Hi!
I started to refurbish QUAD ESL63 panels around 1999 and around 2001 it went more serious, a separate workshop was rented together with a friend.
Hostaphan of different thickness in big rolls was purchased and we developed a "perfect coating"... the coating lasted for approx 10 years, so it was not perfect after all. The Hostaphan film stayed stretched!
A lot of testing was done in a not so scientific way, also heat treatment was done with heat gun. But everything can be done slightly better so thats the background of the information in this tread.
This tread is about the properties of polyester film, BO-PET, and you will see that it is not that simple as you may think.
The film has different properties and tensile strength in X and Y direction.
And this is important if you want to stretch the membrane really hard since it will relax slightly over time....
A good stretched membrane will stay almost straight when there is no signal:
Too much relaxation means that the film will stick to one side and almost touch the stator:
And when the membrane is too relaxed you can get arcing and hiss noise, arcing is producing ozone, and ozone will make the film brittle and it will tear apart and get "rotten". The glue holding the membrane will also decompose or degrade in contact with ozone.
To measure the performance we need to do some testing of the tensile strength and the elongation.
When you apply tensile force to the film it will start to break at the weakest point which usually is where you fixate the film.... Over 20 pieces was tested and 6 of them "worked" by breaking somewhere close to the middle.
Some tape helps out a little bit.
Clamping force is crucial.. and many tests has to be done before you get one that breaks where you want it.
I have a MOV movie but could not download it here of the actual "happening" when it breaks...
So here is some snippings..
Some milliseconds before it breaks..
And finally it gives up.
The test jig is moving 50mm/minute and the force is monitored continuously.
There is a difference between X (width) and Y (length) direction but more sample has to be tested with carefully mounting to get better data.
It is also extremely important how you cut the test pieces. Even a small scratch on the surface introduces a force concentration and it will break long before it´s real maximum.
Here is a table with some more or less successful measurements. Test is performed at 21deg C.
Nr 2, 3, Is in the longest direction, the winding direction of the roll so to say.
Nr 4, 10, 11, and 12 is in the shortest direction, side to side. The spread of data is as you can see to big to draw any conclusions for now.
Wider pieces will be tested to reduce the influence of irregularities from cutting the film.
Here is a graph:
And on 6 samples we can theoretically do a calculation, even though we know that there was some issues when the sample was prepared.
The orange and green curve has a different slope compared to the others, that we can see clearly.
This is a good beginning. Or what do you think?
Next step is to measure the influence of heat, you know about the "heat gun" stretch and the "baking" of the film in an oven as we have seen QUAD is doing.
I do not find it but if you know you can post the reference further down, Thank you!
A little teaser for next experiment before bed time...
With this test jigg you can adjust the chamber temperature (chamber is not attached in this picture, but you get the point.
Same force diagram is plotted as previous, but here you can adjust temperature within +/- 0,2 deg C. Good enough.
After stretching at a certain force, you heat up the chamber/membrane to a certain very specific temperature where the polyester molecules actually starts to tensioning the diagram even harder! I will explain the physics later on.
And if you exceed this temperature point it will relax instead and then if you heat it even more it will break.... game over.
We need to characterize this better than the attempt above, and also better than the stretch jig i developed >20 years ago, and actually is still working very good by the way!
If you want to make a good long lasting electrostatic speaker, maybe you can learn something here.
I started to refurbish QUAD ESL63 panels around 1999 and around 2001 it went more serious, a separate workshop was rented together with a friend.
Hostaphan of different thickness in big rolls was purchased and we developed a "perfect coating"... the coating lasted for approx 10 years, so it was not perfect after all. The Hostaphan film stayed stretched!
A lot of testing was done in a not so scientific way, also heat treatment was done with heat gun. But everything can be done slightly better so thats the background of the information in this tread.
This tread is about the properties of polyester film, BO-PET, and you will see that it is not that simple as you may think.
The film has different properties and tensile strength in X and Y direction.
And this is important if you want to stretch the membrane really hard since it will relax slightly over time....
A good stretched membrane will stay almost straight when there is no signal:
Too much relaxation means that the film will stick to one side and almost touch the stator:
And when the membrane is too relaxed you can get arcing and hiss noise, arcing is producing ozone, and ozone will make the film brittle and it will tear apart and get "rotten". The glue holding the membrane will also decompose or degrade in contact with ozone.
To measure the performance we need to do some testing of the tensile strength and the elongation.
When you apply tensile force to the film it will start to break at the weakest point which usually is where you fixate the film.... Over 20 pieces was tested and 6 of them "worked" by breaking somewhere close to the middle.
Some tape helps out a little bit.
Clamping force is crucial.. and many tests has to be done before you get one that breaks where you want it.
I have a MOV movie but could not download it here of the actual "happening" when it breaks...
So here is some snippings..
Some milliseconds before it breaks..
And finally it gives up.
The test jig is moving 50mm/minute and the force is monitored continuously.
There is a difference between X (width) and Y (length) direction but more sample has to be tested with carefully mounting to get better data.
It is also extremely important how you cut the test pieces. Even a small scratch on the surface introduces a force concentration and it will break long before it´s real maximum.
Here is a table with some more or less successful measurements. Test is performed at 21deg C.
Nr 2, 3, Is in the longest direction, the winding direction of the roll so to say.
Nr 4, 10, 11, and 12 is in the shortest direction, side to side. The spread of data is as you can see to big to draw any conclusions for now.
