I'm toying with the idea of a hybrid esl.
To keep the D/S spacing down in order to keep efficiency up I guess bass is handled "best" by a conventional woofer.
The question is where to put the x-over frequency?
300Hz would make it easier (150-200Hz would make it really easy) to build a bass modul but I've seen recomendations to keep it between 500-600Hz in order to keep the SPL up?
How would you do it?
To keep the D/S spacing down in order to keep efficiency up I guess bass is handled "best" by a conventional woofer.
The question is where to put the x-over frequency?
300Hz would make it easier (150-200Hz would make it really easy) to build a bass modul but I've seen recomendations to keep it between 500-600Hz in order to keep the SPL up?
How would you do it?
Hi,
I´d say it depends....
.....on panel dimensions in first place. The lower the panel can reach down in frequency the lower You can place the CR-frequency. So it´s a matter of area and stator-stator distances and the mechanical tension of the diaphragm. A larger sized panel will show less drop out through phase cancellation and allows lower CR-freqs. I wouldn´t place the CR lower than the frequency where the dropout reaches -3dB, maximum -6dB, otherwise you need too much equing, hence too much power and You run into excursion problems and decreased dynamic range. If the panel can accept the high power, because it is built very precise and with a tightly tensioned diaphragm You might test the -6dB-point. If it is built just mediocre and with lower tensioned film You might rather take the -3dB-point.
Sonically I like the lower CR-freqs better. With very precise built large panels You may reach down below 200Hz (and some special features which I´m not allowed to disclude here) and still can have an outstanding dynamic range.
- CR-frequency may also be dependant on the choice of bass system. Dipoles are probabely the best partners for ESLs sonically but their upper bandwidth limit stays typically below 400Hz and the useful freq-range rarely exceeds 300Hz.
Closed boxes et al don´t face the upper bandwidth limit problem, but don´t integrate in such a smooth and seemless way with the ESL as dipoles can.
To give You an idea.
My ´small´ system features a panel of 3.000cm² area, accompanied by 6x 6.5" bass-drivers in a dipole bass tower, crossed over at ~250Hz.
The ´big´ system features a panel of 6.000cm² area, accompanied by 8x 6.5" drivers in a dipole bass tower and a 2x15" Subwoofer in CB, crossed over at ~170Hz and 50Hz. Dynamic range reaches ~110dB@4m.
jauu
Calvin
I´d say it depends....
.....on panel dimensions in first place. The lower the panel can reach down in frequency the lower You can place the CR-frequency. So it´s a matter of area and stator-stator distances and the mechanical tension of the diaphragm. A larger sized panel will show less drop out through phase cancellation and allows lower CR-freqs. I wouldn´t place the CR lower than the frequency where the dropout reaches -3dB, maximum -6dB, otherwise you need too much equing, hence too much power and You run into excursion problems and decreased dynamic range. If the panel can accept the high power, because it is built very precise and with a tightly tensioned diaphragm You might test the -6dB-point. If it is built just mediocre and with lower tensioned film You might rather take the -3dB-point.
Sonically I like the lower CR-freqs better. With very precise built large panels You may reach down below 200Hz (and some special features which I´m not allowed to disclude here) and still can have an outstanding dynamic range.
- CR-frequency may also be dependant on the choice of bass system. Dipoles are probabely the best partners for ESLs sonically but their upper bandwidth limit stays typically below 400Hz and the useful freq-range rarely exceeds 300Hz.
Closed boxes et al don´t face the upper bandwidth limit problem, but don´t integrate in such a smooth and seemless way with the ESL as dipoles can.
To give You an idea.
My ´small´ system features a panel of 3.000cm² area, accompanied by 6x 6.5" bass-drivers in a dipole bass tower, crossed over at ~250Hz.
The ´big´ system features a panel of 6.000cm² area, accompanied by 8x 6.5" drivers in a dipole bass tower and a 2x15" Subwoofer in CB, crossed over at ~170Hz and 50Hz. Dynamic range reaches ~110dB@4m.
jauu
Calvin
You give me hope. 
Of course I am comfident in my ability to keep up the building quality.
