RD-75 Active Dipole Modular System

[gregw]

I wanted to share some details of a speaker system I put together for my basement listening room and home theater. I've been reading here and at other speaker building sites for many years but my participation was essentially all vicarious. I followed along and built systems on paper and in my head but work and other projects always kept me from actually building anything.

Along the way, I bought drivers when deals came along that fit into various projects I had in mind. I have accumulated a nice collection. I bought a pair of BG RD-75 drivers when a pair was available used. I became interested in the RD-75 when it was a hot topic on the old Bass List. Many plans then were formulated around the Carver Amazing series which used a 60" planar magnetic driver manufactured by BG I believe. It used high Q 12" drivers in an open panel dipole to achieve good low bass response.

When someone from TC-Sounds offered some 15" aluminum cone drivers with specs suitable for dipole use at a great price, I bought 10 of them. Of course, after I bought them, I read many threads about suitable drivers to pair up with an RD75, and 15" woofers were rarely on the short list. I was not deterred.

My first plan for these drivers was to create a modular panel arrangement attached to a metal frame. I did a sketchup design using unistrut components. I had never heard of unistrut but saw it mentioned in a comment someone made on this forum. It looked like an ideal modular system to provide the frame for a modular speaker. That way, I thought, if I was unhappy with any component, I could put a line of something else in its place. If the 15" woofers proved unsatisfactory, I could swap in an array of smaller drivers. If the RD75 was superseded by something better, like say an array of Neo10s, I could substitute. As it turned out I didn't use any unistrut components, but it was a good idea. I kept it much simpler. More details to follow.

[gregw]

So it all looked good on paper. I picked up a Driverack PA to use for crossover and eq. I bought a Behringer A500 amp and an EP4000. Still, it all sat for a couple more years.

More projects around the house were getting done and the basement space was starting to actually be space instead of boxes and junk. Then I saw Stigerik's RD75 thread over on Diyaudio and I thought I can do that. What he had done was hang the RD75 from the ceiling with no baffle at all. He crossed it over at 200 Hz to some 21" dipole woofers.

So I decided to do something similar to see how it sounded. My basement ceiling is only 79" high to open joists. A picture is worth a thousand words.



I hung it from a cross piece that could swivel that attached to another piece that spanned between the joists and could move side to side. So I could adjust the toe-in and the width of the drivers.

With the 15" woofers, I decided to mount them on a flat panel, 24" wide and 75" tall. I wanted the panel to be stiff so I used two sheets of OSB and one of MDF. I cut all the holes for the drivers with a Jasper jig, then I glued them together with PL premium and used a whole bunch of screws to clamp them. I also glued and screwed some 3/4" oak along each side for additional stiffness.

I bought some oak stair tread to use as a base. The oak stair tread comes in 4 ft lengths and is 1 1/4" thick if I remember right. Very stiff stuff. I didn't want these heavy panels falling over so I made the base pieces 24" long and used some heavy duty shelf brackets to bolt them to the panels.

I didn't care what these looked like since I planned to use them behind black grill cloth and an acoustically transparent screen. So no paint or special finishing.





The bags of cellulose insulation are temporary attempts at bass trapping. The floor is just bare concrete. I have my projection screen hanging from chains and I swung it up and clamped it to the ceiling to take the picture.

I set the crossover at 400 Hz then used the auto eq function built into the DBX Driverack PA. It uses pink noise to set the eq automatically and has several house curves you can choose from. I expected to see a boost needed at low frequencies due to the open baffle. Instead, the auto eq made significant cuts, like 9 or 12 dB all the way down to 31 Hz where it then boosted the output by 9 or more dB.

In measuring with REW, the bass response was good down to about 35 Hz where it then dropped off. I did probably hundreds of measurements with REW and tweaked the eq and amp levels. For music, it really didn't seem like I needed deeper bass. There was bass in abundance. But for movies and to insure I didn't overdrive the open baffle woofers, I dragged my Velodyne ULD-18 down the stairs and incorporated that into the mix. I crossed at 50 Hz to the Bass panels.

