see your post earlier.

I was trying to have one of those work-work stuff days, but it ended with some traces evaporating, a dead MOSFET, and an irreparable board. Traces just literally not there anymore. Oh well, that system was flawed anyway. Back to the Class E RF generator. I ordered a Type 2 powdered iron core from Amidon - a 4" OD x 2.25" ID x 1" tall doughnut of an RF choke. Winding it is going to be a pain. It's sitting on my desk here and looks like one of those weights you put on a baseball bat for your practice swings.

[Amphiprion]

Just a random question, but what does a time averaged spectrum RTA measurement look like?

[chasw98]

A new day, new measurements. After the morning lawnmowers and leaf blasters died down, I was able to get some 4 meter and 2 meter measurements. Here is what it looked like and an impedance sweep.



Impedance

[chasw98]

These are measurements taken at 4 meters distance, 90 dB SPL, 10 ms gate at heights of 30 inches, 44 inches, and 60 inches.

30 inch height.



44 inch height.



60 inch height

.

[chasw98]

These are measurements taken at 2 meters distance, 90 dB SPL, 10 ms gate at heights of 30 inches, 44 inches, and 60 inches.

30 inch height.



44 inch height.



60 inch height.

[ThomasW]

You're using the ECM-8000?

Just for comparison do a nearfield/farfield plot using TrueRTA

[chasw98]

You're using the ECM-8000?

Just for comparison do a nearfield/farfield plot using TrueRTA

Yes, with Kim G calibration file.

Now you ask, I will set it up again in a little while! :B

[chasw98]

Here is TrueRTA at 4 meters and 40 centimeters.

4 meters.



40 centimeters.

[ThomasW]

Ok shows similar stuff. Not an issue for what you're building given the intended 650Hz XO point.

If I get a chance later today I'll mount a single driver in a baffle and run some sweeps.

[Hank]

Chuck, you're doing good stuff there - lots of work but an educational experience. I think measuring a LA close doesn't work. Depending on aiming a mic at a driver center or between drivers and different distances up and down the LA will give different, confusing results. As you've found, you have to back up to where the half-cylinder wavefront has developed to get the mic to "hear" what your ears will hear. Took care of that false 900 Hz dip
Those B-G Minimum listening distances are pulling me towards the RD-50. My current listening distance is 12 ft and the RD-75's Minimum is 15 ft. I've been busy and in a funk lately (my new car has hail damage - meeting with my insureance co. adjuster this afternoon :cry:
Also, I'm still leaning towards one baffle for the Peerless and B-G RD's (or Neo 8's).
Regarding support, I may just do a "spine" of 3/4" MDF veneered on both sides with some of that cool WilsonArt Decorative Metals laminates - should be quite stiff.

[chasw98]

Chuck, you're doing good stuff there - lots of work but an educational experience. I think measuring a LA close doesn't work. Depending on aiming a mic at a driver center or between drivers and different distances up and down the LA will give different, confusing results. As you've found, you have to back up to where the half-cylinder wavefront has developed to get the mic to "hear" what your ears will hear. Took care of that false 900 Hz dip
Those B-G Minimum listening distances are pulling me towards the RD-50. My current listening distance is 12 ft and the RD-75's Minimum is 15 ft.

I will have 8 Neo8-PDR's and 22 Neo3-PDR's by Friday this week. I have the Neo8's on a lend lease program where I can return them if they don't work out. I will be getting them into a baffle and going through the motions just like I did with the Peerless drivers to see what happens. I agree with you about having to 'back up' to get a somewhat realistic measurement of a LA. Not at all like an ordinary box monopole! Just like you my grand hall is not very grand in size but is big enough to accomadate what we are trying to build.

I've been busy and in a funk lately (my new car has hail damage - meeting with my insureance co. adjuster this afternoon :cry:
Also, I'm still leaning towards one baffle for the Peerless and B-G RD's (or Neo 8's).
Regarding support, I may just do a "spine" of 3/4" MDF veneered on both sides with some of that cool WilsonArt Decorative Metals laminates - should be quite stiff.

