Hello everybody,
I am planning to make a wire stator full range ESL.
Planned width 198mm, height 1600mm, distance between stator and film 2mm.
I'm also thinking about segmentation using "esl_seg_ui".
I made a full-range ESL using Perforated metal in the past.
I used a 2x50VA 6V/230V power supply toroidal transformer and it didn't produce much bass.
I think the culprit is the step-up transformer.
What are the requirements for a transformer that can be used for a full range?
Required core cross-sectional area. What is the number of turns on the primary side? , what is the required turns ratio? , the gauge of the primary and secondary coils?
It was written in detail in the following URL, but there was something I didn't understand.
https://www.diyaudio.com/community/threads/diy-bass-transformer-for-esls.186011/
What value is "E" in the following formula?
The following transformer specifications seem to be suitable for full range.
The cross-sectional area of the core varies in size between inches and centimeters. Is it my reading problem?
which one is correct?
thank you.
(I am using Google Translate)
I am planning to make a wire stator full range ESL.
Planned width 198mm, height 1600mm, distance between stator and film 2mm.
I'm also thinking about segmentation using "esl_seg_ui".
I made a full-range ESL using Perforated metal in the past.
I used a 2x50VA 6V/230V power supply toroidal transformer and it didn't produce much bass.
I think the culprit is the step-up transformer.
What are the requirements for a transformer that can be used for a full range?
Required core cross-sectional area. What is the number of turns on the primary side? , what is the required turns ratio? , the gauge of the primary and secondary coils?
It was written in detail in the following URL, but there was something I didn't understand.
https://www.diyaudio.com/community/threads/diy-bass-transformer-for-esls.186011/
What value is "E" in the following formula?
The following transformer specifications seem to be suitable for full range.
The cross-sectional area of the core varies in size between inches and centimeters. Is it my reading problem?
which one is correct?
thank you.
(I am using Google Translate)
Hello wout31
In the ESL I made in the past.
Using 6 micrometer polyester film. elongation by 2%.
I wiped the surface with isopropanol.
I applied Licron Crystal.
In the ESL I made in the past.
Using 6 micrometer polyester film. elongation by 2%.
I wiped the surface with isopropanol.
I applied Licron Crystal.
E is RMS value of the voltage applied across the winding of N turns. Using Tesla instead of Gauss change 10^8 to 10^4. This comes from cross section in sq centimeters instead of meters (1sq cm = 10^-4 of a sq m). And you have 10^4 Gauss in 1 Tesla.
hello alexberg,
"E" is the effective value of the voltage.
Due to the difference in the unit system, it may be replaced by 10^4 or 10^8.
I was able to gain knowledge. Thank you very much.
"E" is the effective value of the voltage.
Due to the difference in the unit system, it may be replaced by 10^4 or 10^8.
I was able to gain knowledge. Thank you very much.
Keep in mind that Soundlab uses two transformers to make their full range signal. Your description is for the bass transformer only.
A single large EI transformer with a 1:200 step-up is likely to have limited high frequency response because of high leakage inductance.
Soundlab schematic here showing dual transformers:
https://www.diyaudio.com/community/threads/acoustat-mk-series.254723/post-3911097
And some more description in this post of the thread you linked to:
https://www.diyaudio.com/community/threads/diy-bass-transformer-for-esls.186011/post-2522064
"Their interface unit uses a nearly identical setup to Acoustat where they have a large bass transformer with high turns count and high leakage inductance mixed with a toroidal treble transformer with low turns count and low leakage inductance. The toridal transformer is only driven above about 250Hz."
Last edited:
hello mattstat,
I see, Soundlab and Acoustat were dual transformers!
Thank you for introducing the link. informative.
I wanted to try the full range without using the network.
Poor high frequency response. I wonder if I should reconsider.
Can leakage inductance be avoided by lowering the turns ratio, for example to 1:120?
What would an ideal step-up transformer for a full range sound be like?
Is this just a compromise? Is high frequency response degradation unavoidable in full range transformers?
I see, Soundlab and Acoustat were dual transformers!
Thank you for introducing the link. informative.
I wanted to try the full range without using the network.
Poor high frequency response. I wonder if I should reconsider.
Can leakage inductance be avoided by lowering the turns ratio, for example to 1:120?
What would an ideal step-up transformer for a full range sound be like?
Is this just a compromise? Is high frequency response degradation unavoidable in full range transformers?
