I wonder if considering a rational problem solving approach for anyone with this sort of interest "building an ESL. Specifically a full range one that doesn’t require a conventional woofer to produce descent bass"; might save time, money and sanity?
Something like:
1. listen to plenty of ESL models contrasting pure "full range" designs like the Quad ESL-2912 versus woofer assisted hybrid designs like many ML's. Find out if the bass experience with panels is what you like (maybe better for acoustic then hard rock for some?).
Quad ESL-2912 bass can be terrific driven by the right amps, in the right room with some tracks as Stereophile reported: "Double basses were reproduced with tunefulness and subtle dynamic variations"
2. Decide on your ESL design wishes.
3. Buy used commercial ESL panels that allow you to combine panels in various configurations and only then: 4. build your own ESL panels.
5. When or if DIY panels sound better and are reliable over the years then sell off the old commercial panels.
Or is that all too boring?
Okay so you Still want to build a Giant (or Small) Full Range ESL?!!!
How Do You Plan On Driving that thing??
You have to ask yourself, "What does it take to get in to the lowend ?", Ya lots of surface area gets you there plus displacement, " But then what allows them to get so low and so clean ?!!! " ....................
Big Expensive Iron..............
That is the only way !!!
Or, Eventually learn HV amplifier circuitry and go the DD amp route
Some manufacture's like "Martin Logan" and maybe Quad not sure, go as far as using multiple compounded diaphragms (isobaric style) to gain more authoritive control at the super low frequency end of the spectrum some (ML) use a separate bass panel with a smaller width semi-fullrange panel
Well, That is about all I have on this without going in to actual construction tips, But at least this gives you somethings to look at when you are deciding how to go about doing your build.
Anything you decide we can help you.
Once you get there, Please!!! ..... Feel Free, to pop in here and Ask if you have any Questions at all about anything DIY ESL !!
We/I never get bored seeing Newcomers jump in to the world of DIY ESL's and succeed with their Own creation !!
Cheers and Good Luck !!!
jer
Thanks jer, I appreciate it.
I just called my local metal supplier today and I’m trying to get some perforated steel in order! I’m going off of Roger Sanders’ guidelines for now
As a matter of fact I do have one more question (and more to come):
What kind of transformer do I need to high fidelity? The audio transformers I look at never give a turn ratio, what gives? Blocked
I have never used those 70V constant line jobbers before, I have read Rod Elliot's article about using them but got nothing out of it.
After researching them I was going to get a few but never did.
First off they have a very low turns ratio, like 1:20 or so int hat range, and you would need at least 4 (1:80) of them if not 8 (1:160) to suffice enough efficiency to make the system enjoyable to listen to.
And at >$20 a pop they aren't cost effective at all !!
I don't know exactly what Rogers dimension guidelines are that you are following.
Just tell me your dimensions and we can go from there.
Anyhow take a look at CharlieM's site "The Jazzman's Blog" and his page explains the electronics involved very nicely, to simply get it up running very quickly, it is a lot to type out.
But not very complicated at all.
I will compile up all of the links I can on all of our transformer testing threads here in this forum as well so that way they, too, will be in one spot !!
If you wish to search for them they are in the time era of 2010 to 2013 or so, it was the last time I did any testing of any sorts that extensively !!
How it worked for me is I got lucky back in 2008 and got a super deal on some 35V 200 watt toroid transformers for $20 each at Parts Express, so I got 10 of them, was gonna build some amps.
Right About Exactly at that same time I checked back in to DIYAudio world just to See if anything new was going on with ESL's since I was thinking about getting them back out again after hiding in the closet for 5 years.
At this point I had just got back in town from a DUI after 10 months only to discover that my dear Mom had thrown out a bunch of stuff including my 14 stage power supply and stereo 6V6 and 6L6 power amps chassis's and more...... All Gone !!! :'(
And in November 2008 as i was working on my Very First 200v Class A amp to try and drive them with, I stumbled upon a conversation that Calvin and Charlie were having about using low voltage power toroid transformers as step-ups for ESL's !!!
And I thought "Well That is just Fine !!!!"
You see prior to that in 2003-2006 I was searching High and Low for a solution to that big iron all of the way down to that I had finally found a local source for some Large E-I cores material in bulk by the Lbs. Cheap !!!