Sorry to hear about your car. Don't you hate it when the new shiny good looking machine is now changed forever by a circumstance out of your control. What a drag. I empathise with you.

I have come to the conclusion that an 'artsy' spine out of MDF will do just fine for my needs also.

[ThomasW]

I was thinking there was something inherently wrong with Chuck's plots until I ran across this one from Mark K's RS225 line array....

[Johnloudb]

Comb Filtering?

[chasw98]

It's the weekend and I'm back with more stuff..................

I got 6 out of 8 Neo8's on Friday and also received 22 Neo3's. I cut up some 1 1/2" mdf this morning and just mounted them up spaced 1 inch apart on a 12 inch wide slab. Here are the pictures. I set them up with a DCX2496 and have played with some settings. They seem to be very happy with LR12 slopes.

Coffee, tea, or Neo3's!



Fitting the drivers into the slab.



Drivers temporarily mounted.



Standing up and wired on the porch.



BTW, they sound really, really good for just a quick setup. Definitely worth pursuing further. No measurements yet. I should have some for tomorrow.

[brucemck2]

I'm sure they'll sound terrific when done.

These sorts of measurment differences as a function of height / distance / angle are what make the "plug and play" DEQX type of solution tough (but not impossible) in practice.

[Johnloudb]

BTW, they sound really, really good for just a quick setup. Definitely worth pursuing further. No measurements yet. I should have some for tomorrow.

That's all I needed to know ... I'm gonna get started now. Do they need much equalization, other than Bass EQ to get them to sound good?

[chasw98]

12 db at 80 Hz without using a subwoofer (which I plan on adding later). No other EQ except level matching. -3 dB lo, 0 dB mid, +6 db high. Crossover set at 400 and 4,000. One interesting setting occured. The DCX2496 has an 'auto-align' function for setting phase/delay and polarity. I plugged in my mic and ran it. In all the times I have used that function the unit has always set some sort of phase/delay no matter how small. It did not set any phase/delay at all and no polarity change either. I am still waiting on 2 more Neo8's to fill up the mids.

PS - For the bass EQ I will eventually build a small passive inline circuit similar to what JonMarsh did for the Arvo so that everyone building this array can take advantage all it has to offer even when using all passive or somewhat analog crossovers that don't allow for EQ.

[chasw98]

I have a question.
Since this is an open baffle loudspeaker, should the back side be as open as possible? I mounted the Neo8's and Neo3's in some MDF that was glued together to form a 1 1/2" baffle. The Neo's are not very thick. As a matter of fact they are maybe 1/4" thick or not much more. When the Neo's are mounted in a hole that is 1 1/2 inches deep, it certainly looks like the back side is not as open as maybe it should be. Even with roundovers or 45 degree angles on the back it still doesn't seem open enough or as open as it could be. I am thinking that it might help to mount the Neo's on 3/4" MDF rather than double thick 1 1/2" MDF. Actually I was thinking about using 1/4" black plexiglass and surface mounting that into the baffle with threaded mounting holes for the Neo's............but maybe not. Might look cool. Possibly $pendy.

Does anyone have any facts or intelligent opinions about this or am I splitting hairs? :E

Chuck

[ThomasW]

I was going to mention the thick baffle but didn't since this is a test baffle.

Yes you want to avoid creating a 'cavity' behind the little planars.

[chasw98]

You're right, it is a test baffle. I used what was convenient. And how would you define 'cavity'? Is 3/4" depth OK? (Obviously better than 1 1/2" but not as good as 1/4", right?). I am still thinking that 1/4" plexiglass inset into the face of a 3/4" baffle might work very well, both mechanically and acoustically. Then make the side with the woofers double layer (1 1/2") with aluminum struts for rigidity. Add a rear support and voila!

[ThomasW]

Yes a plex or alu mounting plate sounds like a plan.

