Beyond the Ariel

If anyone does the 2.5 way system, I'd just caution that you can screw up the response using the extra woofer. It really depends on what acoustic crossover slopes you use for the main crossover and what slope you use for the extra woofer. For example, if I use 4th order Linwitz-Riley slopes, a 1st order butterworth lowpass on the extra woofer screws up the response above the crossover, but a second order butterworth sums pretty nicely. That's just looking at the on-axis response - I'm sure you're doing even more fun stuff when you consider the polar response of the drivers and how the new nulls you may or may not create effect the power response and ultimately the reverberant field you're creating in your room.
 
Good point about phase summation at the main acoustical crossover, Mr. Sheerin. The lower woofer is 45 degrees away from the upper woofer at the -3 dB point, and moves towards 90 degrees as output decreases. Offsetting the lower woofer in the Z plane phase-corrects the on-axis response, while tilting the lobes and nulls of the off-axis response (no free lunch principle).

While we're at it, the floor image has to be included (if the frame of the lower woofer is no more than a couple of inches (50mm) from the floor). That effectively creates a set of three woofers playing together, resulting in a complex vertical polar pattern over the working bandwidth. At 1/2 octave below the main acoustical crossover, there are no less than four effective drivers working together ... many lobes and nulls in the vertical plane, resulting in a substantial departure between total room spectra and direct-arrival spectra.

Hmm. This might explain why systems with 2.5-format stacked woofers, with the lower woofer with a typical 1st-order additional LPF around 150~250 Hz, can sound not-so-great in marginal show conditions. I've even seen expensive commercial systems where both woofers share a common cabinet volume, which doesn't sound like such a great idea.

When considering the complexity of the polar pattern in the absolutely critical 200~800 Hz region, the 2.5 approach kind of looks less attractive. The floor reflection can't really be avoided in the analysis, since most carpets absorb very little in that frequency range.

Maybe a triad of three drivers (MF horn on top, two 15" drivers directly below and side-by-side) creates the most coherent polar pattern at crossover, which is what is most important.
 
Last edited:

ra7

Member
Joined 2009
Paid Member
Not sure it quite works like that, Lynn. If you assume that the woofers are working together because the wavelengths are large and everything sums, then there will be no peaks and nulls. Below a certain frequency this is true and you would get "floor gain." Above a certain frequency, the floor image becomes a floor bounce and it can destructively add with the main woofer, creating peaks and nulls. But you would not have peaks and nulls throughout the range of operation of the woofer. By placing the woofer close to the ground and "getting out" before it starts to get directional, you can avoid most of the floor bounce problem. This means the upper horn has to be pretty large.
 
Maybe a triad of three drivers (MF horn on top, two 15" drivers directly below and side-by-side) creates the most coherent polar pattern at crossover, which is what is most important.

Or the twin pair on the bottom, floor coupling, with a real midrange driver - like an altec 414 centered above, with upper mid horn above. Westake has built some pretty respected loudspeakers in this fashion

Reference Series

TAD was pretty successful with the triad but normally used a bit bigger horn than featured in this thread. Check out the TSM1 and TSM2.

maxresdefault.jpg
 
If there's just a single woofer with the frame close to the floor, I'd expect it to act like a pair of woofers hanging in free space, with the expected set of lobes and nulls. The measurements I've made on the floor-bounce showed less than 1 dB of attenuation in the 200~800 Hz frequency range, so it's nearly a perfect mirror, even with carpeting.

The problem with using a single woofer close to the floor is that the centerline of the horn then has a height of about 26"~30", which leads to an odd too-close-the-floor impression, which does peculiar things to the overall stereo image. I try to get the centerline of the MF/HF radiator in the 36" to 42" range, which seems to result in a more natural stereo image.
 
The measurements I've made on the floor-bounce showed less than 1 dB of attenuation in the 200~800 Hz frequency range, so it's nearly a perfect mirror, even with carpeting.

I use a throw rug placed over a futon, which yields about 3-5 dB of attenuation. Yes, a typical rug on a hard floor is not very effective at all.