At that point I didn't care what the grade was or if it had any for the most part !! He,he,he
But Because of my Issue and problem/s i never got as far a being able to go get some, But I had these nice cores, So, Lets try them out, and the testing began in 2010 here in the threads.
Check Charlies parts list he usually has part numbers of transformers that he is using or have found cheaper the better results than some others, i will be getting back to it more as we go as i am just waking up out of a long long Sieste' !!
Jazzman's DIY Electrostatic Loudspeaker Page
Cheers !!
jer
P.S. If you are planning on taking these full range down to 20Hz or even too 40Hz Flat, There are A LOT of things to consider, and the simple 1 or 2 core setup is not gonna get you there.
Check out the two links from the link in post#5 under transformers and that is where some of those were it started, I will compile the rest of the links later.
Step-up transformer design
My 2:nd ESL-attempt
After researching them I was going to get a few but never did.
First off they have a very low turns ratio, like 1:20 or so int hat range, and you would need at least 4 (1:80) of them if not 8 (1:160) to suffice enough efficiency to make the system enjoyable to listen to.
And at >$20 a pop they aren't cost effective at all !!
I don't know exactly what Rogers dimension guidelines are that you are following.
Just tell me your dimensions and we can go from there.
Anyhow take a look at CharlieM's site "The Jazzman's Blog" and his page explains the electronics involved very nicely, to simply get it up running very quickly, it is a lot to type out.
But not very complicated at all.
I will compile up all of the links I can on all of our transformer testing threads here in this forum as well so that way they, too, will be in one spot !!
If you wish to search for them they are in the time era of 2010 to 2013 or so, it was the last time I did any testing of any sorts that extensively !!
How it worked for me is I got lucky back in 2008 and got a super deal on some 35V 200 watt toroid transformers for $20 each at Parts Express, so I got 10 of them, was gonna build some amps.
Right About Exactly at that same time I checked back in to DIYAudio world just to See if anything new was going on with ESL's since I was thinking about getting them back out again after hiding in the closet for 5 years.
At this point I had just got back in town from a DUI after 10 months only to discover that my dear Mom had thrown out a bunch of stuff including my 14 stage power supply and stereo 6V6 and 6L6 power amps chassis's and more...... All Gone !!! :'(
And in November 2008 as i was working on my Very First 200v Class A amp to try and drive them with, I stumbled upon a conversation that Calvin and Charlie were having about using low voltage power toroid transformers as step-ups for ESL's !!!
And I thought "Well That is just Fine !!!!"
You see prior to that in 2003-2006 I was searching High and Low for a solution to that big iron all of the way down to that I had finally found a local source for some Large E-I cores material in bulk by the Lbs. Cheap !!!
At that point I didn't care what the grade was or if it had any for the most part !! He,he,he
But Because of my Issue and problem/s i never got as far a being able to go get some, But I had these nice cores, So, Lets try them out, and the testing began in 2010 here in the threads.
Check Charlies parts list he usually has part numbers of transformers that he is using or have found cheaper the better results than some others, i will be getting back to it more as we go as i am just waking up out of a long long Sieste' !!
Jazzman's DIY Electrostatic Loudspeaker Page
Cheers !!
jer
P.S. If you are planning on taking these full range down to 20Hz or even too 40Hz Flat, There are A LOT of things to consider, and the simple 1 or 2 core setup is not gonna get you there.
Check out the two links from the link in post#5 under transformers and that is where some of those were it started, I will compile the rest of the links later.
Step-up transformer design
My 2:nd ESL-attempt
Last edited:
Another red flag on that transformer is the core size. Small core = poor bass performance. Toroidal transformers also tend to have small cores, which often make them suitable only for midrange and up.
Unfortunately, good transformers tend to be expensive. Output transformers from tube amplifiers are the closest "normal transformer" you're likely to find for full-range ESL use. Periodically you can find old ones for sale. New, something along these lines is likely what you're after:
http://www.hammondmfg.com/pdf/5c0051-52.pdf
Impedance ratio is often stated instead of turns ratio, but you can calculate from that.
You definitely make some good points. Participating in this thread has reminded me over and over how much there is to learn, how many details are involved in doing this from scratch, and how many prototypes I built to figure a lot of different things out.
Over the years, I also bought several kinds of used panels from ebay to judge my work against what I thought were some of the better designs (Janszen, Acoustat, Quad). In this instance, a few Acoustat panels are probably most like what the original poster is trying to build.
