I agree. Although not impossible to do, the structure holding all the elements would be a real challenge. I do like the single sheet of perf segmented by double sided tape. I have no idea how it would perform in the real world, maybe v-bro or another aspiring ESL builder will give it a try.
Is that the one with different colored wires that looks like a spiderweb?
I wonder if that could be done today with digital delay and an amplifier for each section?
It sure could be done, might even work much better than the esl63's 'analog' solution. I agree that this network of transformers, coils,resistors and ceramic C's has it's weaknesses.
Another way would be to build the delaylines at line signal level instead of after the transformer where it has to handle several kV.
But the price is a separate amplifier+stepup trannie for each section.
6 or more amps + transformers for each channel might take the fun out of DIY for most of us...
Interesting fellas!
But why not use just two segmentations, a larger say 2' midrange ring around a say 3" diameter round treble unit. The ring produces from say 200-500hz upto 2.5-3Khz and the centre produces from 2Khz-up...
Would perhaps make it easier to find step-up tannies too, only having to handle the narrow bandwidth.
The ESL-57's I have I got for free from an old couple that bought new amps and speakers, they knew I loved to experiment with them. Funny thing is they gave me three units, two of them still worked fine, one had membrane problems.. Yesterday I put the spare step-up device in the oven to melt out the sort of paraffin grease. It looks like a pretty neat tranny
in there...
Maybe I can give it a try with this one, and either use one from another panel later, or buy another single unit from ebay...
Any thoughts on this?
On the other hand I don't care if it is necessary to use more amps and trannies to get good results, and I'm very interested in a possibility to put delay lines (if necessary) in the line signal path.
Thanks again..
But why not use just two segmentations, a larger say 2' midrange ring around a say 3" diameter round treble unit. The ring produces from say 200-500hz upto 2.5-3Khz and the centre produces from 2Khz-up...
Would perhaps make it easier to find step-up tannies too, only having to handle the narrow bandwidth.
The ESL-57's I have I got for free from an old couple that bought new amps and speakers, they knew I loved to experiment with them. Funny thing is they gave me three units, two of them still worked fine, one had membrane problems.. Yesterday I put the spare step-up device in the oven to melt out the sort of paraffin grease. It looks like a pretty neat tranny
in there...
Maybe I can give it a try with this one, and either use one from another panel later, or buy another single unit from ebay...
Any thoughts on this?
On the other hand I don't care if it is necessary to use more amps and trannies to get good results, and I'm very interested in a possibility to put delay lines (if necessary) in the line signal path.
Thanks again..
Attachments
It might work. My own preference would be a larger treble unit, I like the sizzle. The larger size is a good idea, bigger panel = plays louder. I may be old, but I still like to rock.
Most ESL's are a form of line source. I don't know the reason for it, but they're very common. From a practical standpoint, it's usually easier to fit a pair of tall skinny speakers in a room.That's quite a find. Have you got them working? Here's a page from Sheldon's website that you may find interesting. Take a look at his Quad ESL Book for service info and schematics.
Yes I've got them working.. Only thing is I find the treble panels really great (not to say incredible 
), but the bass panels don't quite thrill me.
I changed the bass panels for a dynamic speaker in a transmissionline enclosure (got these subs from a collegue who designed them). The basspeaker measures 10" and is from the brand "precision devices". My collegue had the drivers custom built with custom thiele and small parameters (and rubber surround, a real hifi speaker). He sold me the drivers for 150 EUR. and I got the cabinets for free...
It's the sound I'm listening to at this very moment that inspires me to experiment with it more myself.
I've looked at Sheldons page, he also uses a transmissionline sub with the DIY ESL 1.0 , very interesting info on the quads too!
I always say the more you study, the smoother the build...
Here's a pic from my present speakers (only the ESL and sub underneath are connected offcourse):
), but the bass panels don't quite thrill me.I changed the bass panels for a dynamic speaker in a transmissionline enclosure (got these subs from a collegue who designed them). The basspeaker measures 10" and is from the brand "precision devices". My collegue had the drivers custom built with custom thiele and small parameters (and rubber surround, a real hifi speaker). He sold me the drivers for 150 EUR. and I got the cabinets for free...
It's the sound I'm listening to at this very moment that inspires me to experiment with it more myself.
I've looked at Sheldons page, he also uses a transmissionline sub with the DIY ESL 1.0 , very interesting info on the quads too!
I always say the more you study, the smoother the build...
Here's a pic from my present speakers (only the ESL and sub underneath are connected offcourse):
Attachments
Regarding section size....
