If I were to bypass the segmentation resistor network in my speaker, it would sound very much like Roger Sanders' model 10 speaker. Especially since I followed Roger's guidelines for matching the woofer to the panel.
But whereas Sanders speaker is purposely designed to project a narrow/planar wave front, my symmetrically segmented wire panel is designed to project a curved cylindrical wave front. It's dispersion pattern is actually wider and smoother than a curved Martin Logan panel.
A cool option with segmented wire panels is the ability to configure them with selectable wide and narrow dispersion modes. This is quite easy to do by merely installing a multi-pole switch to defeat (jump over) or enable (leave engaged) the segmentation resistors. I actually did this with my previous welding rod panels; although I didn't bother adding the switch feature to my current wire panels (I may yet do that).
Switching between modes is a bit of a hassle though because the panel's response is naturally flat with the segmentation resistors engaged (wide mode) but has a rising response which requires EQ'ing back to flat when the resistors are defeated. The EQ'ing hassle is reduced significantly if you have a DSP with memory presents for the EQ's. Since the mode switch is on the high-voltage side of the transformer, switching modes has to be done with power off, lest arcing destroy the switch.
Last edited:
While curved esls can have better dispersion it doesnt mean you get a 'polluted first arrival'. Certainly not when keeping them away from a sidewall.
It has been suggested that wide dispersion speakers have better imaging as the indirect soundfield is more balanced and becomes less obtrusive as a consequence.
Also many flat panels use electronic methods to increase dispersion, for example electric segmentation (audiostatic), delay lines (quad)
Most speakers will sound and image better when secondary reflections are minimized IE. Pulled away for side walls etc. And imaging cues come generally from higher frequencys that paint the outline. A 1” dome “wide” dispersion tweeter in comparison to 9” wide curved panel producing the same signal is physically impossible IMHO to produce tight sound image outlines coming off a piece of curved 9” Mylar. A small physical point source just does imaging better overall , ala, mini monitors, so wide dispersion is not an enemy in and of itself.
Attempting to make a big sheet do the same thing and get the same sound field
Illusion is not what I personally hear.
At our Newport hiend show a few years ago ML had their new Neolith playing and right down the hall the Sanders system and it was interesting to go back and forth listening.
The Ml sounded big and vague in outline textures and still sounded great but the Sanders was clearly better defined and detailed and I can only come to the conclusion that keeping all secondary reflections at bay any way you get it just gets you closer to accuracy.
That “triangle” of setup or headphone experience just sounds more detailed. Some like the Bose type effect I guess
Been going to CES for over 30 years and have heard just about everything and start to narrow down what impresses you and the reasons why and it’s all good fun and you make your choices.
(BTW , I have yet to apply your sauce to the Quad panels but getting there)
Regards
David
Attempting to make a big sheet do the same thing and get the same sound field
Illusion is not what I personally hear.
At our Newport hiend show a few years ago ML had their new Neolith playing and right down the hall the Sanders system and it was interesting to go back and forth listening.
The Ml sounded big and vague in outline textures and still sounded great but the Sanders was clearly better defined and detailed and I can only come to the conclusion that keeping all secondary reflections at bay any way you get it just gets you closer to accuracy.
That “triangle” of setup or headphone experience just sounds more detailed. Some like the Bose type effect I guess
Been going to CES for over 30 years and have heard just about everything and start to narrow down what impresses you and the reasons why and it’s all good fun and you make your choices.
(BTW , I have yet to apply your sauce to the Quad panels but getting there)
Regards
David
Jazzman or anyone... can you say why" SPECTRA" ... is not a better way to setup the Diy segmented panel.....using just (two res)..one front an one on the back for just half of the panel?...
Would this give us a "real curved" output to any segmented panels?
I have had the Audiostat segmented panels...Setup like jazzmans panels
an the Spectra type...that give a more coherent sound ...to my ear .
Thanks.... for any an all input on getting better sound from ESLs
THEORY & PRACTICE OF SPECTRA OPERATION
Spectra 11-1100 represents a very high state of perfection
of electrostatic hybrid loudspeakers.
Theintedacetechniques allowing the Symmetric Pair
Electrically Curved Transducer (SPECTRA) mode
of operation were only perfected in late 1986.
At that time Acoustat made a quantum leap in the
productof STEP-UP EFFICIENCY and BANDWIDTH
which allowed for fhe first time a combination of
traditional Acoustat high pedormance with the halfcentury
old ideal of VARIABLE GEOMETRY operation.