I'd also caution talking about woofers in the modal region as if they are in free space (i.e. directivity, 1/2 space loading, etc.) In the modal region the woofers see ALL of the walls, not just the nearby ones. That makes the problem much more complex to consider and simple thought experiments are usually not very accurate.
 
I use a throw rug placed over a futon, which yields about 3-5 dB of attenuation. Yes, a typical rug on a hard floor is not very effective at all.

I'd also caution talking about woofers in the modal region as if they are in free space (i.e. directivity, 1/2 space loading, etc.) In the modal region the woofers see ALL of the walls, not just the nearby ones. That makes the problem much more complex to consider and simple thought experiments are usually not very accurate.

Once again, I agree 100% with Dr. Geddes. At the minimum, in the absence of heavy absorption (several inches of soft material), the most important reflections are the floor (the first reflection to arrive after the direct sound from the drivers), and the wall behind the speakers and the nearest side wall (the second and third reflections). After that, the reflections from the ceiling arrive, followed by the more complex two-bounce paths. And then the rest of the room chimes in with a large array of modes. Impulse response is not good, yet this is what we accept as a "normal" acoustic for listening to music.

This throws simplistic concepts of a fixed bass-lift compensation for "baffle step" out the window. It's also clear that speakers that are dipoles, or have some degree of directivity thanks to bass horns, are going to interact with the room in a different way than a direct-radiator that is effectively omnidirectional below 500 Hz.
 
Last edited:
What is the actual problem you're having with your subs?

I have a very similar system (sealed 15" subs, JBL 2225's in sealed boxes, LC horn crossed over at 800Hz, etc.). The best method to integrate a sub with a main speaker if you're going to do stereo subs in my experience is to measure the sub and speaker in an anechoic environment (outdoors ground plane would work) with the correct relative positioning and work out the crossover so it sums correctly (get the same slope and crossover point on each driver, then make sure they're time aligned). It sounds like you don't know how or where your mains rolloff, so integrating correctly with a sub is basically a guessing game for you. I tried to do it in room first and found it impossible. Once your system is working correctly, put it in your room, measure again at the listening position (averaging multiple locations is nice) and see how much room gain you have. EQ the entire system (not just the sub - you'll potentially screw up the phase response at the crossover) from there to get your desired in-room response curve. I usually have something that's flat below 100Hz to as low as I can get it to go, although based on what level you like to listen at I suppose you could change that. In my case, room gain almost perfectly matches the LF rolloff of a sealed box with the 15" drivers I'm using, and I get nice response to below 20Hz from a cabinet that only went down to around 40Hz anechoic.
 
Member
Joined 2012
Paid Member
Altec Midwoofer’s Fc and/or Slope: Makes Sub’s Plate Amps Too Audible?

Lynn, It's been a long time. I built your design using the GPA 416 alnico , Radian 745neo BR , and (2) Rythmik subs. I originally built the posted crossover and later with Dave Slagel's measurement assistance went to a 2nd order low pass and first order high pass( Jupiter copper foil) with a Slagel autoformer (-18 db at 700 hz )I'm very happy with the speakers and the synergy with the Line Magnetic 518 ,845 amp . I primarily listen to vinyl. However, when going digital I'm still using the NM 24 with Mundorf Silver Oil caps, Amprex tubes and the Monarchy AC regenerator as suggested by Mr. Poon. Thanks again to You and Gary and all others that helped create this excellent speaker. :)Greg

GregOH1 that's a wet dream system!

The subs are the Rhythmic FG12/sealed/audiophile subs listed on the site. The amps allow for numerous tweaks not always found on your typical sub such as damping, rumble filter etc. Check out Rythmik for all the details. I 'm happy with subs set at 50Hz/ 24db, 20 hz cut, High damping. I also have extensive GIK room treatments . Sealed/ no eq/ beryllium driver (I also have the standard 745 and yes there is a difference)

I understand the desire not to use a 1st order HP. However with a Slagel autoformer it can work.Hey, it doesn't hurt to try- More info available on the Intact Audio site/ forum.