If you can get in front of some, listening to full range ESLs to judge their limitations is definitely worth doing also. Quads in particular, since ESL-63 and newer include circuits to limit their drive to levels that are safe for the speaker. Depending on where you live, that's getting harder to do though.
Last edited:
As discussed previously, you typically don't find a 200:1 ESL transformer at all. When you do, it's likely for bass only. Full range commercial ESL's have typically used two transformers. And rarely does anyone try to design one for use at 20 Hz, since ESLs can't typically make appreciable output at 20 Hz, transformers that can handle those frequencies get huge at high power levels, and you wind up with conflicting needs when trying to maintain high frequencies, low frequencies, and high power.
The Martin Logan CLS used one transformer, but it's the only recent speaker I can think of that did. I'm not sure if they'll sell you a transformer without owning their speakers or if they still make parts for that model, but might be worth a call.
From what I remember of tests from years ago, the Hammond 1650E was roughly comparable to a MagneTek Triad S-142A. Seems like Sanders suggested S-142As in his book.
Last time I was seriously working with commercially available transformers, I liked 3 Hammond 1609's. With both 0-4 ohm windings in parallel as the primary and the full blue-brown as secondary, you get 50:1. With all 3 transformers in parallel on the primary side, and all 3 full secondaries in series, that gets you to 150:1 (red wire of center transformer is center tap in this configuration). I don't remember if I tested distortion, but they sounded good. They've also increased in price since I was doing this, so not exactly a cheap route anymore. I was using this setup with an electrically segmented panel also, so don't know if it would have sufficient high frequency extension on a full unsegmented panel.
The Martin Logan CLS used one transformer, but it's the only recent speaker I can think of that did. I'm not sure if they'll sell you a transformer without owning their speakers or if they still make parts for that model, but might be worth a call.
From what I remember of tests from years ago, the Hammond 1650E was roughly comparable to a MagneTek Triad S-142A. Seems like Sanders suggested S-142As in his book.
Last time I was seriously working with commercially available transformers, I liked 3 Hammond 1609's. With both 0-4 ohm windings in parallel as the primary and the full blue-brown as secondary, you get 50:1. With all 3 transformers in parallel on the primary side, and all 3 full secondaries in series, that gets you to 150:1 (red wire of center transformer is center tap in this configuration). I don't remember if I tested distortion, but they sounded good. They've also increased in price since I was doing this, so not exactly a cheap route anymore. I was using this setup with an electrically segmented panel also, so don't know if it would have sufficient high frequency extension on a full unsegmented panel.
Last edited:
Hi,
Matt and Ger already listed the most points You need to know.
I just liked to add/underline a few tips.
Generally its a good advice to select a as large membrane size as possible .... but within the given sensible size constraints.
15" width is probabely the upper level of what I´d call reasonable.
Keep in mind that with growing width of the panel the listening distance where the panel develops a natural stage impression also increases.
For a 15" panel calculate with 4-5m distance and 2-3m for a 10" wide panel.
Sitting closer the panels sound rather like oversized headphones.
The best compromise will probabely be between 8"-12" panel width.
Since You are at the beginning of Your ESL adventure I recommend to follow proven concepts that guarantee a high chance of success.
The imho best in this regard is the use of a flat segmented wire stator, because of:
- easyness and low cost to source building materials
- unburdons You from the important(!) requirement to create a safe and reliable and well functioning stator insulation.
- building requires no special skills, but working precise and patient and every now and then following the tips of the experienced (even if they might differ with Your intuition)
- building doesn´t require highly specialized jigs and tools ... a large plane table will be Your most important and used tool
- allows for easiest ways to segment ... besides the positive effects on distribution character and amplitude response this effectively reduces the burdon on the amplifier. Non-segmented panels when driven through high-quaility audio trannies can be a real killer-load to the majority of amps.
As general tips (probabely repeating Matt & Ger):
- don´t chase for lowest resonance frequency.
A ESL is no prime performer below ~150Hz.
With every Hz lower the problems increase exponentially and You have to seriously compromise on performance.
A reasonable compromise could be the subwoofer assisted panel with a not too low fs of 70-80Hz, x-overed around 100-120Hz.