This breakthrough solved the long standing impasse
which had not allowed high SPL electrostats to have
optimum behavior at all audio frequencies.
You will find Spectra has wide, pleasant dispersion at
all frequencies, razor-sharp high-frequency time
alignment, and clean low bass.
Spectra achieves this by effectively changing size
and shape at different frequencies. Spectra is about
3" wide at highest frequencies and about 9" wide at
middle frequencies. The lowest frequencies aresmmothy transitioned to an 8" dynamic woofer.
Spectra's excellent midrange results from the magic
of wave physics which causes the 9" of array operatiinal
at these frequencies to act dispersively for
waves leaving both the front and the rear of the array.
Spectra is devoid of the midrange beaming common
in planar speakers using large flat midrange radiators.
Last edited:
Sorry Tyu but I’m not familiar with the SPECTRA configuration. Even if I were, it would probably take someone with better math skills than me to analyze it. I do recall, however, reading a post on this forum (I forget whether it was Golfnut or Bolserst) which indicated that, for symmetrically segmented panels, more/narrower segments gives smoother/wider dispersion -- but there is no advantage in making the segments narrower than the shortest treble wavelengths (about 12mm).
Just speculating here-- but I gleaned from your description of the SPECTRA that it may consist of just one 3” HF section and one or two 9” LF sections; which would equate to just two electrical segments of differing widths. If so, the SPECTA would have a different and perhaps narrower dispersion pattern compared to my 15-section / 8-segment panels.
Many (including me) would say that a narrow dispersion panel actually gives better imaging, coherency and slam when listening from the exact focal sweet spot—because there are fewer secondary reflections bounced from room surfaces. Of course; this very characteristic makes an extreme directivity panel (like the Sanders model 10 for example) great for solo listening at the focus but not so good for entertaining guests.
It could be that the SPECTRA panel has enough dispersion to alleviate the ‘head-in-a-vice’ effect, yet still has relatively narrower dispersion than other segmented panels, and that sounds better to you (it might sound better to me too). I would love to hear a SPECTRA speaker.
You may compare a curved speaker with a flat one, but what makes it sound different? I mean you are comparing two esls with probably entirely different components like transformers, different construction materials, crossover components, mylar thickness etc. etc. In the past I build unsegmented esls; quite beamy as a consequence. They also had more powerfull high mids and trebles so yes I could here 'more detail' compared to the unsegmented version.
You could compare 2 esls which only differ in having their panels bend or not but as the bending changes the frequency response it still is tricky.
You could compare 2 esls which only differ in having their panels bend or not but as the bending changes the frequency response it still is tricky.
I find the SPECTRA type setup to fall under the ...Less is more-better....panels sound.
The idea of... one wire flat panels... with one cute in the middle of the wire... to give two electrical segments on each side ...that's is 4ea inputs wires....on one 9"wide panel..this give a mylar output area of 6" per side.... like the Acoustat SPECTRA panel...we take.. two input, one front, one same side on back... feed from the same one step up trans secondary... thought two res....the other two wire segments inputs... feed the full output...this puts the full output on all 6" this is a the trick...
So what do the res do...setup up a time delay...that the ear must hear as a 6"curved panel...from ea side front an back..... this gives the best flat panel eaze of drive...an to my ear the most natural tone...
But in the end....ML curved panels to my ear...with what ever trans..crossover...amp...room....wins ...the less is more ...argument...
yes the curved panel... it can be called low-mid-high fi....becase of it weard output from the back side being...I don't know but being concave..an not convex...that would give the full Circle like the spectra....
But Man...don't we know.... all these ESLs are a lot of Fun... if you are in this hobby ....an live for great sound every day like I do..
thanks for all your input
The idea of... one wire flat panels... with one cute in the middle of the wire... to give two electrical segments on each side ...that's is 4ea inputs wires....on one 9"wide panel..this give a mylar output area of 6" per side.... like the Acoustat SPECTRA panel...we take.. two input, one front, one same side on back... feed from the same one step up trans secondary... thought two res....the other two wire segments inputs... feed the full output...this puts the full output on all 6" this is a the trick...
So what do the res do...setup up a time delay...that the ear must hear as a 6"curved panel...from ea side front an back..... this gives the best flat panel eaze of drive...an to my ear the most natural tone...
But in the end....ML curved panels to my ear...with what ever trans..crossover...amp...room....wins ...the less is more ...argument...
yes the curved panel... it can be called low-mid-high fi....becase of it weard output from the back side being...I don't know but being concave..an not convex...that would give the full Circle like the spectra....