I designed my sealed upper cabinets for use with the AE TD15M's, but Lynn talked me into trying the 416's; the parameters were similar enough to make this plausible. I re-cut the driver openings for the Altec frames, and tried 416's as well as 515's. Much to my surprise, it was the 416's that really brought the music to life. And I preferred both to the TD15M's.

I have made a change in the bass system that I haven't yet mentioned, though. The plate amps have been replaced by a modified (Big Sky Audio) Parasound HCA-1500A. Low-pass filtering and room correction for the subwoofers are now handled by a DSPeaker Anti-Mode 8033SII. I have two of the Anti-Mode units, actually, and have been experimenting with dual-mono vs stereo bass. Surprisingly, DSPeaker recommends the dual-mono configuration in two-subwoofer systems. My results so far suggest that they have a point.
Gary Dahl
Gary has provided lots of invaluable info and guidance as I pursue completing his speakers; also for helping me make the best amplifier choices among several designs.

But though respectably digitized, uncompressed WAV files of ripped CD tracks of many of my favorite recordings exhibited one or more kinds of original analog-based damage (e.g. clipping distortion, phase asymmetries, hard panning) through his now even more mercilessly revealing system. Accordingly, Gary rather emphatically suggested use of the industry grade restoration software that I was later lucky to find and buy on sale last year. https://www.izotope.com/en/products/repair-and-edit/rx.html And a much earlier version of Izotope RX demonstrating its Declip module for removing distortion. Izotope RX |
iZotope RX5 |
Phase Correction Plugin ? InPhase | Waves
Waves InPhase |

But before riding out the software's learning curves (while penning a lot of annotations to its manual downloads), I must now confront what looks like an insurmountable problem involving the sound quality of the subwoofer plate amps which I will likely be locked into using. However, these are predicaments which both Gary and GregOh1 had cleverly and/or luckily escaped, though by different means. I hope that I’m very wrong but it doesn’t look to be an equally successful fix in my case.

I, or rather my builder Jim Salk Salk Sound cloned a pair of Gary’s GPA Altec 416-8B midwoofers in his original sealed 3 cu ft cabinets. The box cuts the 416 off below 70Hz, however sharply I don’t know.

Below the Altecs Gary uses his pair of 15” Acoustic Elegance woofers. Unless mistaken, I believe Gary had long driven those woofers with plate amps, though I don’t know the model. But throughout all of our chats I never recalled Gary lamenting his use of these plate amps. So I had assumed that the Fc he had dialed in with the plate amp’s crossover was low enough, its the slope sharp enough and/or the plate amp’s sound quality was good enough for there to be no audible coloration by these amps of the main speakers and Amity 300B p-p mono blocks.

After many months of health, family, financial issues and system-related fact finding (though possibly still short of some important facts), I am finally ready to pick up where I left off with Jim and resume the build. But visiting this thread last month did I see that Gary has recently swapped out the plate amps with his upgraded Parasound HCA-1500A stereo power amp. http://www.diyaudio.com/forums/multi-way/100392-beyond-ariel-1442.html

And though Gary once expressed reluctance about adding a second layer of digital conversion to his system, the pair of DSP units he installed obviously includes the crossovers needed between the Altec midwoofers and the subs.
http://www.diyaudio.com/forums/multi-way/100392-beyond-ariel-1442.html

That Gary wanted to see if the virtues of room correction would be more noticeable than any DSP converter artifacts seems clear. And if hours spent with the system now has it actually sounding far better without the plate amps, how fortuitous that he could implement such an elegant solution.

But what most concerns me about why Gary ditched the plate amps after all this time was if it was done primarily because the extensive upgrades that he recently completed on the Amity 300B amps had made the plate amps’ presumably lesser sonics too noticeable, than they have those of the upgraded Parasound amp.

Meanwhile, I’m locked into the pair of Jim’s version of Brian Ding’s Rythmik 12" sealed servo subs. Jim uses the same 12” Rythmik aluminum or GR Research paper drivers and the Rythmik plate amp, but- better built, braced and slightly larger cabinets.