I would think that Peter Walker et all would have looked into this thoroughly. I would suggest that if you fashioned your surface radiating areas using the Quad 63 as a model you could not go wrong. Walker was after all one of the sharpest knives in the ESL drawer.
I think that making the middle section larger will result in the opposite result as it will also behave as a midrange size driver. So I think that you want to keep it small.
Remember that if you build this like a Quad 63 then you are fashioning a point source. This means that you will be exciting all possible room modes (floor, ceiling and side walls) are you sure you really want to do that? I suggest a horizontally segmented line source. This way you will get the desired dispersion that you are after without exciting the floor and ceiling modes of the room. With a line source you can make your tweeter section as narrow as you want to obtain excellent dispersion at high frequency and not be concerned with area as the line is long. Walker did not have this luxury as he chose a point source and so had to make the central section larger than is desireable for wide dispersion at high frequencies.
The fewer room modes that you bring into affect the better. That is one of the reasons a panel like an Acoustat 0ne plus 0ne can image so very well (when you are in the sweet spot). The wider the horizontal dispersion the more you will bring side wall reflections into play.
Dipoles like horns have controlled dispersion and so direct more of the wavefront directly at the listner which minimises room interaction as well as increasing apparent output level at the listening position. That's why ESL's sound subjectively louder than they measure. Just some thoughts to keep in mind. Regards Moray James
I would think that Peter Walker et all would have looked into this thoroughly. I would suggest that if you fashioned your surface radiating areas using the Quad 63 as a model you could not go wrong. Walker was after all one of the sharpest knives in the ESL drawer.
I think that making the middle section larger will result in the opposite result as it will also behave as a midrange size driver. So I think that you want to keep it small.
Remember that if you build this like a Quad 63 then you are fashioning a point source. This means that you will be exciting all possible room modes (floor, ceiling and side walls) are you sure you really want to do that? I suggest a horizontally segmented line source. This way you will get the desired dispersion that you are after without exciting the floor and ceiling modes of the room. With a line source you can make your tweeter section as narrow as you want to obtain excellent dispersion at high frequency and not be concerned with area as the line is long. Walker did not have this luxury as he chose a point source and so had to make the central section larger than is desireable for wide dispersion at high frequencies.
The fewer room modes that you bring into affect the better. That is one of the reasons a panel like an Acoustat 0ne plus 0ne can image so very well (when you are in the sweet spot). The wider the horizontal dispersion the more you will bring side wall reflections into play.
Dipoles like horns have controlled dispersion and so direct more of the wavefront directly at the listner which minimises room interaction as well as increasing apparent output level at the listening position. That's why ESL's sound subjectively louder than they measure. Just some thoughts to keep in mind. Regards Moray James
i have a few thoughts after reading this thread ..
Why do you guys all settle for CHEAP STUFF
i mean, i ain't doing an ESL project because i want it to perform CHEAPLY ... i want it to be the BEST POSSIBLE
of what i will be able to do ..
So why bother with stupid old E-Core transfos
Just go out and pay some real money for Plitron ESL's
or other correctly designed ESL toroidal transfos
and be set with it ...
it's not like it's a 1 use shot transfos, u can reuse em in all ur projects ( if you discard the one before of course)
Then, i am considering a design with wires,
simply because i will not accept the compromises of a perforated metal shee(i)t
i do not care about how it looks, because i am making an ESL, not a fashionable object to me ...
Then i truly think that perforated sheet is only good for first time try, where u are only experiencing with cheap components and EXPECT a cheap result ..
i believe that the cheapest element in ur design will result in the end quality !!
Then, why are everybody using wires on the vertical with long narrrow designs ?
The way i have thought of what i want is a narrow
( probably 3 to 5" wide ) assembly with wires running
on the horizontal plane so that they can remain very stiff because of their total length , without having to resort to extreme tensions,
steel frame a la piano style but all around ( obviously)
with a few vertical division of some sort( probably teflon strips ) at desired intervals
then a large wing span + back wings
of a large enough size so that i can lower dipole effect
( 24 to 36" total width )
Teflon coated steel wires in approx 1mm total diameter
( maybe aluminum wires if i can find..or SS if not too $$)
i wouldn't use anything else than teflon on a high voltage project ( i use it everywhere since it's probably the best dieletric matter i can find )
what do you guys think ?
Why do you guys all settle for CHEAP STUFF
i mean, i ain't doing an ESL project because i want it to perform CHEAPLY ... i want it to be the BEST POSSIBLE
of what i will be able to do ..