My initial thoughts were something along the lines of a 8"x48" (2400cm^2) panel , pretty close to your small system but not quite. (Phase cancellation will start early)
I get what you mean when you say your big system...
The larger options doesn't have the same WAF as the small one and I imagine the difficulties increase when you opt for big screens?
Of course I am comfident in my ability to keep up the building quality.
My initial thoughts were something along the lines of a 8"x48" (2400cm^2) panel , pretty close to your small system but not quite. (Phase cancellation will start early)
I get what you mean when you say your big system...
The larger options doesn't have the same WAF as the small one and I imagine the difficulties increase when you opt for big screens?
Hi Markus,
My hybrids have 12”x 48” panels mated to a 10” transmission line bass and they are bi-amp’d using an active crossover with 24db slopes. The panels are sectioned with 3.5” span between the diaphragm supports, 1/16” d/s spacing and the diaphragm is mechanically tensioned to just over 1% elongation. For my setup, a 275hz crossover frequency works very well without excessive EQ'ing.
Good luck with your project!
Charlie
My hybrids have 12”x 48” panels mated to a 10” transmission line bass and they are bi-amp’d using an active crossover with 24db slopes. The panels are sectioned with 3.5” span between the diaphragm supports, 1/16” d/s spacing and the diaphragm is mechanically tensioned to just over 1% elongation. For my setup, a 275hz crossover frequency works very well without excessive EQ'ing.
Good luck with your project!
Charlie
What d/ of sensitivity would one expect from a ESL with a <300Hz crossover frequency?
Lets say 2'x5'? Or 12"x48" like you already have CharlieM.
I will have to defer to others to answer that question, as I don't have the means to actually measure the output of my panels. My impression, however, is that a 4ft2 panel with 1/16 d/s spacing and 1:70 or so step up ratio driven by moderately powered amps will easily play loud enough to suit most people.
I'm bi-amping my hybrids with a pair of 225W/channel Carvers and 1/3 throttle is about all I can stand (they'll blow women's panties right off).
My friend Mavric has similar hybrids bi-amp'd with a pair of 100W/channel Adcoms and they rock n' roll pretty good.
If you look at Sheldon Stokes' ESL 1.0's, his panels are almost certainly less efficient than mine, since he's using twice the d/s spacing (1/8"), yet he powers them with a pair of 30W tube amps and they apparently play loud enough to suit him.
When I was building my first set of panels I was very worried that 12x48 might be too small but my worries turned out to be unfounded, as they are much more efficient and dynamic than I had imagined. In fact, my 4ft2 panels can easily overpower the 10" woofers past a certain volume level so I later augmented them with subwoofers for when I want to play my jazz really loud.
I guess it's in the eye of the beholder.
Fair enough though, ESL's are never small but it sounds like the perfect size for me.
I was thinking 2mm D/S spacing myself and it really does sound like you have built the speaker I was thinking about building.
Great to hear you're happy with it.
It sounds like the perfect beginner size. If needed one can always build larger later on.
Fair enough though, ESL's are never small but it sounds like the perfect size for me.
I was thinking 2mm D/S spacing myself and it really does sound like you have built the speaker I was thinking about building.
Great to hear you're happy with it.
It sounds like the perfect beginner size. If needed one can always build larger later on.
Important question, only unpalatable answers. Only partly in my personal/direct experience but:
problems multiply as you try to go low. Right place, at least for me, is down where you can have a mixed bass, with all north of there handled by ESL. See my sig. otherwise, forgedaboutit.
But crossovers (even biamps) are major hurdles and doubly so blending cones and ESLs. Maybe sensible compromise on a first project is like classic Janzen: 1000+ Hz. You can use slow slopes since you can build an ESL to handle down to say, 250 and the cone woofer might do fine up to 2000. You'll get 80% of the ESL benefit.
problems multiply as you try to go low. Right place, at least for me, is down where you can have a mixed bass, with all north of there handled by ESL. See my sig. otherwise, forgedaboutit.