The sub’s servo boards are integral to Rythmik (Class AB and D) plate amps-not available as separate modules-so a standalone power amp cannot be used. And the plate amp schematics, like for Rythmik A370-XR2 Class AB amp that I chose to have installed, are not available at the website. Only a general description of its functionalityhttp://www.rythmikaudio.com/download/XLR2_sealed_quickguide.pdf and its design features Rythmik Audio • Subwoofer plate amplifiers are disclosed.

However, I did learn from someone at audiocircle.com who offers upgrades that:

"There's no capacitors in the signal path in these amplifiers so you've got no need to pay for expensive capacitors to get the amp up to snuff in that department.

The opamps, first off the TL072 are not operating the purpose you'd probably suspect, and are not a determinant to performance. It's not a signal device for amplifying the signal in these units I don't believe. The OPA2134 is good because it's very low noise, lower than many well known audio opamps. But for fullrange capabilities it's not as good in tone etc. For bass that's not really relevant and the servo aspect sort of robs it getting the chance to be anything but good. However I'll give you a quote on what it'd take to replace it with discrete units or if I find a find a direct replacement.

I don't believe it's to your benefit to have the work done because if they're not easily accessible on the back it really adds a lot for maybe no reason to price. The fact that the discrete units have to "hang" off the board means I have to make supports or they'll break the solder joint or tracing off of the main board.”

GregOh1 and virtually all Rythmik sealed sub users I’ve encountered claim they deliver excellent performance in most if not in all respects, particularly for music.

But Greg, who may have made his Altec midwoofer sealed cabinets bigger than 3 cu ft and/or crosses the 416s lower than Gary and/or I can, has his Rythmik subs playing at “……under 60 Hz”. :: View topic - 1st order/ Azurahorn
http://www.diyaudio.com/forums/multi-way/100392-beyond-ariel-1437.html

Thus, as the ear’s sensitivity rapidly declines with frequencies below 200Hz
https://en.wikipedia.org/wiki/Equal-loudness_contour , the sonic signature of Greg’sRythmik plate amps might therefore be far less noticeable than his very sweet sounding Line Magnetics i518 845 SE main amp. http://www.diyaudio.com/forums/multi-way/100392-beyond-ariel-1436.html

Indeed, having subwoofer amps-particularly the arguably compromised design of most plate amps-handle less of the more audible bandwith-intentionally or not was what may have strongly contributed to Greg’s and perhaps Gary’s now total satisfaction. However, in pursuit of this apparently very important sound quality goal, as explained, as best I could, I may not be so lucky.

Instead, what might the final system midbass sound quality resemble, due to what Fc and slope of the Rythmik plate amp crossover AND/or that of the main speaker’s crossovers blend with Altec midwoofers-which in Gary’s 3 cu ft boxes only extend to ~ 70Hz?

That is, what, if anything, can be done to best maximize overall sound quality and minimize plate amp sound in my case?


1.) Should the Altec midwoofer be in a sealed box larger than 3 cu ft?

2.) Build Gary’s crossover as is? Or must it be appropriately modified to produce the right Fc and slope, not necessarily for the Altecs to play lower than Gary's 70Hz cut off, but so that the Rythmik plate amp can cross the Altecs such that the subs start playing "…..under 60 Hz.”-as Greg's do-thereby making any plate amp coloration much less audible?

3.) Keep Gary’s cabinet sizes and crossovers but upgrade the plate amp’s circuitry, done however cost-effectively? Indeed, has anyone here done or knows of someone who’s had success modding a plate amp to sound way better than it did in its stock form?

This Rythmik plate amp design may or may not be better sounding than most. But would it be worthwhile to swap out that TL072 op-amp and use Mundorf EVO oil caps for op-amp feedback, inter stage coupling and supply decoupling in the crossover, LP filter and driver output stages? Please suggest any sensible plate amp upgrades.

Jim Salk’s cabinet interiors are larger than Brian Ding’s stock Rythmik F12 cabinets, so they might accommodate any added outboard circuitry.

