Perhaps this is the post you were thinking of?
http://www.diyaudio.com/forums/plan...con-dots-resonance-control-3.html#post1958582
From your pics, I think your thread count estimate is about right ~ 100/inch.
Unfortunately, the thread size looks pretty small resulting in the pore size being relatively large so it will likely not provide a whole lot of damping.
To a first approximation, acoustic resistance for thin monofilament mesh is inversely proportional to the square of the pore size. Mesh thickness and % open area are important terms as well, but the pore size is the dominant term for the range of resistance we are interested in. The 110 and 160 count mesh I measured in the post linked above had acoustic resistance of about 8 and 18 Rayles respectively. I would estimate your cloth would be less than 5 Rayles. For good LF damping, you need something more in the 30 – 40 range. If you use mesh on front and rear stators like I do, the resistive damping is additive. It also provides a decent amount of protection from dust and cat hair getting into the air gap.
As SyBorg mentioned, you can get mesh from art supply stores. For example:
Silk Screen Printing Mesh | Silk Screening Supplies
Monofilament Polyester Screen Fabric - BLICK art materials
There are plenty of other sources, including ebay I’m sure.
Mesh made for silk screening with 160 – 200 threads per inch would be a good starting point.
Attachments
There was an Excel spreadsheet posted over in the ESL simulator thread that plots the falling LF response behavior golfnot(and Bazukaz) mentioned when you try to build a full range segmented ESL that is not floor-to-ceiling in height. You can also specify the diaphragm resonance frequency and Q to play with the tuning options for counteracting the falling LF response.
http://www.diyaudio.com/forums/plan...r-esl-simulator-esl_seg_ui-2.html#post2913884
cutting rods
Simple method for cutting large amount of rods.
I used 4 hose clamps to bundle the 10 lbs of rods about 1 1/4" in dia and placed them in my cut off saw.
Now I have 32" long precision cut lengths.
Now
have to decide on 32 or 64". Gonna try rod end fusion
using my big 400 amp welder. Gettin the current set will be the big trick.
Simple method for cutting large amount of rods.
I used 4 hose clamps to bundle the 10 lbs of rods about 1 1/4" in dia and placed them in my cut off saw.
Now I have 32" long precision cut lengths.
Nowhave to decide on 32 or 64". Gonna try rod end fusion
using my big 400 amp welder. Gettin the current set will be the big trick.
A question that both Charlie and I will face shortly regards the division of the diaphragm. The rule of thumb I heard was to have no more than 1:100 d/s to open diaphragm (and preferably less). In my case, I will have a d/s of 0.125 (1/8") so at 1:80 that means no more than 10" of open diaphragm.
The stator segments are running lengthwise in a 23"x47" panel, so I would need to have at least 5 diaphragm segments (maybe more). The question is: Should they all be the same spacing, or should they be different spacings. I have heard arguments on both sides.
Varying spacings will result in different diaphragm resonances for each section, but, the Q of the panel resonance as a whole will be less. Identical spacing will result in a single resonant frequency of higher Q, but perhaps a single frequency can be more easily damped (silk screen cloth), giving better overall LF performance.
I seem to recall some discussion about this before but could not find the thread. I think it was Bolserst who was surprised that the varying resonances actually sounded and measured better when engineering sense should have pointed otherwise.
So, is there a consensus? Identical divisions or varying ones?
The stator segments are running lengthwise in a 23"x47" panel, so I would need to have at least 5 diaphragm segments (maybe more). The question is: Should they all be the same spacing, or should they be different spacings. I have heard arguments on both sides.
Varying spacings will result in different diaphragm resonances for each section, but, the Q of the panel resonance as a whole will be less. Identical spacing will result in a single resonant frequency of higher Q, but perhaps a single frequency can be more easily damped (silk screen cloth), giving better overall LF performance.
I seem to recall some discussion about this before but could not find the thread. I think it was Bolserst who was surprised that the varying resonances actually sounded and measured better when engineering sense should have pointed otherwise.
So, is there a consensus? Identical divisions or varying ones?
No, the rods (and electrical segments) will go lengthwise, the 47" dimension - the diaphragm will be sectioned width-wise, the 23" dimension.
I'm taking photos and describing the progress as I go along on this build at:
Full Range ESL Build
in case anyone is interested. So far, I'm working on the stator panels, just thinking ahead to the diaphragm.
I'm taking photos and describing the progress as I go along on this build at:
Full Range ESL Build
in case anyone is interested. So far, I'm working on the stator panels, just thinking ahead to the diaphragm.
Hi Syborg
The <100d/s rules applies only to the smallest of the two dimensions so three ~9" segments is about right. On my ESL for example, the diaphragm is partitioned as 1050mm x 185 mm segments, my diaphragm-stator spacing is about 2.5 mm.