But crossovers (even biamps) are major hurdles and doubly so blending cones and ESLs. Maybe sensible compromise on a first project is like classic Janzen: 1000+ Hz. You can use slow slopes since you can build an ESL to handle down to say, 250 and the cone woofer might do fine up to 2000. You'll get 80% of the ESL benefit.
Assuming your building for the love of it
Dear MarkusG,
I assume you are building electrostatics as a hobby rather than to save money. If you are trying to save money by building electrostatics then I think this is a poorly considered move as I think you will need to build many designs to get the best of the engineering compromises, and second hand reconditioned units may be surprisingly cheap from the makers. For the rest of this email I will assume your building electrostatics for the fun building many is all part or the fun.
Personally I would first suggest buying an old broken electrostatic speaker and rebuilding it and then design around what you learned form looking at that speaker. Broken quads ESL 57's in the UK go for around 100 UKP, their construction and parts could be reused for your own designs, and the design is simple and a good set of compromises but aimed at approximately 20 Watt valve amplifiers (though the 30 Watt Quad 303 sounds lovely with them)
If I really wanted to build from scratch (first rather than doing a rebuild) I would first start with some tweeters from about 1kHz and up. Then I would build some mid ranges going down to around 400Hz. This should be crossed over using active cross overs, as passive cross overs are in my opinion cost savings only suitable for people planning mass production. This would be good enough to get most of the accuracy and stereo imaging benefits of electrostatics without tricky SPL issues and directivity issues with high frequencies that can be difficult to live with.
Having a pair of reconditioned Quad ESL 63's from the factory, If I was building the ultimate music solution, I would aim for a final solution for a low frequency of around 100Hz, This is because with higher cross over frequencies on my Quads I still prefer the electrostatic to moving coil speakers.
I would also recommend you look into Quads idea of having concentric rings, with delays for each ring. This idea could be nicely done, and since 24 bit 192Khz A/D converters exist and cheap around 30-40 Euro and the majority of my music is in digital form.
I would then have 5-6 D/A converters with a computer tuned delay each driving a separate power amp to each transformer for each stator ring. You could then compromise on bass extension for the center rings and compromise on treble extension on the outer, larger rings. This would nicely emulate a modern Quad ESL speaker and potentially beat it in both performance and sound dispersion. I think below 100Hz I would then cross over to the moving coils anyway as electrostatics for me are not some thing I want to pack away when not listening and the Quad ESL 63's are already quiet large.
All this said I think that
a full price dealer bought Quad 2905 might be good enough to prevent you feeling the need to do this and would not need a subwoofer.
The other thing I would consider is building a pair of electrostatic headphones.
So in summary 400Hz is good enough, but with my ESL 63's but I to put the cross over frequency appropriately at 100Hz as electrostatics sounds better. My final cross over design will have a switch so I can compromise between SPL and quality. I expect you will also find that different cross over frequencies are good for different volumes too.
Dear MarkusG,
I assume you are building electrostatics as a hobby rather than to save money. If you are trying to save money by building electrostatics then I think this is a poorly considered move as I think you will need to build many designs to get the best of the engineering compromises, and second hand reconditioned units may be surprisingly cheap from the makers. For the rest of this email I will assume your building electrostatics for the fun building many is all part or the fun.
Personally I would first suggest buying an old broken electrostatic speaker and rebuilding it and then design around what you learned form looking at that speaker. Broken quads ESL 57's in the UK go for around 100 UKP, their construction and parts could be reused for your own designs, and the design is simple and a good set of compromises but aimed at approximately 20 Watt valve amplifiers (though the 30 Watt Quad 303 sounds lovely with them)
If I really wanted to build from scratch (first rather than doing a rebuild) I would first start with some tweeters from about 1kHz and up. Then I would build some mid ranges going down to around 400Hz. This should be crossed over using active cross overs, as passive cross overs are in my opinion cost savings only suitable for people planning mass production. This would be good enough to get most of the accuracy and stereo imaging benefits of electrostatics without tricky SPL issues and directivity issues with high frequencies that can be difficult to live with.