4.) Build another pair of either sealed or open baffle Rythmik subwoofers. Those drivers would have paper diaphragms-which Rythmik also sells, supplied by open baffle sub advocate Danny Richie’s GR Research.

“The paper driver is lighter than our standard aluminum drivers and has a more extended response. It can handle a higher crossover point, however this is only significant when crossing above 80Hz. “

Rythmik offers sealed 12” paper driver servo subs. GR Research offers servoed dual 12” paper drivers, in an H or W frame OB enclosure per side. Those OB subs also use the same Rythmik plate amps but include a shelving filter.
Rythmik Audio servo subwoofer 12" F12G • Features
Rythmik Audio • Servo subwoofer products

However, I'd then be back to questionable Rythmik plate amp sound quality that would again pose the problem of how to cross them low enough to avoid being as noticeable as the main system amp. And if the boxes need to be > 3 cu ft and/or I’ll need Slagle autoformers and/or inductors different than Gary’s, I'd be in for some hefty losses.

OTOH, what chance, however slim, that a paper subwoofer diaphragm may somehow blend better than the standard aluminum driver in my Rythmik subs crossing to take over below the Altecs 70Hz cut off, even if both use subwoofer drivers use the same plate amp?

Or would I somehow cut my losses better by using non-servoed paper driver subs with standalone power amps and active crossovers. http://www.firstwatt.com/pdf/prod_b4_man.pdf

Then, assuming that Y-connectors are NOT sound quality taboos IF correctly constructed and electrically compatible to use between typical line level input and output impedances and output voltages (??), my DAC’s unbalanced outs would feed my main amp and the stereo standalone sub amp for non-servoed (paper cone?) subs. And the DAC’s balanced pair would feed the plate amps of my other Rythmik subs. Those aluminum drivers could be crossed lower to handle, say, from 40Hz to 20Hz.


Thus, would those plate amp sonics be much less noticeable at those much lower frequencies-and that using two pairs of subs also minimizes standing wave modes in my 14 ft x 19 room? http://www.harman.com/EN-US/OurCompany/Innovation/Documents/White Papers/multsubs.pdf

Please advise on these crucial subwoofer format (e.g. sealed or OB), diaphragm and amplifier design and application solutions.


As Gary’s surely preparing for Easter concerts and other musical festivities, he likely hadn’t time to reply to my earlier questions about his Amity amp upgrades and how the system sounds now.

But at this point I’m hoping that readers may have enough info to help steer the quality of the Altec 416’s 700Hz to 70Hz response and the sonics of my main amp (Gary suggested a well-designed 300B SET amp or mono blocks) out of harm’s way of the plate amps-as Greg was able to do, albeit with different cabinets and/or crossovers and no DSP.

Please advise accordingly for best performance, even if it means having to mod or sell off some parts and/or gear.
 
Member
Joined 2012
Paid Member
Lynn,
Thanks love purple. The subs are the Rhythmic FG12/sealed/audiophile subs listed on the site. The amps allow for numerous tweaks not always found on your typical sub such as damping, rumble filter etc. Check out Rythmik for all the details. I 'm happy with subs set at 50Hz/ 24db, 20 hz cut, High damping. I also have extensive GIK room treatments . My wife tolerates my madness .

Fred
Sealed/ no eq/ beryllium driver (I also have the standard 745 and yes there is a difference)

I understand the desire not to use a 1st order HP. However with a Slagel autoformer it can work.Hey, it doesn't hurt to try- More info available on the Intact Audio site/ forum.
Greg¸ are your Altec midwoofers cabinets vented or sealed? And 3 cubic feet, like Gary’s, or are they larger?

Here’s Gary’s crossover attached.

Are your Slagle autoformers and inductors much different in value and/or composition, where even with the appropriate capacitor and resistor values, it couldn’t come close enough to duplicating the Fc and slope that you chose for crossing between the Radians and Altecs?

What do you like better about the Radians’ sound without EQ?

And what, if any performance aspect (s), might you have willingly sacrificed by omitting the EQ?