So why bother with stupid old E-Core transfos
Just go out and pay some real money for Plitron ESL's
or other correctly designed ESL toroidal transfos
and be set with it ...
it's not like it's a 1 use shot transfos, u can reuse em in all ur projects ( if you discard the one before of course)
Then, i am considering a design with wires,
simply because i will not accept the compromises of a perforated metal shee(i)t
i do not care about how it looks, because i am making an ESL, not a fashionable object to me ...
Then i truly think that perforated sheet is only good for first time try, where u are only experiencing with cheap components and EXPECT a cheap result ..
i believe that the cheapest element in ur design will result in the end quality !!
Then, why are everybody using wires on the vertical with long narrrow designs ?
The way i have thought of what i want is a narrow
( probably 3 to 5" wide ) assembly with wires running
on the horizontal plane so that they can remain very stiff because of their total length , without having to resort to extreme tensions,
steel frame a la piano style but all around ( obviously
)with a few vertical division of some sort( probably teflon strips ) at desired intervals
then a large wing span + back wings
of a large enough size so that i can lower dipole effect
( 24 to 36" total width )
Teflon coated steel wires in approx 1mm total diameter
( maybe aluminum wires if i can find..or SS if not too $$)
i wouldn't use anything else than teflon on a high voltage project ( i use it everywhere since it's probably the best dieletric matter i can find )
what do you guys think ?
Well my room is triangular 10x10 mtr. square floor, 4.5 mtr. high (tip of the triangle), don't know if room modes are much of a problem as I'm sitting much nearer to the panels as the distance to the "ceiling".
On the other hand I would like to keep them universal to many other rooms as well.
My aim is to recreate at least what I have now, preferrably a lot more compact for not too much money. I know a lot of people that will be interested in them...
How would a good small (like a 6") dynamic speaker in a vented enclosure do (upto 1khz) in combination with an ESL panel?
All this I inquire for a couple of friends with very small rooms, in my place I have space for double ESL-57's if I want to..... But I would prefer not to sacrifice a lot of space....
In my room I wonder how a ripole sub would do the job?
Is a '57 a line-source? I must admit the sound very much has a "sweet spot" and would prefer it if things would spread a little bit more....
Tanks to all!
On the other hand I would like to keep them universal to many other rooms as well.
My aim is to recreate at least what I have now, preferrably a lot more compact for not too much money. I know a lot of people that will be interested in them...
How would a good small (like a 6") dynamic speaker in a vented enclosure do (upto 1khz) in combination with an ESL panel?
All this I inquire for a couple of friends with very small rooms, in my place I have space for double ESL-57's if I want to..... But I would prefer not to sacrifice a lot of space....
In my room I wonder how a ripole sub would do the job?
Is a '57 a line-source? I must admit the sound very much has a "sweet spot" and would prefer it if things would spread a little bit more....
Tanks to all!
JinMTVT said:
Agreed, but all I want is to diy most parts, also to learn from it, not only to be cheaper off. Some very cheap items have surprised me more than many expensive ones....After all this IS diyaudio.com isn't it?
JinMTVT said:
Is it proved that E-cores are inferior to toroidal trannies?
JinMTVT said:
My ESL-57's are (I believe) made with perf. sheet stators, why are wires that much better? I don't have much to complain about their performance.
JinMTVT said:
Neither do I, but many of my friends do....
I'm starting to look at a different design for myself than for the others....
JinMTVT said:
Then, why are everybody using wires on the vertical with long narrrow designs ?
The way i have thought of what i want is a narrow
( probably 3 to 5" wide ) assembly with wires running
on the horizontal plane so that they can remain very stiff because of their total length , without having to resort to extreme tensions,
steel frame a la piano style but all around ( obviously
Isn't a design like that troubled by room modes even more? I'm not sure what weighs heavier in the end though....
hi V-Bro
of course a well designed toroidal will outperform a well designed E-Core everytime ..
it is more expensive to do, that is the only reason
why it is less used ...
I am not too sure that i understand the thing about tensionned wires VS room modes??
please elaborate
Then, have you read Sanders's book ?
Have you done research and maths about wire VS perf.?
wires will outperform perf. in everything except construction simplicity ( alot more time consuming and requries an extreme bracing )
U cannot match capacitance and surace distribution of wire with a perf.
And then, to get a perf. to perform as good on voltage VS D/S spacing is nearly impossible, or tenfold more $$$
( perf always have some edges, not uniform coating...
u must know about all those , and that is where wire are almost perfects... )
what do you think ?
of course a well designed toroidal will outperform a well designed E-Core everytime ..
it is more expensive to do, that is the only reason
why it is less used ...