You confirmed my doubt.
I have tested many many coatings in past years (only for Quad ESL) and never found Licron Crystal to be a good candidate.
I also know that many/some people restoring Audiostatic ESL's use Licron and say it works fine (no proof of compare with others).
So maybe there is a difference in what design it is used, bias voltage applied, way of applying bias voltage etc.
If you intend to make your own transformers, the Radio Data Handbook 4 is still a very good source of information. Google RDH4.
In general it is benificial to use large cores and to split the transformer in two seperate units, as is done by Quad on the 63 and later models and by Audiostatic.
There will always be compromises. It is the combination of max output voltage and octaves of bandwidth that is limiting things.
In general it is benificial to use large cores and to split the transformer in two seperate units, as is done by Quad on the 63 and later models and by Audiostatic.
There will always be compromises. It is the combination of max output voltage and octaves of bandwidth that is limiting things.
Leakage inductance can't be avoided, but it can be managed.
From what I've seen, a turns ratio of 1:150 is about the upper limit for a single transformer of reasonably wide bandwidth with good low frequency performance. This will typically still have slightly limited high frequency performance, so a segmented speaker is usually required/the load capacitance has to be reasonable. The Tranex 2-0404 was a pretty good example of this. I may have a datasheet on that one at home (I'm at work at the moment).
And as the others have said, multiple identical smaller ratio transformers can be ganged together to improve high frequency performance verses a single transformer of the same composite turns ratio, as the Quad ESL-63 and others have done. This avoids the frequency dividing network between transformers, but still gives you some of the benefits of using more than one transformer.
The tricky part is getting good low and high frequency performance from the same transformer with a high turns ratio. The requirements tend to oppose each other. So it's always a balance.
From what I've seen, a turns ratio of 1:150 is about the upper limit for a single transformer of reasonably wide bandwidth with good low frequency performance. This will typically still have slightly limited high frequency performance, so a segmented speaker is usually required/the load capacitance has to be reasonable. The Tranex 2-0404 was a pretty good example of this. I may have a datasheet on that one at home (I'm at work at the moment).
And as the others have said, multiple identical smaller ratio transformers can be ganged together to improve high frequency performance verses a single transformer of the same composite turns ratio, as the Quad ESL-63 and others have done. This avoids the frequency dividing network between transformers, but still gives you some of the benefits of using more than one transformer.
The tricky part is getting good low and high frequency performance from the same transformer with a high turns ratio. The requirements tend to oppose each other. So it's always a balance.
Last edited:
hello maudio,
Even old books are useful.
The basics of the transformer may not have changed.
Certainly, it is better to use a step-up transformer for each of the front and rear stators.
It may be good to aim for a direction like Audiostatic.
Even old books are useful.
The basics of the transformer may not have changed.
Certainly, it is better to use a step-up transformer for each of the front and rear stators.
It may be good to aim for a direction like Audiostatic.
hello mattstat,
Thank you for your message while you were at work.
The turns ratio of a full range step-up transformer should be up to 1:150.
For good frequency characteristics, it is better to separate segments and transformers by frequency band.
But this time I will try the full range.
Everyone is so knowledgeable and detailed.
Thank you for your message while you were at work.
The turns ratio of a full range step-up transformer should be up to 1:150.
For good frequency characteristics, it is better to separate segments and transformers by frequency band.
But this time I will try the full range.
Everyone is so knowledgeable and detailed.
I thought this sheet had more specifications on it, but here it is anyway. I had a note that the leakage inductance is 5.9 microhenries. I didn't measure it myself though, so I don't know if that's accurate.
These were made years ago, and Tranex was bought by a larger transformer company, so as far as I know they are no longer available. It's just an example of what a single high-ratio, wide bandwidth transformer is like.
If by saying you will "try the full range" you mean you don't intend to segment the panel, you may have trouble getting sufficient high frequencies with a single high-ratio transformer. That can be calculated if you know the leakage inductance. Also, with a wire panel, it's a lot easier if you wind everything with the intention to segment. If you decide you want to try unsegmented, putting all those wires on one terminal is a lot easier than trying to go the opposite way and having to cut wires after the fact and splice into them. I have done that, and it was a multi-hour job.
These were made years ago, and Tranex was bought by a larger transformer company, so as far as I know they are no longer available. It's just an example of what a single high-ratio, wide bandwidth transformer is like.
