Planar Speaker Asylum

Welcome! Need support, you got it. Or share your ideas and experiences.

Return to Planar Speaker Asylum

Planar Speaker Asylum

RE: Apologies for the long post, answers and some speculation on the ribbons

64.252.13.238

In Reply to: RE: Apologies for the long post, answers and some speculation on the ribbons posted by Satie on September 09, 2010 at 01:22:35

Thanks, that was an unusually informative food fight. Among other things, I had no idea that ribbon corrugations helped to reduce torsional motion, although in retrospect it's obvious.

The sonic differences between the true and Kapton-backed ribbons don't surprise me, they seem very much in line with the generic differences between the technologies.

Still no idea how and to what degree Graz has solved the torsion problem, but my interest was essentially academic, anyway. Though I gather he could run off a true ribbon midrange if you were wanted one. While reading more about the RD-75, I came across an interesting project by someone who put an Acoustat 1.1 on either side of the RD-75 and a line of leaf tweeters. He crossed the RD-75 over at about 650 Hz, which kind of makes me wonder whether his setup doesn't suffer from lobing. Still, he said it's the equal of $75,000 speakers, and I believe him.

I'm not sure what to think about the RD-75 -- I'm tempted by that 150 Hz crossover, which would allow me to run it through the midrange and much of the treble. OTOH, I've come across criticism of its performance below 500 Hz, and it does beam. I understand that Wilson put three traces on it, presumably so that the highest frequencies can be limited to the center trace for better dispersion. I'm wishing that BG made their drivers that way, particularly the Neos 8 and 10 (assuming, of course, that you could get your hands on the Neo 10). In an array, the impedance variations of the various trace lengths wouldn't present a problem (I'm still wondering how Magnepan handles the lower impedance of the supertweeter trace in the 1.7 -- thinner foil?).

Follow Ups:

RE: Apologies for the long post, answers and some speculation on the ribbons - Satie 03:35:32 09/13/10 (4)
I saw a couple of attempts to solve the problem with acoustats as midbass/mid, which is really easy to do with an active crossover. The particular leaf tweeter contraption along the RD75 and acoustat panels I thought was a little much. But hey, probably got it right, no reason it shouldn't work.

The 1.7 does use thinner foil for the tweeter, and a resistor.

Re RD75. The bottom of the RD75 is really limited by dipole cancellation at frequencies with 1/2 wavelengths significantly below its width. Which is why Genesis place their beefed up RD75 in a giant baffle, and BG offshoot Wisdom put theirs in a sealed back box - wince. VMPS, in their RM40 etc. do the same with their Neo8 like drivers, and on the V baffled top model they give plenty of baffle for the back and a shallow wave guide for the front.

The idea of using narrower portions of the driver for high frequencies is exactly what BG did with their Neo8 PDR and Neo3 PDR. Though what they did is via damping of the sides of the diaphragm rather than an electric filter circuit. Works very well and there seems to be no problem up to 14-15 khz as opposed to the plain Neo8 that becomes progressively beamy-er as you go up from 5 khz. But you give up quite a bit of output, lower power handling, and max spl. The damping also reduces the main cavity resonance.

I have this idea that came from playing with horns on the Neo8, Horn loading the entire structure with concentric line source horns - Doing this with the tympani in particular. The bass gets the deepest horn, the midrange a shallower one, and the tweeter a shallower one yet. The back waves would have to be attenuated with some acoustic wool and possibly a solid structure, to divert the remaining backwave sideways.
Of course, you get giant cabinet sized speakers, but it might actually work.
- RE: Apologies for the long post, answers and some speculation on the ribbons - josh358 10:29:49 09/13/10 (3)
  The problem I had with the Neo-8 PDR for this application is that below about 600 hz, the frequency response falls off by comparison to the Neo-8, reaching - 6 dB below 400 Hz. I assume this is mostly a consequence of the tapered drive. With a segmented coil approach, that wouldn't occur, since the diaphragm would be driven across its entire width at low frequencies. Dispersion would be more uniform at the high end as well. And it would be possible to maintain the high frequency response and dispersion in a physically larger driver that could be crossed over at 80 Hz.
  
