Good evening all,
Please forgive my ignorance, but having read through this very interesting subsection of DIYAudio, I am left with an interest in ESL panels and many questions! If I might ask a few questions to those of you in the know, with regards to these large DIY ESL panels:
1. How does the distortion performance compare to a more conventional system?
2. What is a typical frequency response of such a system?
3. What sensitivity levels are typically achieved and is maximum SPL output limited by mechanical or electrical limitations?
Thank you for your patience!
Jim
Please forgive my ignorance, but having read through this very interesting subsection of DIYAudio, I am left with an interest in ESL panels and many questions! If I might ask a few questions to those of you in the know, with regards to these large DIY ESL panels:
1. How does the distortion performance compare to a more conventional system?
2. What is a typical frequency response of such a system?
3. What sensitivity levels are typically achieved and is maximum SPL output limited by mechanical or electrical limitations?
Thank you for your patience!
Jim
Hi,
You can find somegeneral answers under http://www.diyaudio.com/forums/showthread.php?t=104829 and specially in http://www.diyaudio.com/forums/showthread.php?threadid=64620.
1) The level of distortion depends on several factors.
- drive system: constant voltage ESL-->non linear drive--> raised distortion level| constant charge ESL-->linear drive-->low distortion level.
- the bandwidth: As long as the membrane doesn´t need to move, i.e the strokes are kept small, distortion values can be extremely small. Below 200Hz distortion rises quickly. So its best to restrict the frequency range to >200Hz and leave the bass to dynamic drivers which are superior in their range.
- transformer: the distortions generated by the audio-transformer can be of orders higher than the distortions from the panel itself. The lower the frequency is chosen the higher the distortions are (increasing flux values within the core). Different core types and the execution of the windings define the quality of this part. A badly chosen tranny or badly executed panel show distortion values similar or higher to those typically found with dynamic drivers. Well executed trannies and panels show distortion values much lower than dynamic drivers and often even lower than the driving electronics. At time ESLs are probabely the lowest distortion drivers available (within their territory of frequency range...there´s no real good Bass-ESL!)
2) ESLs are wide bandwidth drivers. There is nearly no other system on the market that allows for such good behaviour over such a wide frequency range. But You almost always need some equalization measurements. Following the above link (multidocs 1 and 2) You see that the ESL exhibits a strong basic resonance, which has to be dealt with electrically or mechanically. When installed as an open ´boxless´ system there will be a suckout above the Fs up to several hundreds of Hz (incresing with measuring distance). There´s no suckout when installed into a box, but this kind of installation features other even more serious problems. In a range from 1kHz to 4kHz most panels show a slightly pronounced frequency response. The highs will depend on membrane weight (thickness), openness of the stator sheets and the audio tranny and its associated circuitry. The panels frequency response is normally eqalized to a steady linear falling response (with rising frequency) when measured under nearfield conditions (see multidocs 3 and 4).
3) ESLs achive less power per membrane area as dynamic drivers. To get comparable levels of SPL they need more membrane area (since the electrical field drive only allows for small strokes You simply can´t design larger stroke capability into the panel as You can do with dynamic drivers). Luckily its easy to construct large membrane areas into an ESL without them looking big and ugly. When designed properly ESLs can execute comparably levels of SPL as dynamic drivers and may even come close to horn-loaded drivers. Especially tall and thin panels that exhibit a cylindrical wave pattern can be very loud over long distances. The SPL-distribution over distance is more even over distance than with a global or lobed distribution character.