[penngray]

Actually I was thinking about using 1/4" black plexiglass and surface mounting that into the baffle with threaded mounting holes for the Neo's............but maybe not. Might look cool. Possibly $pendy.

Chuck, I would love to see the look of that. I have 3 extra Neopro5is sitting around with zero projects.

[chasw98]

And here are this weekends measurements. These are of the Neo8's in the array.

Neo8 at 100 centimeters.



Neo8 at 100 centimeters 1/6th octave



Neo8 impedance

[chasw98]

And here are this weekends measurements. These are of the Neo3's in the array.

Neo3 at 100 centimeters.



Neo3 at 100 centimeters 1/6th octave



Neo3 impedance

[chasw98]

And just for grins here is a True RTA plot of the Peerless woofers, Neo8's, and Neo3's.

[looneybomber]

Chas, I hope this wasn't already addressed, but why use the Neo8's? What about a 2k XO between the Peerless and Neo3's?

[ThomasW]

The idea here is to run each driver in it's sweet spot



That means running the Neo3 with a 3.5kHz or higher XO point so they can handle some power and don't require brickwall filtering.

Also we want to avoid any comb filtering issues associated with running a line array of 6.5" midwoofers up high.

[chasw98]

Chas, I hope this wasn't already addressed, but why use the Neo8's? What about a 2k XO between the Peerless and Neo3's?

As noted by ThomasW 2K is little too high for the Peerless woofers to operate in their happy range and it is a little low for the Neo3's to handle. Right now I am 'testing' the Neo8's as a mid and am finding that a 400 Hz and 4K xover point allow each driver to run quite effortlessly in their given range. An active cross fromm woofer at 400 to mid will allow for level matching between the cone drivers and the planars, and a passive at 4K for mid to high will afford very good integration between the 8's and the 3's plus a little buffer for the 3's in case of more power being applied. With all three, it is starting to sound very sweet, with just two, I believe it would be straining. You also have to remember that this has been in the planning process since last October so a lot of thought has gone into the components before any money was ever spent.

Chuck

[chasw98]

OK, the Neo8's are sounding so good that I am taking them to the next step and getting ready to build another baffle that will hold them and the Neo3's properly. That being said, I am tired of not having a jig to make the mounting holes and indents for these things so I built a jig. I am able to route out a hole for an 8 in seconds now.





It is fast, accurate, and easy to use. Once I am done using the jig on my project I am willing to lend it out to anyone seriously building these. Just PM me as time goes on and we will work it out.

Chuck

PS - I will probably build a Neo3 jig as well to go with this one.

[JonMarsh]

OK, the Neo8's are sounding so good that I am taking them to the next step and getting ready to build another baffle that will hold them and the Neo3's properly. That being said, I am tired of not having a jig to make the mounting holes and indents for these things so I built a jig. I am able to route out a hole for an 8 in seconds now.





It is fast, accurate, and easy to use. Once I am done using the jig on my project I am willing to lend it out to anyone seriously building these. Just PM me as time goes on and we will work it out.

Chuck

PS - I will probably build a Neo3 jig as well to go with this one.

Sweet- then you can "rent them out" to anyone going down the same path- sounds like you won't be needing any comparisons done with RD50's, eh? I've got a pair here that can be measured, but I've been might busy- as we sometimes say back in HQ, slow work takes time. :W

[CraigJ]

Excellent job Chuck,

My question might be premature, do you think a person could get close to what your are making with 6 Neo8s and 12 Neo3s per side in a smaller room? Perhaps combining these with 4 RS270s or 6 RS225s? I'm sure your systems already sounds fantastic, just considering something "smaller" and a little more waf.