The sound quality is just amazing. Prime listening distance seems to be about 10 feet away for best stereo imaging. The most striking impression is that the musicians seem to be in the room. Victor Wooten on bass guitar sounds like he has his Hartke bass rig set up right over there. Vocals sound totally natural. I don't notice anything missing on the high end. Where a cymbal is prominent in the recording, it sounds like a cymbal.

I have to think that it is the large radiating surface area of these speakers that accounts for their effortless low distortion sound. I generally listen at about 80-85 dB sound level, which seems plenty loud enough to get their full impact. I have had the Velodyne sub for many years so I am very familiar with low bass, but the bass from these speakers is of a different character entirely. It fills the room and can have real power to it.

One thing I like about it is that it can be a never ending project if I choose. I have thought about buying 24 of the Aura 6" PE buyout speakers and trying an array of smaller drivers with a higher crossover point to the RD75. I have a pair of Raven R2 ribbons that I plan to add to the mix to see if I prefer it with a dedicated tweeter.

I would like to replace the Velodyne sub with a false wall infinite baffle behind this setup to get better low extension for movies. I want to also work on acoustic treatments for the room.

I am sure improvements are possible, but I could not be happier right now with how it sounds.

[Mr.Ed]

Now THAT is DIY at it's best. Very cool.
-Ed

[gbegland]

Cool man! Keep going. I think you'll find a difference of opinion around these parts on how low to run the Rd-75. I'm pretty sure Thomas and Jon don't use them below 500-600Hz, but as you say the other Stig(top gear anyone?) runs them down to 300 and is in love. I'm guessing the helmet has something to do with his hearing... :W

Greg

[JonMarsh]

Considering the two way system and the woofers you're using, 400 Hz is a pretty fair compromise for the crossover frequency. OTOH, my experience is that the best distortion performance and "ease" with the RD's will be obtained at around 600Hz crossover, and as you might expect with 6" or 8" mid woofers, that is a good crossover target. ThomasW uses Acoustant 1+1's as dipole line source midwoofers between 100 Hz and 600 Hz.



If you have a narrow long room, than wide dispersion in the highs may not buy you much, and running the RD's alone is pretty OK. I'm considering that for a cross firing corner placement at my GF's.

On the other hand, above the frequency of the cavity resonance where the notch filter is required, the dispersion is narrow, so a crossover at 3-4 kHz to a super tweeter line source can yield definite benefits in the right setup with the right drivers.

I mention that caveat, because many ribbon tweeters really fall apart below 3 kHz, as regards distortion, and you have to be able to handle the transition region cleanly- for that reason, pushing the crossover to 4 kHz may be necessary.

Chirp sine sweeps such as Praxis or Fuzzmeasure can do are good tools for evaluating the distortion performance at playback levels you want to use.

[gregw]

Thanks for the comments. I experimented with different crossover points between the woofers and the RD75. Stigerik uses 180-200 Hz I believe but he crosses them at 48dB per octave, while I am limited with the Driverack PA to 24dB per octave, so I never attempted to go that low.

On the other hand, the woofers I am using have a measured peak in response at around 700 Hz so I wanted to make sure that I was well below that. So I experimented a bit between 300 and 500 Hz but the differences are pretty tough to discern within that range.

One pleasant surprise was how good the 15" woofers sound. They were very inexpensive and we all know that doesn't always work out. But I have absolutely no complaints. Low saxophone that dips down into the woofer range sounds amazing. It is hard to imagine drums and bass guitar sounding any better. Of course I realize that without a direct comparison it is tough to say that a line of 6-8" woofers wouldn't be even better. Or a pair of Acoustat panels might blow it away.

I was horrified at one point as I was unpacking the woofers and discovered that even though I had asked for dual voice coils, they shipped me half duals and half single voice coils. Since I had them sitting in storage for so long, I never knew. It occurred to me that the two sides don't have to have the same impedance so long as the amp can handle it. So one woofer panel is wired for 4 ohms and the other is wired for 2 ohms. I have them running off an ep4000 with no apparent ill effects.