But Man...don't we know.... all these ESLs are a lot of Fun... if you are in this hobby ....an live for great sound every day like I do..
thanks for all your input
Right now I'm debating whether or not to go with a transmission line design. My initial concept was to have a minimal design like a ML or Acoustat. With these subwoofers:Dayton Audio SD270A-88 10" in each speaker, I'm unsure of how much an improvement a transmission line design would be since their low end already goes down to 26Hz. Attached is my current thinking. (36"x12" panel and 14"x14" MDF box for the subwoofer at the bottom.)
Attachments
The 26Hz you quoted for the woofer is its free air resonance (FS). Its F3 (3db down point) in a box would be significantly higher. A TL would provide somewhat lower extension vs a typical sealed box. But in my estimation, the advantage of a TL is not so much its low end extension but its lower box resonance and reduced compressive effects on cone motion.
Whenever I build a speaker box I like to hum into the woofer opening and listen to the sound rebounding back to my ears. When the pitch of my humming matches the box resonance, the abrupt increase in loudness is unmistakable (rattles my skull, in fact). Using this method I can detect the resonance easily in a sealed or ported box but when I hummed into my TL I nothing back (I’m talking grave yard dead).
Modeling a transmission line leads to insanity but if you're up for the math you might try Martin J. King's TL calculator. King's math was over my head so I just used the generic guidelines in Sanders' ESL Cookbook; which are basically as follows:
- Line length should be about 1/4 wavelength of the lowest desired frequency
- For a folded line the wall behind the woofer should be curved such that standing waves cannot rebound to the woofer cone.
- Line section area at the woofer should be at least 125% of the woofer's piston area (s/d) and should taper to 100% of s/d at the terminus.
- Stuffing density should be about 0.5 lbs/ft3.
- In order to best blend with the ESL's near-zero Q, choose a woofer with low Q, low inductance and a strong motor.
I have a TL design on my website that may interest you and I have a CAD Drawing:
https://4.bp.blogspot.com/-hX8QHoaX...PoC0O4Jy3Pkl6yQCLcBGAs/s640/Eros+Clone+7.jpeg
Whenever I build a speaker box I like to hum into the woofer opening and listen to the sound rebounding back to my ears. When the pitch of my humming matches the box resonance, the abrupt increase in loudness is unmistakable (rattles my skull, in fact). Using this method I can detect the resonance easily in a sealed or ported box but when I hummed into my TL I nothing back (I’m talking grave yard dead).
Modeling a transmission line leads to insanity but if you're up for the math you might try Martin J. King's TL calculator. King's math was over my head so I just used the generic guidelines in Sanders' ESL Cookbook; which are basically as follows:
- Line length should be about 1/4 wavelength of the lowest desired frequency
- For a folded line the wall behind the woofer should be curved such that standing waves cannot rebound to the woofer cone.
- Line section area at the woofer should be at least 125% of the woofer's piston area (s/d) and should taper to 100% of s/d at the terminus.
- Stuffing density should be about 0.5 lbs/ft3.
- In order to best blend with the ESL's near-zero Q, choose a woofer with low Q, low inductance and a strong motor.
I have a TL design on my website that may interest you and I have a CAD Drawing:
https://4.bp.blogspot.com/-hX8QHoaX...PoC0O4Jy3Pkl6yQCLcBGAs/s640/Eros+Clone+7.jpeg
Here's the latest thinking on my speaker:
12" x 41.5" insulated wire panel
3/4" MDF Subwoofer enclosure, painted gloss white
10" subwoofer, quoted F3 of 43.77Hz in sealed enclosure
1 1/2" wood frame to hold the panel, stained a dark black
1 1/2" x 1/2" MDF frames wrapped in black grill cloth (Velcro'd to front and back of stained wood frame)
MiniDSP for crossover and EQing
The wood frame will have 4 gussets that allow it to be bolted onto the front of the MDF box.
My questions regard the placement of the subwoofer.
Should it be angled slightly upward so that the sound is more likely to reach the listener rather than the floor?
Is it placed too far back? It's about 2" behind the panel, meaning that the sound could be out of phase. I'm planning on using a MiniDSP so that I have a 48dB curve and there should be little overlap in frequencies with the panel and sub, but it's still a slight concern of mine.
Sorry if the diagram is confusing, it was the best I could do in Illustrator.