Depending on the design the panel limits either mechanically by the membrane touching the stators or electrically by a flashover. Anyway these limits are no soft limits. SPL increases very linear up to this maximum level, whereas a dynamic driver shows thermal compression, BL-decrease, progressive stiffening of the supension et al effects. Sonically this gives the impression of ease and limitless reserves quite similar to a big horn speaker. On the other hand resolution of lowest-level details is far better than with any dynamic driver. So playing ESLs at very low levels is still very listenable while dynamic drivers suck (side note: When I coat the membranes of my panels the liquid develops small pinhead-sized bubbles. You can actually hear singles of those tiny bubbles pop up when the liquid dries out! It´s because the membrane reacts to microscopic small forces. Even a leightweighted dome tweeter wouldn´t hardly notice anything under these conditions, let alone a midrange- or bass-driver). So while the maximum SPL can be (rem: can!..because often this is not the case, especially not so with fullrange ESLs) comparable to dynamic drivers, You ´gain´ dynamic headroom at the low level SPL end.
jauu
Calvin
You can find somegeneral answers under http://www.diyaudio.com/forums/showthread.php?t=104829 and specially in http://www.diyaudio.com/forums/showthread.php?threadid=64620.
1) The level of distortion depends on several factors.
- drive system: constant voltage ESL-->non linear drive--> raised distortion level| constant charge ESL-->linear drive-->low distortion level.
- the bandwidth: As long as the membrane doesn´t need to move, i.e the strokes are kept small, distortion values can be extremely small. Below 200Hz distortion rises quickly. So its best to restrict the frequency range to >200Hz and leave the bass to dynamic drivers which are superior in their range.
- transformer: the distortions generated by the audio-transformer can be of orders higher than the distortions from the panel itself. The lower the frequency is chosen the higher the distortions are (increasing flux values within the core). Different core types and the execution of the windings define the quality of this part. A badly chosen tranny or badly executed panel show distortion values similar or higher to those typically found with dynamic drivers. Well executed trannies and panels show distortion values much lower than dynamic drivers and often even lower than the driving electronics. At time ESLs are probabely the lowest distortion drivers available (within their territory of frequency range...there´s no real good Bass-ESL!)
2) ESLs are wide bandwidth drivers. There is nearly no other system on the market that allows for such good behaviour over such a wide frequency range. But You almost always need some equalization measurements. Following the above link (multidocs 1 and 2) You see that the ESL exhibits a strong basic resonance, which has to be dealt with electrically or mechanically. When installed as an open ´boxless´ system there will be a suckout above the Fs up to several hundreds of Hz (incresing with measuring distance). There´s no suckout when installed into a box, but this kind of installation features other even more serious problems. In a range from 1kHz to 4kHz most panels show a slightly pronounced frequency response. The highs will depend on membrane weight (thickness), openness of the stator sheets and the audio tranny and its associated circuitry. The panels frequency response is normally eqalized to a steady linear falling response (with rising frequency) when measured under nearfield conditions (see multidocs 3 and 4).
3) ESLs achive less power per membrane area as dynamic drivers. To get comparable levels of SPL they need more membrane area (since the electrical field drive only allows for small strokes You simply can´t design larger stroke capability into the panel as You can do with dynamic drivers). Luckily its easy to construct large membrane areas into an ESL without them looking big and ugly. When designed properly ESLs can execute comparably levels of SPL as dynamic drivers and may even come close to horn-loaded drivers. Especially tall and thin panels that exhibit a cylindrical wave pattern can be very loud over long distances. The SPL-distribution over distance is more even over distance than with a global or lobed distribution character.
Depending on the design the panel limits either mechanically by the membrane touching the stators or electrically by a flashover. Anyway these limits are no soft limits. SPL increases very linear up to this maximum level, whereas a dynamic driver shows thermal compression, BL-decrease, progressive stiffening of the supension et al effects. Sonically this gives the impression of ease and limitless reserves quite similar to a big horn speaker. On the other hand resolution of lowest-level details is far better than with any dynamic driver. So playing ESLs at very low levels is still very listenable while dynamic drivers suck (side note: When I coat the membranes of my panels the liquid develops small pinhead-sized bubbles. You can actually hear singles of those tiny bubbles pop up when the liquid dries out! It´s because the membrane reacts to microscopic small forces. Even a leightweighted dome tweeter wouldn´t hardly notice anything under these conditions, let alone a midrange- or bass-driver). So while the maximum SPL can be (rem: can!..because often this is not the case, especially not so with fullrange ESLs) comparable to dynamic drivers, You ´gain´ dynamic headroom at the low level SPL end.
jauu
Calvin
Last edited:
Hello Calvin,
you answered the questions detailed and seriously.