Was opting for no EQ partly based on your amplifier choice? Might you have chose to use it if your amp were 300B SETs monos? Perhaps as good or better designed than these? Coincident Speaker Technology Frankenstein II Tube Monoblock Amplifier | The Absolute Sound

You had mentioned that you use “multiple” Rythmik 12" sealed servo subs.
:: View topic - 1st order/ Azurahorn If you were using two or more pairs then, why are you apparently only using only one pair now?

Wouldn’t using two pairs of (sealed and or open baffle subs) more likely produce a flatter bass response, especially in relatively small rooms, like my 14 ft x 19 room?

As the Rythmik subs have no DSP, what, if any, software-based measurements might you have done for then selecting the GIK Acoustics products, other than perhaps using this form? Acoustic Advice Form

Congrats on your build and thanks for posting!
 

Attachments

  • Gary Dahl's Crossover.jpg
    Gary Dahl's Crossover.jpg
    71.8 KB · Views: 739
What is the actual problem you're having with your subs?

I have a very similar system (sealed 15" subs, JBL 2225's in sealed boxes, LC horn crossed over at 800Hz, etc.). The best method to integrate a sub with a main speaker if you're going to do stereo subs in my experience is to measure the sub and speaker in an anechoic environment (outdoors ground plane would work) with the correct relative positioning and work out the crossover so it sums correctly (get the same slope and crossover point on each driver, then make sure they're time aligned). It sounds like you don't know how or where your mains rolloff, so integrating correctly with a sub is basically a guessing game for you. I tried to do it in room first and found it impossible. Once your system is working correctly, put it in your room, measure again at the listening position (averaging multiple locations is nice) and see how much room gain you have. EQ the entire system (not just the sub - you'll potentially screw up the phase response at the crossover) from there to get your desired in-room response curve. I usually have something that's flat below 100Hz to as low as I can get it to go, although based on what level you like to listen at I suppose you could change that. In my case, room gain almost perfectly matches the LF rolloff of a sealed box with the 15" drivers I'm using, and I get nice response to below 20Hz from a cabinet that only went down to around 40Hz anechoic.

John

It seems to me that doing a crossover in a free field and then bringing it all into a room and redoing it is rather unnecessary. Why not just do it all at the same time in the room since the room is going to change everything anyways. (I assume that we are doing all this in the modal region of the room.)
 
John

It seems to me that doing a crossover in a free field and then bringing it all into a room and redoing it is rather unnecessary. Why not just do it all at the same time in the room since the room is going to change everything anyways. (I assume that we are doing all this in the modal region of the room.)

Earl,
That's what I've done in the past - do everything in-room. I totally agree things will change in the room (80Hz is definitely modal in my room). But I still found it much easier to set up the crossover on my current system and compensate for delay in an environment where the room was not screwing up the response. To be clear, I'm putting my subs right behind my mains which are pulled out into the room. So my base assumption is that the transfer function between the speakers' location and the listening position is at least similar around the crossover for both the mains and the sub, and thus the effect of the room on both will be about the same around crossover. So a crossover I setup in an anechoic condition will still sum correctly in-room. If you're going to stick the subs in corners, then I agree it could be better to do it in the room. I definitely think you need to eq the complete system in-room. I just think that's much easier when you already know the various drivers in the system are summing correctly.
 
John, I'm not saying that what you do won't work, just that I don't see it as necessary. Once the in-room response is optimized, what difference would it make that it was once good in a free field. As a starting position, it might make sense for faster convergence, but in the end the results should be the same either way.

but getting to the end measuring in room only may require treating the room
 
diyAudio Moderator
Joined 2008
Paid Member
To be clear, I'm putting my subs right behind my mains which are pulled out into the room. So my base assumption is that the transfer function between the speakers' location and the listening position is at least similar around the crossover for both the mains and the sub, and thus the effect of the room on both will be about the same around crossover.
In this case I see your point, but I question the need to use subs in this way. Any ordinary woofer (used with the mains) can do 70Hz and better, covering vocals and moving well into the modal region. Subs, with all that we know these days can, and probably should be placed somewhere else.