Having a pair of reconditioned Quad ESL 63's from the factory, If I was building the ultimate music solution, I would aim for a final solution for a low frequency of around 100Hz, This is because with higher cross over frequencies on my Quads I still prefer the electrostatic to moving coil speakers.
I would also recommend you look into Quads idea of having concentric rings, with delays for each ring. This idea could be nicely done, and since 24 bit 192Khz A/D converters exist and cheap around 30-40 Euro and the majority of my music is in digital form.
I would then have 5-6 D/A converters with a computer tuned delay each driving a separate power amp to each transformer for each stator ring. You could then compromise on bass extension for the center rings and compromise on treble extension on the outer, larger rings. This would nicely emulate a modern Quad ESL speaker and potentially beat it in both performance and sound dispersion. I think below 100Hz I would then cross over to the moving coils anyway as electrostatics for me are not some thing I want to pack away when not listening and the Quad ESL 63's are already quiet large.
All this said I think that
a full price dealer bought Quad 2905 might be good enough to prevent you feeling the need to do this and would not need a subwoofer.
The other thing I would consider is building a pair of electrostatic headphones.
So in summary 400Hz is good enough, but with my ESL 63's but I to put the cross over frequency appropriately at 100Hz as electrostatics sounds better. My final cross over design will have a switch so I can compromise between SPL and quality. I expect you will also find that different cross over frequencies are good for different volumes too.
Yes, all projects are hobby projects.
I realised long ago that before you can start saving money on your builds you have to loose money during the learning phase. And the learning phase can be loooong...
My thoughts on this project is to build a simple ESL in a reasonable size.
Measure them and adjust with some filters.
Measure again and then decide on what kind of bass unit to use.
If all goes great and the panels can handle a 200Hz x-over I have a design waiting to be tried. If not, I'll come up with something else.
My main fears with crossing so low has been low sensitivity. Listening to people in this thread, it might not be an issue. Anyway it would be fun to try and see what I can come up with.
I realised long ago that before you can start saving money on your builds you have to loose money during the learning phase. And the learning phase can be loooong...
My thoughts on this project is to build a simple ESL in a reasonable size.
Measure them and adjust with some filters.
Measure again and then decide on what kind of bass unit to use.
If all goes great and the panels can handle a 200Hz x-over I have a design waiting to be tried. If not, I'll come up with something else.
My main fears with crossing so low has been low sensitivity. Listening to people in this thread, it might not be an issue. Anyway it would be fun to try and see what I can come up with.
Dear MarkusG,
The Quad ESL 57 is an interesting construction design (I haven't taken apart my working ESL 63's) and looks like it could be made in any decently equipped garage. For the tweeters and the bass units look like their built out of plastic sheet, which are then recessed using a mold to make the stator gap, with a thermo plastic in an oven maybe. They maybe where pre drilled, or drilled post molding. Then some thing like aluminium foil glued on maybe at the start or end.
Interestingly the Quad ESL 57 has the conductive stators on the inside for the treble and on the outside for the bass. This is a nice idea since the bass units are much less likely to arc and destroy the film.
Good luck, and don't discount buying a stator or two from ebay UK ("a buy it now" is available for a working treble stator for 70 UKP at the moment and a 99p bid on a broken one, also its not me thats selling) for construction ideas. They are much cheaper than Germany where I would not consider the idea and they might fetch a silly price.
I think full range is too ambitious (due to transformers) and would be far to directive in the treble, which is in my opinion the biggest failing for the Quad ESL 57 even with a treble and bass unit.
The Quad ESL 57 is an interesting construction design (I haven't taken apart my working ESL 63's) and looks like it could be made in any decently equipped garage. For the tweeters and the bass units look like their built out of plastic sheet, which are then recessed using a mold to make the stator gap, with a thermo plastic in an oven maybe. They maybe where pre drilled, or drilled post molding. Then some thing like aluminium foil glued on maybe at the start or end.
Interestingly the Quad ESL 57 has the conductive stators on the inside for the treble and on the outside for the bass. This is a nice idea since the bass units are much less likely to arc and destroy the film.