I am not too sure that i understand the thing about tensionned wires VS room modes??
please elaborate
Then, have you read Sanders's book ?
Have you done research and maths about wire VS perf.?
wires will outperform perf. in everything except construction simplicity ( alot more time consuming and requries an extreme bracing )
U cannot match capacitance and surace distribution of wire with a perf.
And then, to get a perf. to perform as good on voltage VS D/S spacing is nearly impossible, or tenfold more $$$
( perf always have some edges, not uniform coating...
u must know about all those
, and that is where wire are almost perfects... )what do you think ?
Even then, i believe that Roger R. Sanders have gone the perf. way because he didn't want to take time and propose a tenfold more complex wired assembly...
but again if designed properly it will outperfom perf.
unless something other than the stator is limiting the ESL performance ...
also you cannot have as high rigidity with a perf. sheet VS a wire stator without sacrifying size of partitions
else it would require a very thick sheet wich would affect air passage alot i believe ..
wires are almost perfect in that way!
but again if designed properly it will outperfom perf.
unless something other than the stator is limiting the ESL performance ...
also you cannot have as high rigidity with a perf. sheet VS a wire stator without sacrifying size of partitions
else it would require a very thick sheet wich would affect air passage alot i believe ..
wires are almost perfect in that way!
When you guys are stretching your diaphragm,
are you doing it more on a particular axis or pretty much equal all around ?
cause with a long rectangle vertical form,
i'd tend to think that it would require a more tensed film on the vertical axis ... what u think ?
is it possible to stretch mylar almost only on 1 axis
( using just enough pressure on the other one to make it flat ) ??
are you doing it more on a particular axis or pretty much equal all around ?
cause with a long rectangle vertical form,
i'd tend to think that it would require a more tensed film on the vertical axis ... what u think ?
is it possible to stretch mylar almost only on 1 axis
( using just enough pressure on the other one to make it flat ) ??
JinMTVT said:
I don't know if you want to have it segmented, but I understood if you do so and create a vertical dispersion your floor and ceiling modes will come to play part. (instead of the left and right walls).
JinMTVT said:
No I haven't, maybe a good idea, thanks. Only thing I worry about with wire is that it seems unsteady, they are said to loosen tension over time and I was more thinking in the direction of strips of metal. Perf. stators just seemed more steady and the easy build factor came in handy...
I'm not in the market for many dollars/enormous effort for minimal improvements, I'm just curious to shortcuts to a good result with easy reproducing possibilities. This should be possible within a reasonable price margin, providing I put in more effort. I don't care much about the effort needed for the first build, if it leads me to the knowledge to reproduce it with less effort.
I think there will be more readers interested in this, won't you agree?
As for your latest post, I think so too. Tension is naturally higher over a smaller area, have no experience with this though....
nice!!
Well i will show you a 3d sketch when i'll have some time to do it ...
i thought about using some big 4130 bars and machine
( milling) the holders for the wires in it ...CNC of course
( talking abou <1mm distance between each wire here)
i wouldn't be any stressed about wire loosening up with time..if you make ur assembly adjustable neway
one could only torque it up a bit once in a while when you feel at touch that something has loosen,
but steel or SSteel wires shouldn't have any problem on that side once tensionned properly
i believe that positionning the wires with the right tension when isntalling is more important if you want a correclty distributed tension on all the wires !
i'll have to see about that
Well i will show you a 3d sketch when i'll have some time to do it ...
i thought about using some big 4130 bars and machine
( milling) the holders for the wires in it ...CNC of course
( talking abou <1mm distance between each wire here)
i wouldn't be any stressed about wire loosening up with time..if you make ur assembly adjustable neway
one could only torque it up a bit once in a while when you feel at touch that something has loosen,
but steel or SSteel wires shouldn't have any problem on that side once tensionned properly
i believe that positionning the wires with the right tension when isntalling is more important if you want a correclty distributed tension on all the wires !
i'll have to see about that
mixing it all together in one reply...
They are pretty good transformers. But I would suggest to pour the stuff back in, as it will do a great job preventing electrical breakdown (it's bee wax).
When you run only the tweeter panels, be sure to load the bass panel connections with a dummy load. Without, very high peak voltages can build up here and the transformer is very prone to failure. Well, you have a spare..
That's my experience too, I completely rebuild a set of 57's. The mids and the highs are among the best I ever heard on any esl. But the bass response is horrible.