This allows to use off-the-shelf power transformers instead of spezialized audio trannies which are cheaper, easy to source and very reliable to use (most tube trannies -if good enough regarding bandwidth and transformation factor- aren´t designed to handle voltages in the kV range and may die due to internal flash-overs. Power trannies are specced to 4-5kVrms, but may lack on primary inductance for fullrange use)
- thermal treatment of the membrane is highly recommended.
It leads to consistent and stable values of Fs as it preages the membrane at the same.
It is always combined with mechanical (pre-)tensioning.
The amount of mechanical pre-tensioning depends on what kind of panel resp. how high the chosen Fs should be.
Low tensioning results in low Fs ... useable for large fullrange panels.
Thermal treatment may even rise the tension and resultant Fs here.
High mechanical tension can result in Fs of up to 250hz (elongation of the film of 1.5% and more), useable for Hybrids.
Thermal treatment reduces the tension a bit here, but releases the internal mechanical stress and preages the membrane, so it settles almost instantly on its final value.
Only mechanically stretched stats like MLs require a long breaking-in time till the membrane settles to the designed Fs.
Fresh unpacked from the cartonage they perform technically and audibly detuned.
Just m2c.
jauu
Calvin
Matt and Ger already listed the most points You need to know.
I just liked to add/underline a few tips.
Generally its a good advice to select a as large membrane size as possible .... but within the given sensible size constraints.
15" width is probabely the upper level of what I´d call reasonable.
Keep in mind that with growing width of the panel the listening distance where the panel develops a natural stage impression also increases.
For a 15" panel calculate with 4-5m distance and 2-3m for a 10" wide panel.
Sitting closer the panels sound rather like oversized headphones.
The best compromise will probabely be between 8"-12" panel width.
Since You are at the beginning of Your ESL adventure I recommend to follow proven concepts that guarantee a high chance of success.
The imho best in this regard is the use of a flat segmented wire stator, because of:
- easyness and low cost to source building materials
- unburdons You from the important(!) requirement to create a safe and reliable and well functioning stator insulation.
- building requires no special skills, but working precise and patient and every now and then following the tips of the experienced (even if they might differ with Your intuition)
- building doesn´t require highly specialized jigs and tools ... a large plane table will be Your most important and used tool
- allows for easiest ways to segment ... besides the positive effects on distribution character and amplitude response this effectively reduces the burdon on the amplifier. Non-segmented panels when driven through high-quaility audio trannies can be a real killer-load to the majority of amps.
As general tips (probabely repeating Matt & Ger):
- don´t chase for lowest resonance frequency.
A ESL is no prime performer below ~150Hz.
With every Hz lower the problems increase exponentially and You have to seriously compromise on performance.
A reasonable compromise could be the subwoofer assisted panel with a not too low fs of 70-80Hz, x-overed around 100-120Hz.
This allows to use off-the-shelf power transformers instead of spezialized audio trannies which are cheaper, easy to source and very reliable to use (most tube trannies -if good enough regarding bandwidth and transformation factor- aren´t designed to handle voltages in the kV range and may die due to internal flash-overs. Power trannies are specced to 4-5kVrms, but may lack on primary inductance for fullrange use)
- thermal treatment of the membrane is highly recommended.
It leads to consistent and stable values of Fs as it preages the membrane at the same.
It is always combined with mechanical (pre-)tensioning.
The amount of mechanical pre-tensioning depends on what kind of panel resp. how high the chosen Fs should be.
Low tensioning results in low Fs ... useable for large fullrange panels.
Thermal treatment may even rise the tension and resultant Fs here.
High mechanical tension can result in Fs of up to 250hz (elongation of the film of 1.5% and more), useable for Hybrids.
Thermal treatment reduces the tension a bit here, but releases the internal mechanical stress and preages the membrane, so it settles almost instantly on its final value.
Only mechanically stretched stats like MLs require a long breaking-in time till the membrane settles to the designed Fs.
Fresh unpacked from the cartonage they perform technically and audibly detuned.
Just m2c.
jauu
Calvin
A few things you need to know about ESLs, especially full range. The sound pressure is directly proportional to the ac current flowing through the stators. Since the stators behave as a capacitor, if you want a reasonable SPL at low frequencies, you need large stators. Something about 1 square metre is reasonable for 50Hz.