Good luck, and don't discount buying a stator or two from ebay UK ("a buy it now" is available for a working treble stator for 70 UKP at the moment and a 99p bid on a broken one, also its not me thats selling) for construction ideas. They are much cheaper than Germany where I would not consider the idea and they might fetch a silly price.
I think full range is too ambitious (due to transformers) and would be far to directive in the treble, which is in my opinion the biggest failing for the Quad ESL 57 even with a treble and bass unit.
I can tell you that the imaging quality of my CLSs got significantly better when I added the DD-15 and raised the lower frequency point on the CLSs to 140 Hz. If I raise the cross over point higher the DD-15 does not follow sufficiently well, If I lower the cross over point, the CLSs tend to have intermodulational distortions.
For Markus's plans as Mitch illustrates, just as well not to struggle to make it to that low freq crossover if there's no way to straddle the crossover point well (and an octave or two beyond) with both higher and lower drivers.
You'd be surprised how wonderful your sound will be just covering down to 1000 Hz with an ESL composed of two 5x5 inch panels angled at 15 degrees per side. Lots of 5x5 plastic frames out there are your dollar-store.
You'd be surprised how wonderful your sound will be just covering down to 1000 Hz with an ESL composed of two 5x5 inch panels angled at 15 degrees per side. Lots of 5x5 plastic frames out there are your dollar-store.
Last edited:
Love the idea of using picture frames. I think maybe smaller still might be a good idea too to make it less directive. Quad ESL57's have about 5" width for the treble panel but are considerably longer in length.
Dear MarkusG,
I assume you are building electrostatics as a hobby rather than to save money. If you are trying to save money by building electrostatics then I think this is a poorly considered move as I think you will need to build many designs to get the best of the engineering compromises, and second hand reconditioned units may be surprisingly cheap from the makers. For the rest of this email I will assume your building electrostatics for the fun building many is all part or the fun.
Personally I would first suggest buying an old broken electrostatic speaker and rebuilding it and then design around what you learned form looking at that speaker. Broken quads ESL 57's in the UK go for around 100 UKP, their construction and parts could be reused for your own designs, and the design is simple and a good set of compromises but aimed at approximately 20 Watt valve amplifiers (though the 30 Watt Quad 303 sounds lovely with them)
If I really wanted to build from scratch (first rather than doing a rebuild) I would first start with some tweeters from about 1kHz and up. Then I would build some mid ranges going down to around 400Hz. This should be crossed over using active cross overs, as passive cross overs are in my opinion cost savings only suitable for people planning mass production. This would be good enough to get most of the accuracy and stereo imaging benefits of electrostatics without tricky SPL issues and directivity issues with high frequencies that can be difficult to live with.
Having a pair of reconditioned Quad ESL 63's from the factory, If I was building the ultimate music solution, I would aim for a final solution for a low frequency of around 100Hz, This is because with higher cross over frequencies on my Quads I still prefer the electrostatic to moving coil speakers.
I would also recommend you look into Quads idea of having concentric rings, with delays for each ring. This idea could be nicely done, and since 24 bit 192Khz A/D converters exist and cheap around 30-40 Euro and the majority of my music is in digital form.
I would then have 5-6 D/A converters with a computer tuned delay each driving a separate power amp to each transformer for each stator ring. You could then compromise on bass extension for the center rings and compromise on treble extension on the outer, larger rings. This would nicely emulate a modern Quad ESL speaker and potentially beat it in both performance and sound dispersion. I think below 100Hz I would then cross over to the moving coils anyway as electrostatics for me are not some thing I want to pack away when not listening and the Quad ESL 63's are already quiet large.
All this said I think that
a full price dealer bought Quad 2905 might be good enough to prevent you feeling the need to do this and would not need a subwoofer.
The other thing I would consider is building a pair of electrostatic headphones.
So in summary 400Hz is good enough, but with my ESL 63's but I to put the cross over frequency appropriately at 100Hz as electrostatics sounds better. My final cross over design will have a switch so I can compromise between SPL and quality. I expect you will also find that different cross over frequencies are good for different volumes too.
Excellent advise ,especially when you consider what used Accustats are being sold for ...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.