First of all, the 57 trannies are not EI cores but double C-cores.
Second, age has little to do with quality. In fact, I consider the 57 tranfo's to be among the best out there. Only problem is the large difference between stepup for the mid/high and bass section, limiting their practical use.
Toroids are not always better. A toroid is better in low leakage and low interwinding caps which gives it better high frequency performance. But they tend to saturate very quickly at low freqs, distort more in the lows, making them less suited for fullrange panels. C-cores are imho the best of both worlds here. EI cores are less suited in all respects.
In fact, teflon is less suited than pvc for stator wires. It will give a far less sensitive panel, due to the lower dielectric constant (~2, pvc has 3.5-4.5).
You can sonsider the esl panel as a number of capacitors in series: from wire core to wire outer surface, from there to the membrame, and again the same to the other stator.
What matters (drives the membrame) is the AC field strength in between the outer wire surfaces. A lower dielectric constant of the isolation (as in teflon) will 'waste' more of the voltage difference in the wireisolation resulting in less field strength to drive the membrame.
I have bad experiences with combining closed enclosures + esl, I would go for the open baffle thing. But it's much more expensive. TL's work out fine too.
No, it's too small. A line source will give a cilindrical radiation pattern when it is large relative to the room heigth. It then 'reflects' in the floor and ceiling making a seemingly endless long source, giving cilindrical radiation.
The 57's tweeter strip is a bit too wide to give good horizontal dispersion. Almost all esl's have this problem. But narrowing the strip to improve dispersion would result in too little high output. I've been thinking about optimizing segmentation for dispersion and use active eq to straighten things out.
Depends on the construction, solid wire core or not, how much tension etc. Should not be a problem.
Plate stators are difficult to isolate, I don't have any experience with that but heard good things about powder coating in several layers.
They are pretty good transformers. But I would suggest to pour the stuff back in, as it will do a great job preventing electrical breakdown (it's bee wax).
When you run only the tweeter panels, be sure to load the bass panel connections with a dummy load. Without, very high peak voltages can build up here and the transformer is very prone to failure. Well, you have a spare..
That's my experience too, I completely rebuild a set of 57's. The mids and the highs are among the best I ever heard on any esl. But the bass response is horrible.
First of all, the 57 trannies are not EI cores but double C-cores.
Second, age has little to do with quality. In fact, I consider the 57 tranfo's to be among the best out there. Only problem is the large difference between stepup for the mid/high and bass section, limiting their practical use.
Toroids are not always better. A toroid is better in low leakage and low interwinding caps which gives it better high frequency performance. But they tend to saturate very quickly at low freqs, distort more in the lows, making them less suited for fullrange panels. C-cores are imho the best of both worlds here. EI cores are less suited in all respects.
In fact, teflon is less suited than pvc for stator wires. It will give a far less sensitive panel, due to the lower dielectric constant (~2, pvc has 3.5-4.5).
You can sonsider the esl panel as a number of capacitors in series: from wire core to wire outer surface, from there to the membrame, and again the same to the other stator.
What matters (drives the membrame) is the AC field strength in between the outer wire surfaces. A lower dielectric constant of the isolation (as in teflon) will 'waste' more of the voltage difference in the wireisolation resulting in less field strength to drive the membrame.
I have bad experiences with combining closed enclosures + esl, I would go for the open baffle thing. But it's much more expensive. TL's work out fine too.
No, it's too small. A line source will give a cilindrical radiation pattern when it is large relative to the room heigth. It then 'reflects' in the floor and ceiling making a seemingly endless long source, giving cilindrical radiation.
The 57's tweeter strip is a bit too wide to give good horizontal dispersion. Almost all esl's have this problem. But narrowing the strip to improve dispersion would result in too little high output. I've been thinking about optimizing segmentation for dispersion and use active eq to straighten things out.
Depends on the construction, solid wire core or not, how much tension etc. Should not be a problem.
Plate stators are difficult to isolate, I don't have any experience with that but heard good things about powder coating in several layers.
v-bro said:
The problem is that the unloaded transformer has a severe resonance peak (caused by Lleak and the transformer's capacitance). Any signal at this freq will be amplified by much more than the stepup ratio. So this peak has to be dampened.
Placing a small load that contains some resistance (to dissipate the energy) will do so. The resistors that are in series with the bass panels do this normally. So replace the panels with a capacitor of approx the same value (leaving the 180k R's in place) and you're fine. Value isn't critical, a few hundred pF. But make sure it is rated at several kV.
Try the search, there was a discussion about it somewhere here recently
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.