  I really don't know why planars, both electrostatic and magnetic, haven't made more use of this approach, which dates back at least to the original Quad. Cost issues, I suppose.
  
  I know that Genesis uses the RD-75 down to (IIRC) 96 Hz, and it looks like it doesn't require a Genesis-sized baffle to use the RD-75 to 125 Hz if you use a high order crossover. In fact, the RD-75t seems to have fairly good low frequency performance even without a baffle, and to be good to 125 Hz with a 15" baffle. My concerns with the RD's low frequency performance stems from the experience of a couple of guys who experimented with it, and found that it sounded better if crossed over at 650 Hz, and with a former Genesis employee who said that at some frequencies below 500 Hz it had harmonic distortion as high as 50%. Still, I may be making too much of these objections, given that it's apparently good enough run full range for Arnie Nudell's flagship product.
  
  I'm also not wild about the 75's waterfall plot, particularly when compared to the Neo-8's. In just about every respect, the Neo 8 is a better choice, doubly so insofar as its cheaper, more efficient, and has higher output. If I didn't have to put the woofer panels off to the side of the room, the Neo 8 would be a no brainer, but I think I'm more likely to get away with that with a crossover of 125 Hz than a crossover at 200.
  
  I'd be leery of horns and waveguides, even small ones. It's hard to make a horn that doesn't color the sound, and even if you get everything right, you're stuck with time smear. So I suspect that the purity of the sound would be compromised. Doubly so if you hang damping material on the back, which will selectively absorb the high frequencies, or build an enclosure, which will almost certainly make the speakers sound boxy. On the other hand, if you increased the efficiency by 2x in front, you'd have a cardioid radiation pattern, which is quite nice for speakers, according to Arnie Nudell better for imaging than a figure eight.
  
  I'm thinking that if you need more output in the low bass, an infinite baffle planar would be ideal. If it's not practical to cut a hole in the wall (or mount the panel in a door), score another pair of IVa's and make your existing bass panels isobaric, or join four of the woofer panels and use them as a sub. Or even extend your IVA's on the woofer side to make them five panels, feeding 3/4 bass to the outer three panels and 1/4 bass below 80 Hz + midbass to the woofer panel closest to the mid/tweeter.
  
  In the mids, sound reinforcement ribbons might do the trick. They're expensive, though. Another possibility would be to score a pair of 1.7s and set them up in a (1.7)-(Neo8+Ribbon) configuration. Then join the bass panels into a four panel subwoofer. Or maybe put them on the other side the 1.7's, you'd be extending the baffle width even more than in the sub configuration. The 1.7's would allow you to use a higher crossover point on the Neo-8's, increasing their output capability, and would I'm guessing sound better in the midbass than the IVA's.
  
  Not sure what you could do about the tweeter, other than replacing it with something more robust. Though with a crossover of 9 kHz, how much energy does it have to handle?
  - RE: Apologies for the long post, answers and some speculation on the ribbons - Satie 19:31:41 09/13/10 (2)
    I tried 3" to 15" 90 degree waveguides at 4 depths/sizes on the Neo8 and actually got a richer and smoother sound and complementary colorations. The problem was that it ended up standing in the way of the tweeter output so it was not useful that way so long as I am not setting up the tweeter in front of it. Besides, it created a 12-15 db sensitivity difference with the rest of the speaker rather than the 5-6 db I had before.
    
    The cardioid pattern seems like a cool way to avoid side reflections - BTW I have the speakers setup roughly 60 degs to the sidewalls and 30deg to the front wall so I get rather late front wall reflections coming as secondary reflections from the sidewalls and secondary reflections through the 1-1.5 ft slot between the bass panel's outer edge and the wall- those are secondary reflections from the front wall to the side wall.
    
    The 9khz is probably going to go back down to the 5-6 khz range if I can do it without losing the cool transient coherency I managed to get and the "walk in it" soundstage that I am sort of getting now and had not managed to get from the Tympani before - but had managed with the Vandersteens after a few months of tweaking.
    