I am running the RD75s baffle-less. I did add one tweak suggested by Stigerik. I bought some foam pipe insulation and put a length of it on each side of the RD75. I used some "clean-removal" duct tape to secure the pipe insulation to the front of the RD75. It adds a bit of width and depth to the sides of the RD75 so acts as a mini baffle. Big spender--I think they were $.99 for each 8' length at Home Depot

One idea I had that I have yet to try is to use a piece of 2" XPS insulation to carve a rounded baffle that fits over the RD75. They carve surf boards out of foam before fiberglassing them so it could be very nicely done if looks are a concern. I believe John Whittaker experimented with foam baffles next to the RD75. At one point I had the RD75 right next to the woofer panel with duct tape sealing the gap. I had to move things around and never put it back that way. Sometimes the differences are so small that all it takes is a small tweak of the eq and you wonder what was the point.

One thing that made a big difference was getting a house curve with declining response with increasing frequency. There are some articles about it at the REW forums. Almost all of my measurements were done at the listening position. It takes some time with the Driverack PA but almost any response curve is possible.

The ease of making changes is sometimes a curse. I had a Behringer A500 amp that lost one channel. Substituting an Adcom amp with no gain controls meant a couple of hours re-configuring the Driverack Pa.

Regarding ribbons on the high end, I exchanged some emails with someone at Raven who suggested trying a crossover at 10 or 12kHz at 6dB per octave. That is something I would not have thought to try. I don't think I can hear much past 12kHz so I expect the effect might be subtle. I realize that with a single ribbon the response will be optimal at only one listening distance. He also said I would need a 2 ohm stable amplifier because the Raven ribbon is 2.5 ohms. I bought a Behringer NU1000 for $148 from Walmart which has way more power than I need but is one of the least expensive amplifiers rated into 2 ohms. It seems you can never have too many amplifiers when you are using active crossovers.

[Paul W]

It might seem counter-intuitive but, for the same crossover frequency, overlaid filter curves show the LR24 high-pass yielding lower excursion just above xo than the LR48. (8th order response extends lower in frequency before rolling off to the ultimate slope.) Sometimes there are reasons for using high-order slopes, but limiting excursion is not one of them.

[gregw]

Paul, isn't the critical issue with regards to excursion half an octave or a full octave or more below the crossover frequency, rather than at the crossover frequency? BG says minimum 150Hz crossover with second order crossover for the RD75.

From BG RD series product information:

Large rooms with high absorption,high playback levels, or long listening distances usually require a somewhat higher crossover frequency (at least 300 Hz second order high pass crossover). In an average or small sized room with low absorption and shorter listening distance, a 150-200 Hz crossover point may be used. At least third order crossover is recommended for this application

Distortion levels and best crossover frequency must also be dependent on baffle width. It would seem my minimalist baffle (pipe insulation) would be a worst case. Again, from the BG literature:

In dipole or open baffle situations, low frequency extension and SPL output is largely dependent on baffle size and geometry. To achieve a sufficient low-end response, a minimum baffle width of 12" is recommended.Generally speaking,a baffle with irregular or sloped side(s) (e.g.trapezoidal) will yield a smoother response due to the spreading out of baffle edge diffraction effects

Jon, you mention Fuzzmeasure and Praxis for measuring distortion. I don't have a Mac and Praxis is very expensive. Any familiarity with other options? REW has distortion measurements. Is that feature in REW worth exploring? It might be interesting to compare distortion figures using different crossover frequencies and baffles. Holmimpulse is another measurement program that measures THD. I have downloaded it but haven't used it yet.

An interesting tidbit that I didn't know about the RD75 from the BG literature: It is a replacement driver for the Genesis 1.1. I wonder what low end crossover frequency they use between the RD75 and the array of 12" servo woofers. The youtube video gives some nice closeup views.

[Paul W]

Greg,
Don't know if BG is talking electrical or acoustic slopes, and dipole baffles really muddy the water...so I'd rather not guess what they mean.