Thanks,
-Ben
12" x 41.5" insulated wire panel
3/4" MDF Subwoofer enclosure, painted gloss white
10" subwoofer, quoted F3 of 43.77Hz in sealed enclosure
1 1/2" wood frame to hold the panel, stained a dark black
1 1/2" x 1/2" MDF frames wrapped in black grill cloth (Velcro'd to front and back of stained wood frame)
MiniDSP for crossover and EQing
The wood frame will have 4 gussets that allow it to be bolted onto the front of the MDF box.
My questions regard the placement of the subwoofer.
Should it be angled slightly upward so that the sound is more likely to reach the listener rather than the floor?
Is it placed too far back? It's about 2" behind the panel, meaning that the sound could be out of phase. I'm planning on using a MiniDSP so that I have a 48dB curve and there should be little overlap in frequencies with the panel and sub, but it's still a slight concern of mine.
Sorry if the diagram is confusing, it was the best I could do in Illustrator.
Thanks,
-Ben
Attachments
Your drawing is very similar to this diy ESL. This speaker and most Martin Logan hybrids have the woofer at floor level and not angled upward -- and it apparently works so I wouldn't worry about it.
I did angle the woofer 6 degrees upward in my beam splitter speaker but I don't know that it made much difference beyond complicating the build.
To the extent that the woofer is out of plane with the panel, the phasing will not exactly match and any audible mismatch would be most noticeable in the crossover band. A crossover frequency of let's say 250Hz equates to a wavelength of about 4.4 feet versus a couple of inches out of plane mismatch may not be audibly objectionable. If MiniDSP has an add on module for time delay you could time align the woofer and panel. I use a DCX2496 and it time aligns the drivers based on mic measurements.
I did angle the woofer 6 degrees upward in my beam splitter speaker but I don't know that it made much difference beyond complicating the build.
To the extent that the woofer is out of plane with the panel, the phasing will not exactly match and any audible mismatch would be most noticeable in the crossover band. A crossover frequency of let's say 250Hz equates to a wavelength of about 4.4 feet versus a couple of inches out of plane mismatch may not be audibly objectionable. If MiniDSP has an add on module for time delay you could time align the woofer and panel. I use a DCX2496 and it time aligns the drivers based on mic measurements.
I posted a polar map comparison between flat-unsegmented, curved, and flat-segmented for the same width panel.
Perhaps this is what you were thinking of?
First time ESL builder #137
All of the Stereophile reviews of curved panels show the hot-spot slightly off-axis in the top octave and then the reduced dispersion in the midrange. Basically once the wavelength of sound gets on the same order of magnitude as the dimension of the arc the sound waves stop being directed by the curved geometry.
Your description of “enough dispersion to alleviate head-in-a-vice” is in agreement with my listening experience.
Preference on the amount of dispersion may be dictated by the reflectiveness of your room.
The Spectra white paper and figure were posted here:
Acoustat Answer Man is here #1185, 1186
Acoustat Answer Man is here #1188
The Spectra 11 configuration is simply two 3.375” wide segments positioned side by side.
From the white paper: “…Sector locations in Spectra also lead to upper midrange being dispersed into the listening area and steered away from the room's side walls….”
Modeling of the Spectra 11(upper-left plot of attached pic) certainly bears that out. Note that the off-axis lobes in the polar plots are alternately in out/in-of-phase with the on-axis sound. My preference is for symmetric segmentation using more/smaller segments such that there isn’t any out-of-phase lobing in the midrange.(lower-right plot) It just sounds more natural to me, and the larger listening area receiving spectrally balanced sound from both speakers is a nice bonus to share with friends and family.
Those curious about pursuing asymmetric segmentation might be interested in the polar modeling data posted here: Another segmented ESL #47
Attachments
On-axis with response equalized flat, the impulse response of segmented and unsegments panels are very similar and quite compact.
Off-axis the segmented panels continue to have quite compact impulse responses, while the impulse response of the un-segmented panels stretches out in time more and more as you move further off-axis.
Even extremely thin (3µm) dust covers will put ripples in the response if they are spaced away from the diaphragm. You can see the ripples in the Quad-ESL63 response, and I also posted some measurements from experimenting with dust covers on my own ESL. Like you said, probably not enough to worry about but there is an influence. The trick is to keep the spacing as small as possible so that the ripples are all at or above the top octave.
All Acoustat panels can give #13
All Acoustat panels can give #16
The one thing that I can't really decide on fully is the subwoofer enclosure. I don't personally know how the F3 of 43Hz would sound with my music. It's hard to translate these numbers into audio. Is there some way I could experience with some software what this would actually sound like? This uncertainty leads me to want to have a lower extension.