The term "electrostatic" describes the way an ESL is driven.
Because of the properties of the motor a typical ESL
is a large area flat panel design with releatively small maximum
excursion.
Now my question:
Is in Your opinion such a typical flat panel ESL also a bending wave
transducer ?
The term "bending wave transducer" refers only to the way sound is
radiated from the membrane and does not imply any assumptions on how
the membrane is driven.
Is the whole area of an ESL radiating in phase over the whole audio
spectrum ?
you answered the questions detailed and seriously.
The term "electrostatic" describes the way an ESL is driven.
Because of the properties of the motor a typical ESL
is a large area flat panel design with releatively small maximum
excursion.
Now my question:
Is in Your opinion such a typical flat panel ESL also a bending wave
transducer ?
The term "bending wave transducer" refers only to the way sound is
radiated from the membrane and does not imply any assumptions on how
the membrane is driven.
Is the whole area of an ESL radiating in phase over the whole audio
spectrum ?
Hi,
thanks for the praise ;-)
In short:
A1: no
A2: it depends
As You said, the drive system is not what matters but the way the membrane radiates sound. With a bending wave transducer adjacent parts of the membrane vibrate out of phase (or better with a certain phase shift). The membrane of an ESL on the other hand is driven in phase, though not with same amplitude (which leads to reflections at the circumference and standing wave modes). An exception to this are segmented panels, where each segment vibrates in phase, but with a phase shift to adjacent segments (Quad et al).
jauu
Calvin
thanks for the praise ;-)
In short:
A1: no
A2: it depends
As You said, the drive system is not what matters but the way the membrane radiates sound. With a bending wave transducer adjacent parts of the membrane vibrate out of phase (or better with a certain phase shift). The membrane of an ESL on the other hand is driven in phase, though not with same amplitude (which leads to reflections at the circumference and standing wave modes). An exception to this are segmented panels, where each segment vibrates in phase, but with a phase shift to adjacent segments (Quad et al).
jauu
Calvin
Interesting. Based on simple diy panels and many commercial samples i am quite surprised with this statement. Could you give an example of a design which can really compete with dynamic speakers? Obviously the entire Quad range is severely deficient in this area, at least to my ears.
Hi,
Full range esls like quad usually have some restrictions in dynamic range, though they may play more than loud enough for most of us.
Hybrid esls ( with woofer added) tend to go much louder, partly because of the higher efficiency but also because the film doesn't have to move that much at higher frequencies and won't hit the stators.
Full range esls like quad usually have some restrictions in dynamic range, though they may play more than loud enough for most of us.
Hybrid esls ( with woofer added) tend to go much louder, partly because of the higher efficiency but also because the film doesn't have to move that much at higher frequencies and won't hit the stators.
Hi,
as MJ pointed out, You are restricted to hybrids if You want to reach highest dynamic headroom values.
Basically every well designed hybrid panel could reach earsplitting levels of SPL. MLs do, PASs do my own panels do and there are certainly others that do too.
Equivalent to nature: Size matters! *lol*
Think of a panel of 2.000-3.000cm² area, +-1mm possible excursion and imagine how much more air volume this can move compared to a 6.5" driver with typically around 130cm² and +-3mm excursion, or even a 1"-dome with meager 6cm² and +-0.5mm excursion. Roughly the possible dynamic range is restricted by the membrane area times the maximum excursion (air volume). The pure dynamic driver features app. 1/500 to 1/20 of the ESL´s membrane area and typically higher excursion (<5:1).
Now think about what a horn is and how it works (rem: horn without a pressure chamber). It can be considered as a mechanical transformer that transforms a small membrane area at the horn throat into a larger ´virtual´ membrane area at the horn mouth. This increases the dynamic range. But as with any transformation this is a compromised process (losses, reduced bandwidth etc).
The ESL on the other hand features a real large membrane and as such there´s no reason why it shouldn´t reach high dynamic values. The only reason against is that the electrostatic forces working directly upon the membrane are considerably weaker than the magnetic forces driving a dynamic speaker. They decrease quadratically with increasing stator-stator-distance.