An enclosed driver with a 2nd order acoustic high-pass maintains constant excursion below xo (-6db point). Less than 2nd order acoustic means excursion is rising even though acoustic output is falling...that can get many drivers in trouble very quickly. Higher than 2nd order acoustic reduces excursion below xo and is far less likely to cause driver problems.

I think "driver friendly" is one of the reasons LR4 acoustic is so popular...relatively low excursion around the knee, coupled with declining excursion below xo.

[cjd]

ARTA does distortion. Free to try if you never want to save any data.

[gregw]

Paul, that makes sense now. My experience with looking at excursion numbers has been primarily related to subwoofers in Winisd where I usually model without a high pass filter. I have modeled the 15" woofers I use with a 50Hz high pass filter and excursion peaks at about 35Hz with a 2nd order filter which is where max SPL is about 6dB down. Of course Winisd does not account for the dipole rolloff.

In reading through the Genesis owner's manual, it says that the low pass to the woofers is adjustable with an initial recommended setting of 98Hz! They recommend fine-tuning the bass crossover by ear saying some "very large rooms may require that you set the lowpass filter very high up to 115Hz." With the 41" wide panels that the Genesis uses, the low frequency performance of the RD75 is enhanced, but I am surprised that they are relying on them down to 100Hz.

By the way Paul, let me add to the chorus of wows in seeing your home theater. I have looked at it before but spent some time this afternoon reading your write-up. Impressive is an understatement. It has me wondering if I shouldn't consider putting the RD75s into a monopole enclosure in the front wall combined with an infinite baffle sub. I just wish I could easily switch back and forth to see which I prefer.

[Paul W]

Thanks for the kind words.

Don't know if yours is a stereo or multichannel rig but, for stereo, my personal preference is dipole to more fully "illuminate" the room. For multichannel, I prefer the clarity of narrow directivity LCRs with wide surrounds for ambience. No right or wrong choice...it's personal preference.

For the next stage of your journey, consider Jon's suggestion to add lines of 7-8" midwoofers...that could free up your woofers for a really good IB sub :T

[gregw]

With this new setup I have been spending more time listening to music than movies. I am just continually amazed at the sound quality. I find myself getting the same kind of excitement and joy from this system that I get from a live performance. It sounds trite but I am re-discovering my music collection. It may be recording quality or sometimes other factors, but some songs sound especially good so I have started tagging my songs in jRiver so I can go back to the best ones. Or to save for demos for friends.

I find lately that I prefer one of the surround sound options on my Onkyo receiver even for two channel stereo. One reason is that I usually get a better stereo image. A problem I have had with this system is that sometimes, especially on some material, the sound is shifted to one side. I want to reach for the balance control, which of course does not exist on the modern receiver. Last night I switched to THX music mode and I thought ahh, there it is, as the image centered and came into focus. It could be that my amp gains and Driverack are not set up quite right, and the surround sound setup corrected it. As I write this I realize that must be the problem. So now it only sounds right in surround mode.

My surrounds are some very inexpensive Insignia speakers that were all the rage a few years ago, picked up on a whim when they were on sale. In fact I was thinking last night that it was surprising that I preferred the surround mode. I thought why aren't the cheap speakers contaminating the sound from the good ones in front?

Sometimes they do. Sometimes sounds are shifted to the surround speakers that should be coming from the front. I prefer not to hear the surround speakers at all except to broaden the soundstage or add ambience. I want to start checking into music mixed for surround.

It is interesting that Jon and Thomas find that the RD75 works best crossed over at 600Hz while StigErik and Genesis have used crossover points that push the lower limits of BG's recommendations. Is it playback level? StigErik says he doesn't listen at high levels and he sits very close. Genesis says their system is not meant to be played LOUD, just at realistic levels.

Several areas for experimentation:

1. Smaller drivers in place of the 15s. What do you think of 12 Aura 6" NS6-255-8A per side?

2. High frequency augmentation. I have the Raven ribbons that I plan to try. Ideally a line of BG Neo3s, or maybe an inexpensive small dome like the Dayton Audio ND16FA-6 5/8" Neodymium Dome Tweeter?