If by saying you will "try the full range" you mean you don't intend to segment the panel, you may have trouble getting sufficient high frequencies with a single high-ratio transformer. That can be calculated if you know the leakage inductance. Also, with a wire panel, it's a lot easier if you wind everything with the intention to segment. If you decide you want to try unsegmented, putting all those wires on one terminal is a lot easier than trying to go the opposite way and having to cut wires after the fact and splice into them. I have done that, and it was a multi-hour job.
hello mattstat,
Thank you for showing us your valuable materials.
Using a wire panel, first make a full range.
Depending on the result, I will devise and make it so that it can be segmented.
Now, let's consider what kind of specifications we will use to order the optimal step-up transformer for a full range.
Buying a commercially available product may be more efficient and cheaper, but there are few options.
Thank you for showing us your valuable materials.
Using a wire panel, first make a full range.
Depending on the result, I will devise and make it so that it can be segmented.
Now, let's consider what kind of specifications we will use to order the optimal step-up transformer for a full range.
Buying a commercially available product may be more efficient and cheaper, but there are few options.
Not using segmentation on a wire esl is imho a very bad idea. A non-segmented panel will beam like crazy, resulting in what basically is an unusable speaker. There is no good reason I can think of why not to use segmentation. It will also lighten the load on the transformer considerably. For a start, separate a high frequency strip of a cm or 2-3 from the rest.
About transformer specs: aim for 8kVpp out @ 50 hz for fullrange operation. Turn ratio 1:100-120, not higher. Limit the ds spacing of your panel accordingly. 1.5 to 2 mm max.
Choose the core as large as possible, limiting factor being the primairy inductance at the lowest frequency you aim for.
Determine core saturation point of the core using a test setup, that will give you an idea of Bmax. This will give you the # of turns on both prim and sec.
For the coils, in general it is benificial to maximize area. Large C-cores are good for this as you can divide the secundairy in 2 parts. Less layers and interwinding is good for capacitance but bad for leakage, you have to experiment to optimize. Same for the thickness of insulation layers in between.
It will take a couple of trials to get it right. Prepare for a lot of winding hours...
As for commercial options, 6/230V toroids are worth looking into. Use 2 or 4 per channel. There are some good threads about this to be found here somewhere.
About transformer specs: aim for 8kVpp out @ 50 hz for fullrange operation. Turn ratio 1:100-120, not higher. Limit the ds spacing of your panel accordingly. 1.5 to 2 mm max.
Choose the core as large as possible, limiting factor being the primairy inductance at the lowest frequency you aim for.
Determine core saturation point of the core using a test setup, that will give you an idea of Bmax. This will give you the # of turns on both prim and sec.
For the coils, in general it is benificial to maximize area. Large C-cores are good for this as you can divide the secundairy in 2 parts. Less layers and interwinding is good for capacitance but bad for leakage, you have to experiment to optimize. Same for the thickness of insulation layers in between.
It will take a couple of trials to get it right. Prepare for a lot of winding hours...
As for commercial options, 6/230V toroids are worth looking into. Use 2 or 4 per channel. There are some good threads about this to be found here somewhere.
Last edited:
Some toroids will do quite well in fullrange application as well. But not all are equal, the only way to find out is buy one and measure. Aim for 50VA types.
Tube output transformers are another option, but stepup is low, 1:60 range. Not ideal for FR.
Tube output transformers are another option, but stepup is low, 1:60 range. Not ideal for FR.
There is a big difference between the transformers for segmented and unsegmented ESLs.
Unsegmented ESLs have a large capacitance, so you must use transformers with a low leakage inductance to prevent the hf attenuation. Larger toroidal transformers, at least 50 VA, or purpose built EI transformers with interleaved windings are required (see amplimo/plitron transformers for examples). For a large full range ESL, you'll have trouble getting the step-up ratio above about 60, and the larger the ESL the greater the problems. As pointed out already, large unsegmented ESLs beam - the sweet spot may be less than 1 degree wide = less than the distance between your ears at 3 m - be prepared to put your head in a vise. Avoid long speaker leads - unsegmented ESLs may be the only speaker where the leads matter.