Regarding the deliberate use of different spacing to spread the resonant frequencies - load of bollocks I think, for several reasons - I'll give two.
The resonance comes about because of the air mass that moves with the diaphragm (like shaking a bedsheet) and the spring action of the diaphragm. Think of the air mass as an inductance and the spring of the diaphragm as a capacitance => LC series resonance.
The first reason its bollocks is that the imaginary part of the radiation impedance for a diaphragm, which is responsible for the air mass attached to the diaphragm, varies over the over the whole ESL panel - the air mass is greater near the centre of the panel.
Secondly, spring systems that are coupled together don't behave independently of each other - for the ESL, the air mass is shared by all segments. This means its incorrect to think of each segment as having its own resonant frequency. Instead if you have a 3-segment ESL, the ESL panel will have three resonant frequencies - called normal modes or eigenmodes. You can see this effect if tap on the panel frame and record the sound and display it on a PC - you can see the different notes beating. On a single segment panel you will see a single pure note.
For example, with one of the resonant modes, all three segments of the diaphragm will move in the same direction. In one of the other modes, the centre segment will move in the opposite direction to the two outer segments - and this mode will have a different frequency. In the third mode the two outer segments will move in opposite directions and the inner segment will note move. The main point is, the three segments are coupled and the nature of the resonances is very complicated physics.
There is a nice video at https://www.youtube.com/watch?v=zlzns5PjmJ4 showing a coupled spring system, first with two masses, and then with three masses (after 2:25). These will give much the same behaviour as two and three segment ESLs.
best wishes
Rod
The <100d/s rules applies only to the smallest of the two dimensions so three ~9" segments is about right. On my ESL for example, the diaphragm is partitioned as 1050mm x 185 mm segments, my diaphragm-stator spacing is about 2.5 mm.
Regarding the deliberate use of different spacing to spread the resonant frequencies - load of bollocks I think, for several reasons - I'll give two.
The resonance comes about because of the air mass that moves with the diaphragm (like shaking a bedsheet) and the spring action of the diaphragm. Think of the air mass as an inductance and the spring of the diaphragm as a capacitance => LC series resonance.
The first reason its bollocks is that the imaginary part of the radiation impedance for a diaphragm, which is responsible for the air mass attached to the diaphragm, varies over the over the whole ESL panel - the air mass is greater near the centre of the panel.
Secondly, spring systems that are coupled together don't behave independently of each other - for the ESL, the air mass is shared by all segments. This means its incorrect to think of each segment as having its own resonant frequency. Instead if you have a 3-segment ESL, the ESL panel will have three resonant frequencies - called normal modes or eigenmodes. You can see this effect if tap on the panel frame and record the sound and display it on a PC - you can see the different notes beating. On a single segment panel you will see a single pure note.
For example, with one of the resonant modes, all three segments of the diaphragm will move in the same direction. In one of the other modes, the centre segment will move in the opposite direction to the two outer segments - and this mode will have a different frequency. In the third mode the two outer segments will move in opposite directions and the inner segment will note move. The main point is, the three segments are coupled and the nature of the resonances is very complicated physics.
There is a nice video at https://www.youtube.com/watch?v=zlzns5PjmJ4 showing a coupled spring system, first with two masses, and then with three masses (after 2:25). These will give much the same behaviour as two and three segment ESLs.
best wishes
Rod
Hi,
IMO good points there. For me it was always interesting to find a way to possibly measure/visualize the behaviour of an ESL membrane in operation. Any one tried to use a strobe light to do this ..?
Another applet to play with, showing some possible ways a membrane could behave :
Rectangular Membrane Applet
A link to a PDF with some math and visualization of both strings, membranes and other stuff :
http://www.thphys.may.ie/Notes/MP205/MP205_Lecture_13-14.pdf
Some questions :
a) Isn't the air load highest at the bottom of a floor-standing ESL where it's an imaginary center of the line source?
b) Was your 3-segment panel sectioned vertically or horizontally?
c) What was the correlation in measured output level between 1,1 and 1,3 modes in your segmented panel ?
Regards,
Lukas.
I think this is probably the post you were thinking of:
http://www.diyaudio.com/forums/plan...nge-electrostatic-question-2.html#post3968866
For a given panel area and D/S, the highest bass SPL output will be had if you use identical divisions so the whole panel area is producing useful output down the bandwidth cutoff. That being said, for good bass quality you will need to damp the resonance appropriately with mesh or thin felt.
In theory I am in agreement with golfnut’s post. However, measurements of the SoundLab ESL indicate that there is more going on here than one might first think. I hope to revisit this sometime in the future to understand why it behaves the way it does, counter to intuition.