Given that Roger Sanders has said ported (or sealed) enclosures aren't the way to go, the only other option is a transmission line design. But 1/4 of the wavelength of 26Hz is nearly 3.3 meters. Is there another type of TL that can lead me to a somewhat reasonable size? I'm just getting into researching them so I'm fairly clueless.
I'm hoping that theres some kind of compromise between a smaller enclosure and one that's 100 inches tall that can be made. I would really like to keep the subwoofer drivers one per panel and have them each be able to be very capable in terms of range and not have an additional sub.
Am I dreaming? Is this just unrealistic?
Thanks,
-Ben
Given that Roger Sanders has said ported (or sealed) enclosures aren't the way to go, the only other option is a transmission line design. But 1/4 of the wavelength of 26Hz is nearly 3.3 meters. Is there another type of TL that can lead me to a somewhat reasonable size? I'm just getting into researching them so I'm fairly clueless.
I'm hoping that theres some kind of compromise between a smaller enclosure and one that's 100 inches tall that can be made. I would really like to keep the subwoofer drivers one per panel and have them each be able to be very capable in terms of range and not have an additional sub.
Am I dreaming? Is this just unrealistic?
Thanks,
-Ben
Assuming 43Hz F3 is correct; that's still rather respectable as speakers go.
As for Roger Sanders' TL; I haven't put a tape measure on it but I have studied its proportions in comparison to the published width/height/depth and white paper sketch and I'm rather certain that its line is between 7 and 8 feet in length (his own literature says it's 8 ft but I estimate a bit less).
I believe the published F3 for the Sanders' model 10D is 22Hz and, if so, he's getting that from a line significantly shorter than 1/4 of a 22Hz wavelength. I can't be certain about the woofer he' using but his literature references a Danish source with aluminum cone, which I think must be a SEAS 10 with FS around 25Hz.
Hopefully someone more knowledgeable with TL's can jump in and shed more light on actual response in relation to line length. I know that stuffing the line slows the speed of sound through it, effectively making the line longer than it actually is, and that may be a means of increasing the low end extension-- just speculating.
Last edited:
In the interest of amplifiers, how do I find a suitable one for these stats? I had no idea about the difficult load that electrostatic speakers presented in my previous projects and I guess i just got lucky with my setup. My last amp was this Sony Reciever. I would probably use this again except for that in stereo it automatically applies a high pass filter (and I'm not exactly sure how/where it applies this). I know to look for a low impedance capability and a high output voltage, but the voltage isn't usually listed. This Onkyo M-282 seems like a reasonable choice, but altogether I'm not exactly sure. I'm also interested in those Class-D amps from PE, but I don't have experience with those and there seems to be limited info on whether they're truly Hi-Fi.
If someone could offer advice on this dilemma it would be greatly appreciated, as I'm starting to lose sleep over making the wrong choice!
Thank you in advance,
Ben
If someone could offer advice on this dilemma it would be greatly appreciated, as I'm starting to lose sleep over making the wrong choice!
Thank you in advance,
Ben
TL sub
This is my current TL sub pair. Still raw ply though and used with Kingsound Kings electrostats crossed over at 50Hz with a minidsp 4x10.
HiVi D8.8 in a TL
The were originally designed for a ER Audio ESL III and were to be placed horizontally slotted in under the panels. I had them crossed over at 200 to 300 Hz. Now they are placed behind the Kingsounds with about 0.8ms delay on the panels I think.
I have 2 configurations on the DSP for with and without the subs. I use a bit of boost under 200hz for the non sub configuration which is good for low level listening (wife and family around) for watching movies or background music.
I got the HiVi D8.8 drivers pretty cheap many years ago now as they were sent to the local importer by mistake. As mentioned in the thread I tried them in sealed and ported boxes first.
Regards,
kffern
This is my current TL sub pair. Still raw ply though and used with Kingsound Kings electrostats crossed over at 50Hz with a minidsp 4x10.
HiVi D8.8 in a TL
The were originally designed for a ER Audio ESL III and were to be placed horizontally slotted in under the panels. I had them crossed over at 200 to 300 Hz. Now they are placed behind the Kingsounds with about 0.8ms delay on the panels I think.
I have 2 configurations on the DSP for with and without the subs. I use a bit of boost under 200hz for the non sub configuration which is good for low level listening (wife and family around) for watching movies or background music.
I got the HiVi D8.8 drivers pretty cheap many years ago now as they were sent to the local importer by mistake. As mentioned in the thread I tried them in sealed and ported boxes first.
Regards,
kffern
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.