    But yeah, with crosses at either 6 khz or 9khz leaves barely 2 octaves - if that. So it is possible to do less.
    
    I don't really need more bass, as 2000+ watts per channel go a long way towards filling out the bass out of the panels, it only loses power in the near subsonic and below. Still plenty. Better than a Revel Ultima Studio (ha ha... that was a nice surprise - you had to see my friend's jaw literally drop when that contrabass note came out, I still can't get over it). Now that the subsonic filter is out I have "what the @%$*" every listening session as stuff comes out at the bottom octave I did not even know was there or suddenly gets me with heavy output.
    
    The Neo 8 when used in a line array of 8+ is just so much better suited for lower mids to 200-250 hz, it is a no brainer to use it rather than the RD75 (I think the driver's FR curves just scare the Genesis folks), and that waterfall plot is outstanding. Anyway, that is why it is in my setup. BTW current crossover is 18db Butterworth LP at 240 hz. 200 hz 1st order and 10.5 khz 1st BP, and 1st 8.5 khz HP. The mids acoustically roll off on the bottom from a +3 db peak at 300 hz at 2nd order and drop from a +3 or 4db peak at 6-7 khz they roll off at apparently 1st order steepening slightly further up.
    - RE: Apologies for the long post, answers and some speculation on the ribbons - josh358 17:46:43 09/14/10 (1)
      That isn't bad at all. I'm thinking I might be able to get by with the bass panels off the side and crossed over at 200 Hz. It's hard to tell, sometimes I notice the position of a sub crossed over at 300 Hz, sometimes not, even if it's off to one side.
      
      There's a very interesting graph in Toole's paper that shows the audibility threshold of audibility of reflections as a function of delay. It looks like below 20 msecs or so, reflections have to be down about 17 dB to be inaudible. I'm thinking that a cardioid pattern would allow you to achieve that with both the side wall and front wall reflections and minimal treatment.
      
      I'll be interested to see the results of your crossover experiment. How far is the centerline of the ribbon from the centerline of the Neo 8's? I'm bothered by both the shift in imaging due to the lateral displacement of line source drivers, as well as the shift in vertical position due to the lateral displacement of point sources. But with the space constraints I have and the available drivers, I can't figure out how to make a laterally symmetric configuration without screwing up the polar response at the crossover point, never a good thing. About the best I've been able to come up with is putting an MGMC1 on either side of a 2.5R ribbon and running the quasi ribbons from 250 to 1K, but I suspect the quality advantage of the Neo-8's would outweigh the advantage in imaging. Anyway, it will be interesting to see if lowering your crossover point has a negative effect on the image, and at what point that happens. If anything, it seems to me that from an imaging perspective you'd want to go the other way, since there's important directional information at about 11 kHz. But the flip side of that is that polar response seems to be important to soundstage depth, so . . . I'm not sure at what frequency that effect falls off.
      - RE: Apologies for the long post, answers and some speculation on the ribbons - Satie 01:39:46 09/15/10 (0)
        
        In Reply to: RE: Apologies for the long post, answers and some speculation on the ribbons posted by josh358 on September 14, 2010 at 17:46:43
        
        I don't know which item you are referring to as "not bad at all" as all issues in the prior post refer to the mid/bass crossover.
        I find that you can tell low frequency direction from the event aspect of their initiation - which we localize perfectly even if we can't hear the actual tone. Which is why I always suggest to people that they get two subwoofers - not one.
        