There are three basic forms of ESL that have a flat frequency response, so do not need any extra equalization. Very briefly (I can give much more detail if you want)…
Plane source: These require a very large panel with no segmentation and are used with the listener in the near field – close to the panel. Think about an ESL occupying an entire wall – not good for stereo. The only practical examples are near field monitors and headphones. A lot of DIY ESLs are built as a single panel, but they are usually not big enough to give a flat frequency response. These things behave electrically like a capacitance of about 1-2 nF.
Point source: These are built with circular segments connected to each other by inductors to form an LC transmission line. The quad 63 family of ESLs operate this way. The LC transmission line means that the behave electrically like a resistor.
Line source: These are built with a tall narrow panel divided into 20 or more vertical segments/strips connected to each other by resistors. These are a relatively recent development, and I know of only one commercial ESL like this – a kit sold by ER audio in Australia. A good number of line-source ESLs of different designs have been built successfully by DIYers here in the last 10 years or so, and some commercial ESLs are built in a similar way, but not enough segments.
If you are serious about a full range ESL, then the line source using a panel about 2m tall and 0.5 m wide is a good choice. Jazzman also used this design as the basis for his hybrid ESL – a good choice too. I suggest searching the DIY forum for segmented ESLs - there are a lot of good posts with helpful links to software for designing them.
Because the three different designs behave quite differently as an electrical load, the step-up transformers must be different for the different designs…
For the plane source, for which the load is capacitive, you need very low inductance transformers – so 50VA toroidal power transformers work well. For a full range ESL, you will struggle to get step up ratios large than about 60 before the transformer inductance and the ESL capacitance start attenuating the high frequencies.
For the point source ESL, the transformers need a low capacitance, so small multilayered C-cores are used – see the Quad 63 transformers. Step-up ratios as high as 250 are achievable.
For the line source, you need something midway, capacitance below 100pF, inductance below 500 mH. Optimized, you can get step up ratios of 150. Easiest build is a stack of 10 or 15 VA toroid’s (230V windings in series). Do not use toroid’s with 230V/115V option – these have a very high capacitance.
Finally – read what you can find written about ESLs. I found Baxandall’s chapter very good, but the Books by Sanders and Wagner are not as helpful as you might hope. If you send me a PM with your email address I can forward some papers.
There are three basic forms of ESL that have a flat frequency response, so do not need any extra equalization. Very briefly (I can give much more detail if you want)…
Plane source: These require a very large panel with no segmentation and are used with the listener in the near field – close to the panel. Think about an ESL occupying an entire wall – not good for stereo. The only practical examples are near field monitors and headphones. A lot of DIY ESLs are built as a single panel, but they are usually not big enough to give a flat frequency response. These things behave electrically like a capacitance of about 1-2 nF.
Point source: These are built with circular segments connected to each other by inductors to form an LC transmission line. The quad 63 family of ESLs operate this way. The LC transmission line means that the behave electrically like a resistor.
Line source: These are built with a tall narrow panel divided into 20 or more vertical segments/strips connected to each other by resistors. These are a relatively recent development, and I know of only one commercial ESL like this – a kit sold by ER audio in Australia. A good number of line-source ESLs of different designs have been built successfully by DIYers here in the last 10 years or so, and some commercial ESLs are built in a similar way, but not enough segments.
If you are serious about a full range ESL, then the line source using a panel about 2m tall and 0.5 m wide is a good choice. Jazzman also used this design as the basis for his hybrid ESL – a good choice too. I suggest searching the DIY forum for segmented ESLs - there are a lot of good posts with helpful links to software for designing them.
Because the three different designs behave quite differently as an electrical load, the step-up transformers must be different for the different designs…
For the plane source, for which the load is capacitive, you need very low inductance transformers – so 50VA toroidal power transformers work well. For a full range ESL, you will struggle to get step up ratios large than about 60 before the transformer inductance and the ESL capacitance start attenuating the high frequencies.
For the point source ESL, the transformers need a low capacitance, so small multilayered C-cores are used – see the Quad 63 transformers. Step-up ratios as high as 250 are achievable.
For the line source, you need something midway, capacitance below 100pF, inductance below 500 mH. Optimized, you can get step up ratios of 150. Easiest build is a stack of 10 or 15 VA toroid’s (230V windings in series). Do not use toroid’s with 230V/115V option – these have a very high capacitance.
Finally – read what you can find written about ESLs. I found Baxandall’s chapter very good, but the Books by Sanders and Wagner are not as helpful as you might hope. If you send me a PM with your email address I can forward some papers.