3. Wider baffle around the RD75 to improve lower level output which should reduce excursion and thereby distortion.

4. IB sub. It occurred to me that if I used smaller drivers next to the RD75 then I could re-use the 15" woofers in an IB. They supposedly have 16mm xmax. They only model well in an open baffle or IB.

5. Improved surround speakers, probably using the BG Neo series.

6. Improved acoustic treatment of the area behind the front speakers and insulation stuffed into the joist spaces for bass trapping and to make the ceiling acoustically dead.

My fear is that I will unwittingly make changes that will degrade rather than enhance the sound quality. Some of the changes above will be easy to experiment with, others should result in obvious improvements. It should be fun.

[Paul W]

By most accounts the Aura is a good driver. Xmax/displacement is low so maybe download and use SL's "SplMax1" spreadsheet before you make any new driver and baffle plans. With some mental gymnastics, the spreadsheet might also give you a very rough idea of what the BG75 can really do down low.

[PMazz]

When I played with my RD50s years ago, I tried several iterations using a single 10-12" woofer below the RD. As soon as I tried a line of mid-woofers, I never considered anything else. I eventually settled on a line of Dayton RS180s crossed over at 650Hz with a 4th order LR. Per Jim Griffin's old line array write up, driver spacing at those freqs is easy on the wallet. I use an IB to fill in the bottom end.

[Paul W]

Here's another tweeter that might make a good line...

http://www.parts-express.com/pe/showdetl.cfm?partnumber=275-010

...sold in 10 packs at a very reasonable price & would provide very low CTC. Currently out of stock, but patience usually pays.

[gregw]

I've just been playing around with winisd and splmax. It is tough for any of the smaller drivers to compete with the 15s at low frequencies. Maybe keep the 15s for low bass. Allows dipole bass to 50Hz or even lower. That pretty much is what Linkwitz is doing with his dipole woofers. He uses the Thor below 40Hz then the Peerless dipoles up to 110Hz. Genesis recommends putting their woofer towers several feet behind the main panels. They recommend staggering that distance to improve in-room response. I might have to move my screen out farther into the room, but I have some room to work with.

Then match up a line of RS180s or the Aura drivers next to the RD75 from 110 to 600Hz. It is interesting comparing the RS180 to the Aura NS6. I compared 12 NS6 per side to 8 RS180 per side. That is the most of each that would fit. It might be possible to squeeze in 10 of the RS180s per side if that number can be wired for a suitable impedance. The NS6 is capable of 7dB higher output at 100Hz. Does the lower distortion of the RS180 make up for the lower headroom? Is this strictly an empirical question that must be answered subjectively?

In the infinite baffle area, there seems to be overwhelming support for the idea that overkill and headroom trumps minor distortion differences. What about as you go up the scale?

Add in 4 FI IB3 18s for bass below 50Hz. Add a line of tweeters like the Neo3s (too expensive now?) or another suitable tweeter option. Then the Genesis owners who are looking to upgrade could come to me to set them up.

I certainly appreciate the fine workmanship that goes into something like the Genesis speakers. They are a work of art. But there is something to be said for the brute force method as well.

[gregw]

Paul, I put my email address in to be notified when PE gets that tweeter in stock. Is that the same tweeter that Don Keele uses in the CBT array?

I was trying to determine if it was capable of adequate headroom and I found this calculator:

http://www.doctorproaudio.com/doctor...ladores_en.htm

which indicates 72 of them are capable of 127dB max SPL. If that is right then that should do.

I think I may order one or two of each of the RS180 and the Aura drivers and try to do my own distortion testing. Somewhere recently I read that Linkwitz had always in the past pursued the lowest distortion drivers, but now believes that once a certain point is reached, lower distortion doesn't contribute much. I'll have to try to find that discussion on linkwitzlabs.com.

[Paul W]

Displacement of 8 RS180s is slightly greater than 12 NS6s, so something is amiss in the SPL calculations...with the same baffle width, max SPL should be ~ the same. Unless IB location would be wonky, consider a line of 8 or 9 RS225s with your 15s in the IB.

Don't know that it is the same tweeter Don uses, but it seems very close and a line of them should nicely complement the BG.