Thanks, and hope your neighbors enjoy them

Craig

[chasw98]

Craig:
I am by no means an expert on line arrays...... yet! But from what I have learned, a lot of the line array theory is based on so called 'floor to ceiling' line of drivers in order to produce the wavefront and to take advantage of the mirror imaging of the floor and ceiling to give the wavefront the 'line array' properties that go into making the sound a line array is associated with. Part of it is the fact that each driver can receive a 10th of a watt and produce a very large sound but the single driver is running basically at idle. As soon as you start decreasing the amount of drivers involved, the efficiency goes down and the drivers are now being run at higher levels. Just what the tipping point is for the line array phenomenon to go away or to appear, I am not sure. I am pretty sure 6 Neo8's and 12 Neo3's would still work, but I think it would be stretching it a little thin with 6 RS225's. Now 8 RS225's might be OK. I just don't know enough yet to tell you. Thomas or Jon can probably make a better guess than me. So far I have tried the array I have setup with 5 Neo8's, 6 Neo8's, and 8 Neo8's. 5 was not enough, 6 was OK, 8 really sounded great. All this is without a polished, finished crossover so take it with a grain of salt.

[Bear]

Craig:
I am by no means an expert on line arrays...... yet! But from what I have learned, a lot of the line array theory is based on so called 'floor to ceiling' line of drivers in order to produce the wavefront and to take advantage of the mirror imaging of the floor and ceiling to give the wavefront the 'line array' properties that go into making the sound a line array is associated with.

I'm not sure how authoritative this is, but it reads well and the math is presented in a relatively easy-to-follow fashion:

http://www.systemsolutions.co.za/Tech%20Zone/Line%20Array%20White%20Paper.pdf

[JonMarsh]

I'm not sure how authoritative this is, but it reads well and the math is presented in a relatively easy-to-follow fashion:

http://www.systemsolutions.co.za/Tec...te%20Paper.pdf

Yup- Jim Griffin's white paper for 2003- the reference 4 from JBL is probably one of the most important he cites.

[chasw98]

This is sort of a bump to get this thread back on the first page.

I have been on a trip to Washiongton, DC and visited a friend who I would call my mentor. He is an oceanographer who uses sonar line arrays to map the ocean floor. I was asking him about the dilemna of measuring an array and showed him the problems I am having. His suggestion was to dig a trench in your yard to lay the array on its back in and then fly a microphone 3 to 4 meters above the array for measurement. He also noted that even with that scenario you might never get valid measurements that are useful for building a crossover. He also suggested for measuring purposes to power taper the array.

I am not sure how well it will go over to dig a trench in the back yard, but maybe supporting the array on a pair of sawhorses and then measuring from above might be worth looking into. Just an idea....

[JonMarsh]

I vote for option #2. :W I bet your wife will, too. :rofl:

In the end, your measurement distance should reflect the way you're going to listen to them- at least as regards measurement distance and height. That way the crossover tuning should be optimized as near as feasible for how you're going to be listening to them. With somewhat nearfield listening, power tapering may be a good idea- I don't recall what Jim Griffin's paper says, might be some nuggets in there for you.

[ThomasW]

"Measuring large dipole speakers is always problematic, because the underlying assumptions about the measurement techniques and their relationship with the device under test are no longer completely valid."

As Chuck knows John Whittaker's research into measuring dipole line arrays is a definitive resource.... http://ldsg.snippets.org/ALSR/index.php#HOME

[chasw98]

As Chuck knows John Whittaker's research into measuring dipole line arrays is a definitive resource.... http://ldsg.snippets.org/ALSR/index.php#HOME

Huh? :E :T ;x(
Oh, yeah, right!

[Johnloudb]

What I found really interesting is that his DynaPleat speaker was crossed over 1st order, between the Dynas and RD75. Hmm... for an RD50 that would be a lot to ask of the driver in the low frequencies, but maybe a dual slope crossover at 500Hz. First-order for an octave.

[ThomasW]

The basis of John W's testing and use was based on B&G RD Series recommended 150Hz XO point. Thing is B&G themselves never used that low a XO point in their own designs. As as example their freestanding 520i has a 350Hz XO point.

Years ago when JonMarsh and I originally measured my RD75s, we found a XO point between 500-650Hz gave the best sound quality. Our experience was validated by other builders in subsequent years.