The RC transmission line segmented ESLs present a much friendlier load to transformers - a mix of resistive and capacitive load with the effective capacitance about 20X smaller for the same sized panel. Small toroids work quite well, 15 VA toroids with the 230V windings in series and low V windings in series and/or parallel to get the input rating you require (~ 25V at 50 Hz) work well. Commercial transformers are normally made to withstand 4kV between primary and secondary, so should withstand use in ESLs. Be sure to avoid transformers with the dual 115V/230V windings - these have a very high capacitance and are no good. With small toroids, you can get step up ratios of 120 easily, and with large purpose-built EI transformers, step up ratios approaching 200 are possible. A pair of the ER Audio (each 1:90, the pair giving 1:180) transformers in a push-pull arrangement work well. With segmented ESLs you get a flat frequency response over about 10-15 degrees of angle - actually a better polar response than many conventional speakers.
In general, the larger the ESL panel and the lower the frequency response you require, the bigger the transformer required and the lower the step up ratio that can be achieved. Also, whatever the ESL design, you need to keep them away from rear walls (> 1m) to avoid low frequency cancellation from the sound reflected from the wall.
Unsegmented ESLs have a large capacitance, so you must use transformers with a low leakage inductance to prevent the hf attenuation. Larger toroidal transformers, at least 50 VA, or purpose built EI transformers with interleaved windings are required (see amplimo/plitron transformers for examples). For a large full range ESL, you'll have trouble getting the step-up ratio above about 60, and the larger the ESL the greater the problems. As pointed out already, large unsegmented ESLs beam - the sweet spot may be less than 1 degree wide = less than the distance between your ears at 3 m - be prepared to put your head in a vise. Avoid long speaker leads - unsegmented ESLs may be the only speaker where the leads matter.
The RC transmission line segmented ESLs present a much friendlier load to transformers - a mix of resistive and capacitive load with the effective capacitance about 20X smaller for the same sized panel. Small toroids work quite well, 15 VA toroids with the 230V windings in series and low V windings in series and/or parallel to get the input rating you require (~ 25V at 50 Hz) work well. Commercial transformers are normally made to withstand 4kV between primary and secondary, so should withstand use in ESLs. Be sure to avoid transformers with the dual 115V/230V windings - these have a very high capacitance and are no good. With small toroids, you can get step up ratios of 120 easily, and with large purpose-built EI transformers, step up ratios approaching 200 are possible. A pair of the ER Audio (each 1:90, the pair giving 1:180) transformers in a push-pull arrangement work well. With segmented ESLs you get a flat frequency response over about 10-15 degrees of angle - actually a better polar response than many conventional speakers.
In general, the larger the ESL panel and the lower the frequency response you require, the bigger the transformer required and the lower the step up ratio that can be achieved. Also, whatever the ESL design, you need to keep them away from rear walls (> 1m) to avoid low frequency cancellation from the sound reflected from the wall.
こんにちは、maudio、mattstat、golfnut、
トランスフォーマーの具体的な目標を教えていただきありがとうございます。
実現できるかどうかはわかりませんが、トランスのメーカーさんにできるだけ大きなコアと巻線比を相談してみます。50Hzで8kVpp出力。高価で注文が難しい場合があります。
私のアイデアは、AudioStatic-DCI のようなスピーカー システムを最初に作成し、結果に応じて改善することでした。
あなたが言ったことから、セグメンテーションは必須のように思えます。
参考までに良い例を教えてください。
念のため、セグメンテーションについて教えてください。
1: esl_seg_ui.exe を使用して生成された抵抗を使用して ESL ビームを回避する方法ではないでしょうか?
2：ESLパネルとトランスをバンドごとに使い、ディバイディングネットワークで分離したスピーカーシステムのことですか？ESL-63のように。
トランスフォーマーの具体的な目標を教えていただきありがとうございます。
実現できるかどうかはわかりませんが、トランスのメーカーさんにできるだけ大きなコアと巻線比を相談してみます。50Hzで8kVpp出力。高価で注文が難しい場合があります。
私のアイデアは、AudioStatic-DCI のようなスピーカー システムを最初に作成し、結果に応じて改善することでした。
あなたが言ったことから、セグメンテーションは必須のように思えます。
参考までに良い例を教えてください。
念のため、セグメンテーションについて教えてください。
1: esl_seg_ui.exe を使用して生成された抵抗を使用して ESL ビームを回避する方法ではないでしょうか?
2：ESLパネルとトランスをバンドごとに使い、ディバイディングネットワークで分離したスピーカーシステムのことですか？ESL-63のように。