But this disadvantage is easily made up by the larger membrane area and the much lower membrane weight-ratio (0.5-2mg/cm² versus 50-100mg/cm²).
So a well designed panel (with regard to dynamics) is one with small stator-stator-distances. This discludes the use for low frequencies that ask for considerably large strokes --> hence the usage as hybrids above 100Hz.
Segmentation of panels (reducing effective membrane dimensions with increasing frequency) used to control the distribution character reduces the dynamic hedroom as well, not only through the smaller effective membrane area but also due to the needed higher transformation factor of the audio tranny (always use the lowest transformation factor possible). Still though some segmented panels can reach more than sufficiently high levels of SPL.
jauu
Calvin
as MJ pointed out, You are restricted to hybrids if You want to reach highest dynamic headroom values.
Basically every well designed hybrid panel could reach earsplitting levels of SPL. MLs do, PASs do my own panels do and there are certainly others that do too.
Equivalent to nature: Size matters! *lol*
Think of a panel of 2.000-3.000cm² area, +-1mm possible excursion and imagine how much more air volume this can move compared to a 6.5" driver with typically around 130cm² and +-3mm excursion, or even a 1"-dome with meager 6cm² and +-0.5mm excursion. Roughly the possible dynamic range is restricted by the membrane area times the maximum excursion (air volume). The pure dynamic driver features app. 1/500 to 1/20 of the ESL´s membrane area and typically higher excursion (<5:1).
Now think about what a horn is and how it works (rem: horn without a pressure chamber). It can be considered as a mechanical transformer that transforms a small membrane area at the horn throat into a larger ´virtual´ membrane area at the horn mouth. This increases the dynamic range. But as with any transformation this is a compromised process (losses, reduced bandwidth etc).
The ESL on the other hand features a real large membrane and as such there´s no reason why it shouldn´t reach high dynamic values. The only reason against is that the electrostatic forces working directly upon the membrane are considerably weaker than the magnetic forces driving a dynamic speaker. They decrease quadratically with increasing stator-stator-distance.
But this disadvantage is easily made up by the larger membrane area and the much lower membrane weight-ratio (0.5-2mg/cm² versus 50-100mg/cm²).
So a well designed panel (with regard to dynamics) is one with small stator-stator-distances. This discludes the use for low frequencies that ask for considerably large strokes --> hence the usage as hybrids above 100Hz.
Segmentation of panels (reducing effective membrane dimensions with increasing frequency) used to control the distribution character reduces the dynamic hedroom as well, not only through the smaller effective membrane area but also due to the needed higher transformation factor of the audio tranny (always use the lowest transformation factor possible). Still though some segmented panels can reach more than sufficiently high levels of SPL.
jauu
Calvin
Last edited:
Good afternoon Calvin and all,
Thank you for taking the time to post up your thoughts on the matter. Very interesting stuff! I have been toying with the idea of building a set of live-performance oriented speakers for my living room. That is to say, something with a large sound stage, and full range frequency response, and then hopefully reasonable SPL capability for headroom transients and overall level. I don't mind if the FR is the flattest ever as long as it is reasonable. Dipoles, from what I have read of them so far, interest me for such an application, as the the room interactions resulting from dipole radiation may lend the program material a live feel, though this might be a special effect of sorts.
All that to say I have a lot of reading ahead of me! Such a system might be a bit ambitious, but I will certainly give it some thought.
Jim
Thank you for taking the time to post up your thoughts on the matter. Very interesting stuff! I have been toying with the idea of building a set of live-performance oriented speakers for my living room. That is to say, something with a large sound stage, and full range frequency response, and then hopefully reasonable SPL capability for headroom transients and overall level. I don't mind if the FR is the flattest ever as long as it is reasonable. Dipoles, from what I have read of them so far, interest me for such an application, as the the room interactions resulting from dipole radiation may lend the program material a live feel, though this might be a special effect of sorts.
All that to say I have a lot of reading ahead of me! Such a system might be a bit ambitious, but I will certainly give it some thought.
Jim
- Status
- Not open for further replies.