[gregw]

Thanks for catching that Paul. I was using Winisd and must have put something in incorrectly. I should have thought about it for a moment with a common sense check. The Dayton has 50% higher XMAX. Surprisingly, the SDs are very close, 126 vs 127.7 considering one is nominally a 6" driver and the other a 7". The numbers are virtually the same in spl_max.

RS225 gives about 6 dB greater output at low frequencies compared to the RS180. (If I did it right.) Comparing 8 TC Sounds 15s to 4 FI IB# 18s, the 15s are within 3dB at 10 Hz. I could use the same panels I have the 15s mounted in and mount them to a false wall. This is looking like something that could be done sooner rather than later.

Sixteen new RS225s cost about half what 4 Fi 18s cost, so this could save me some money too.

[gregw]

The tweeter is labeled part ND13TG-8. It is the same driver used in Don Keele's CBT array, at least according to this thread:

http://techtalk.parts-express.com/showthread.php?232595-OMG-these-are-cool!/page2

That would seem to be an outstanding option for the high end. The price is extremely reasonable, compared to other alternatives.

I am going to CES in 3 weeks. I wonder if Audio Artistry will be there with these arrays. I would love to listen to them.

[gregw]

Just to answer my own question from the Dayton Audio web site:

"Dayton Audio will feature the new and innovative CBT36 loudspeaker at the Venetian Luxury Suite #30-230 during the International CES Show in Las Vegas, Nevada on January 8 -11, 2013."

Cool.

[gregw]

I bit the bullet and ordered 16 RS225 woofers. I did a bit of reading before I did. What convinced me was a comment by Mark K where he said that he thought the RS225 sounded better in the midrange than the RS180 as long as you were crossing over below 1.2 or 1.6KHz. Since the crossover to the RD75 will be 650 or lower, the extra excursion of the RS225 makes it an obvious choice.

It will be interesting to see if the shift upwards in crossover frequency to the RD75 sounds better than my current setup with the crossover at 400. As I was listening to the system last night, I thought I must be a little crazy since it already sounds very good.

As I was demonstrating the system to someone a couple of days ago, we were standing in the back of the room and I noticed that the high frequencies were sorely lacking compared to the sweet spot where I normally sit. So I think a high frequency array may give better sound throughout the room. Unfortunately Parts Express gives no indication when they will have the tweeters in stock.

[gregw]

Good news. Parts Express says the ND13TG-8 tweeter is expected in stock January 8, 2013. That will give me a chance to listen to the CBT array that uses this tweeter at CES before I order it.

Don Keele uses 72 per side in his array in a length of 65". Matching up with the RD75, I come up with 84 per side. Using Jeff Bagby's multiple driver calculator in the PCDC, one way of wiring those would be to wire in parallel 14 groups of 6 wired in series. The net impedance would be 3.2 with 88.9dB/2.83v sensitivity. Is this a reasonable approach? Wiring 168 tweeters does not sound like fun.

It makes me want to think about alternatives. The BG Neo3 is too expensive right now for me to consider a full length array, but perhaps 5 or 6 per side would be worth considering.

I am thinking to make one panel to mount the RS225s, the RD75, and the tweeter array. In keeping with the modular theme though, if I could come up with a simple method of clamping panels together, then it could make sense to make separate panels and make each panel only as wide as needed for the specific driver, with the ability to clamp them together and add wings to the outside.

BG recommends a minimum panel width of 12 inches for the RD75. Genesis uses a 41" wide panel. I assume that the tweeter array must be as close as practical to the RD75.

I was wondering if the tweeter array should be to the outside of the array or the inside. I just looked and Genesis puts the tweeter array to the inside. I'm not sure if it makes much difference but if it works for them then it will probably work for me.

[Paul W]

That tweeter line should be fine.

You may want to download The Edge software and use it to determine baffle width and driver placement, whether the baffles are single piece or assembled from modules. IIRC, Edge will show the baffle peak(s) as well as diffraction.

With a conventional WMT horizontal format, the tweeter line can be switched from inside to outside just by swapping channels.