You can clearly see from the SL data in the post linked to above that the sections with the lowest resonance are the largest once at the bottom of the panel, not the ones in the middle.(or perhaps the floor acts like a mirror and these are at the middle of the real/virtual panel?) In any case you can see a spread of resonances as you move up the panel, all of which appear to be lower in Q than is typical for an ESL without any acoustic/resistive damping.
Just had a thought…the mass loading distribution you describe is only true for a uniformly moving piston/diaphragm. Since you have different sections of the diaphragm with different sizes, perhaps their differing motions upsets this general rule of acoustics. It seems kind of like a chicken-and-egg thing though…radiation impedance local to one section is caused by motion of the all the diaphragm sections, but each diaphragm section resonance/motion is altered by its local radiation impedance. Makes my head hurt...
Last edited:
Hi Lukas
First - answers to your questions...
1. Yes - subject to caveats below
2. Yes - actually the ESL is about 2.3 m high in total, built with two separate ~1150 x 380 mm panels split vertically down the middle - so strictly there are four sections. Knowing what I know now I would make them slightly wider and shorter (curtain rail height to improve WAF
), with each of the two panels split 3 ways. vertically3. Not measured.
Actually the situation is very much more complicated than suggest so far. Firstly, as discussed in the other thread, each diaphragm section has multiple resonances anyway - similar to rectangular plate with the edges tied down.
Secondly, each of the various eigen modes has its own radiation impedance which may be quite different from the others. This makes the whole problem hugely messy (sorry for making your headache worse bolserst!) If you look at the video I linked in the previous post you will see that with the three-mass system on the air track, the mode with the lowest frequency is the one where all the masses move synchronously. I'm not sure this is the case with ESLs. I really have no idea which might be at the lowest frequency.
However, independent of all these complications, almost certainly there will be one dominant resonance. I'm guessing that for all of the different modes and radiation impedances, the radiation resistance falls very quickly at low frequencies (e.g. certainly does for a line source - falls as f^3, for a square or circular source it falls as f^4). The radiation resistances damp the various oscillations, and the one with the lowest largest wL/R will dominate. If you damp that one to useful levels (Q up to 2 maybe), then all the others will be damped more and cause no problems. This is why the high-order resonances in the diaphragms are not really relevant.
A further complication, some of the modes will directly excited by the ESL drive, others wont. There is an interesting experiment, put a pair of ESLs side by side. Swap the leads to one so the two are out of phase. Now SLOWLY (!!! this mode has very low resistance and diaphragm displacement is huge) apply signal. If you walk close to them you will hear a very distinctive bass boom that falls off rapidly with distance - this is a mode that would not normally be excited in my ESL (with the split down the middle), but must be there.
best wishes
Rod
Hi Al
13 segments of 5 rods each looks pretty good to me.
0.060" rods spaced at 50% would give segment width of 0.6" (15 mm) . The rods will be a bit thicker once you include the dielectric coating and pull the open fraction of the area down to 40%, which is about right.
According to Bolserst's rule of thumb for good polar response (horizontal dispersion):
lowest frequency=100 kHz/4N^2 (=149Hz),
so it should comfortably perform down to 200 Hz
Just need the panel (or segment) capacitance now to determine the correct intersegment resistors.
Regards
Rod
13 segments of 5 rods each looks pretty good to me.
0.060" rods spaced at 50% would give segment width of 0.6" (15 mm) . The rods will be a bit thicker once you include the dielectric coating and pull the open fraction of the area down to 40%, which is about right.
According to Bolserst's rule of thumb for good polar response (horizontal dispersion):
lowest frequency=100 kHz/4N^2 (=149Hz),
so it should comfortably perform down to 200 Hz
Just need the panel (or segment) capacitance now to determine the correct intersegment resistors.
Regards
Rod
Just to clarify, the “N” in the rule of thumb equation is number of electrical segments(not physical segments) in a Symmetric ESL using Configuration 1. If you have 13 physical segments of 5 rods in a symmetric ESL, this would be a Configuration 2 symmetric ESL with 7 electrical segments. (see Attachment #1). Recall from post #39, that using Configuration 2 provides nearly the same dispersion as a Configuration 1 ESL with twice as many electrical segments. So, I believe your conclusion is still correct that 13 segments of 5 rods is about as good as you can get with fL = 200Hz.
Attachment #2 attempts to clarify the width and number of segments(physical and electrical) for the two symmetric configurations from the Excel spreadsheet calculator posted here: http://www.diyaudio.com/forums/planars-exotics/48120-experiences-esl-directivity-9.html#post2218526
The width and segment numbers in Attachment #2 corresponds to the "W(sec):" and "N:" row entries in the spreadsheet's design table.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.