        Thanks for the figure on reflection audibility threshold. I remember that the 20ms rule is the threshold for separating "events". so if I can manage the reflection's path to go beyond 20 ms, or decay that much (17 db) then I effectively controlled them and the reflections become "reverberant field cues". So all you need is to have the speakers over 10 feet away from the first surface behind them. So that the distance back and forth would be over 20 feet, and thus have a 20ms delay for the straight back wave, and all the side bounces would have a longer delay than that. So if the mid/treble is 30 deg from the front wall with the tweeter's corner 6 feet away from it, then I get 7.4 feet to the first reflection, which bounces away from the listening seat axis to the side wall and from there back to the speaker - in my room the path comes to 17 feet, so I would just need to move the speakers just about 1 foot further into the room. And people with narrow long rooms can position their speakers straight ahead 10 feet from the front wall, or toed in a little and have the speakers 9 feet from the front wall and get no event smear.
        Will try that out.
        I don't think you got my point, the locations of the line sources are not that important. They just need to be:
        1. located symmetrically to their counter channel line source in the room,
        2. located more than 15 deg from the listening axis to either side (I have not managed to get good stereo at separations narrower than that)
        3. Be roughly the same distance from the listening seat and in phase or orthogonal (makes for better transient behavior). OR
        4. at least be in phase from the listening seat (by placement or electronic delay).
        5. Radiate towards the listening seat from the front
        I have had the speakers set up insie and outsie and got nearly the same soundstage I put them one in front of the other at 3 foot distance from each other and played with different positioning of the mid/treble panel just keeping the distances 3 feet different. Outside of wall effects I got substantially the same soundstage.
        At the current crossover, there is a 1.5 inch wavelength. When the tweeter was crossed over in third order (same Fc), there was a serious head in a vice problem and things shifted if I tilted my head one direction or another. With the broad overlap of the 1st order crossover there is no problem, theory be damned. I wondered why Magnepan keeps the QRs first order on the tweeter. Well I see now. I also see what PG was after (even if he didn't know himself). By my calcs, this should be roughly transient friendly if not transient perfect.
        The point of lowering the tweeter HP was in part to get the 11-12 khz area with less phase. The line array drops rather steeply around that area, so I wanted to give it an extra push like magnepan does with the second order LP for its midranges . since with a 1st order electronic slope on top of a 1st order acoustic I get a second order - but less phase.
        Re the horn loaded idea, the lobing effect is interesting and makes me more curious to experiment further at some point, since getting a 3db gain with a horn is not a big deal.
        Looking at your confines, I think there should be no downside to having a line array of Neo8 in front of a bass panel, and crossing over at a frequency that works for you to make the distance between the panels right for positioning in your room, or use a Behringer DCX2496 on the bass only and adjust the delay for the distance. I am inclined to go that way myself if I can't make a PLLXO work 2nd order or quasi 3rd (3rd with low Q). The ribbon tweeter (I am assuming you want one) can go next to the Neo8 array like the Tympani IV and IVa as I have it. But don't fear putting the ribbon in front of the Neo 8 line.
        BTW there are about 6" between the center of the neo8 midrange and the tweeter. At the current crossover it is a 5 wavelength distance. Using my Marchand crossover at 5khz LR4 it is still 3 wavelengths. There is one cool thing to remember, that you are getting output from the entire array both on the tweeter and the midrange, so that lobing does not work as it does in point source drivers. There is the interaction of the center of the midrange and the center of the tweeeter, but most of the output comes from the rest of the array where the distance between say the output from the midrange area 5 feet up interacts with say the tweeter output 2 feet up (ref. floor). Meaning that there is a broad range of interaction distances rather than the narrow range in point sources, so while the average would be like a point source at the center of the mid or tweeter, the actual cancellation and construction are very broadly distributed - to the point where you can ignore them if the crossover slope allows it - the shallower the slope the broader the interaction. am I thinking right? is this sensible?
        
        Edits: 09/16/10

Post a Followup:

FAQ

Post a Message!

Forgot Password?

Moniker (Username):
Password (Optional):
	Remember my Moniker & Password (What's this?) Eat Me
E-Mail (Optional):
Subject:

Message:	(Posts are subject to Content Rules)

Optional Link URL:
Optional Link Title:
Optional Image URL:
Upload Image:
E-mail Replies:	Automagically notify you when someone responds.

To view your new posting or follow-up, click on the RELOAD or REFRESH button on your browser.

This post is made possible by the generous support of people like you and our sponsors:

The Cable Cooker