Wider pieces will be tested to reduce the influence of irregularities from cutting the film.
Here is a graph:
And on 6 samples we can theoretically do a calculation, even though we know that there was some issues when the sample was prepared.
The orange and green curve has a different slope compared to the others, that we can see clearly.
This is a good beginning. Or what do you think?
Next step is to measure the influence of heat, you know about the "heat gun" stretch and the "baking" of the film in an oven as we have seen QUAD is doing.
I do not find it but if you know you can post the reference further down, Thank you!
A little teaser for next experiment before bed time...
With this test jigg you can adjust the chamber temperature (chamber is not attached in this picture, but you get the point.
Same force diagram is plotted as previous, but here you can adjust temperature within +/- 0,2 deg C. Good enough.
After stretching at a certain force, you heat up the chamber/membrane to a certain very specific temperature where the polyester molecules actually starts to tensioning the diagram even harder! I will explain the physics later on.
And if you exceed this temperature point it will relax instead and then if you heat it even more it will break.... game over.
We need to characterize this better than the attempt above, and also better than the stretch jig i developed >20 years ago, and actually is still working very good by the way!
If you want to make a good long lasting electrostatic speaker, maybe you can learn something here.
Attachments
Last edited:
Just a quick note on your first figure: Unless your tension is infinite, the diaphragm will always deflect from the center position when charged up. The amount of deflection limits the effective xmax of the panel since it's now biased to one side. The tension has to be high enough that the diaphragm does not collapse into a stator and you don't eat up all your available xmax. But in a real world ESL there will always be deflection from the non-charged position.
Sheldon
Sheldon
Actually, I have a couple more comments:
The tensile strength of bopet isn't uniform, it's stretched as it hardens and the machine direction has about 80% of the strength of the transverse direction. So let's say you were building a curved martin logan type panel where you want to primarily stretch in one direction, it would be useful to orient the transverse direction of the mylar to support that.
The stress strain curve is very strongly temperature dependent. That could be effecting your pull test results.
Sheldon
The tensile strength of bopet isn't uniform, it's stretched as it hardens and the machine direction has about 80% of the strength of the transverse direction. So let's say you were building a curved martin logan type panel where you want to primarily stretch in one direction, it would be useful to orient the transverse direction of the mylar to support that.
The stress strain curve is very strongly temperature dependent. That could be effecting your pull test results.
Sheldon
BOPP is a thermally quenched film, so is BO-PET.
It is quenched in water after extrusion to retain its properties, IIRC.
BO means Bi axially Oriented.
Process parameters at extrusion will affect the properties, and the settings for production may have been for adhesive tape, or plastic bags, both are common end uses for this type of film, and not intended for speakers, which need to move in X and Y directions, and not stay stretched.
First check it is not BO film, then proceed with your repair project.
Magnepan use special equipment for factory rebuilds of their planar speakers. The film is attached to the speakers using special jigs, and part of that process is a trade secret.
It is quenched in water after extrusion to retain its properties, IIRC.
BO means Bi axially Oriented.
Process parameters at extrusion will affect the properties, and the settings for production may have been for adhesive tape, or plastic bags, both are common end uses for this type of film, and not intended for speakers, which need to move in X and Y directions, and not stay stretched.
First check it is not BO film, then proceed with your repair project.
Magnepan use special equipment for factory rebuilds of their planar speakers. The film is attached to the speakers using special jigs, and part of that process is a trade secret.
Sure you are right, the film will always move to one side for sure, I will write that more clearly, thanks for the feedback!
Check with the supplier, I think it is oriented, so not very useful for you.
Oriented film has better properties in tension, so in the coating / printing process, it is easier to process.
Oriented film has better properties in tension, so in the coating / printing process, it is easier to process.
This film has been in use for many years and many membranes with great success.
Stretching procedure is important, that's why a lot of effort was put in to the development of my stretch jig that i developed >20 years ago.
Bi axial Orientation was taken in consideration, much by trial and error, and it has worked fine since then!
I will try to explain the physics behind Bi-axial Orientation, and also how the PET molecules actually is stretching the film even after mechanical stretching, not so intuitive maybe.
Last edited:
There can be quite some difference in specifications between different kinds/brands of PET film.
Mitsubishi Hostaphan, Dupont Teijin Melinex, Dupont Teijin Mylar, nameless Chinese production.
Looking at the mechanical (and electrical) specs there can be significant differences.
And then there are some that claim that different brands of film have different sound.
Mitsubishi Hostaphan, Dupont Teijin Melinex, Dupont Teijin Mylar, nameless Chinese production.
Looking at the mechanical (and electrical) specs there can be significant differences.
And then there are some that claim that different brands of film have different sound.
I am eager to learn and am still waiting for the promised explanations ... and more precisions.
From infos on the net, I concluded that there are three important temperature ranges for Mylar (PET): Glass transtion temperature, Crystallization and Melting. These "decimally precisely measured temperature spots" however can vary strongly from paper to paper and from graph to graph. Take three nice examples:
So, what temperature points are you talking about? And how exactly do you heat treat your mylar today after 4 km of experience? At which temperature do you heat up your mylar? And for how long? And do you let it gently cool down afterwards, or do you chill it down? Chill it down, because otherwise the molecules would have too much time to rearrange as if there had be no heat treatment?
And under which tension do you heat your mylar up? Does this tension then relax stlightly under this given heat treatment, so you would have to re-tense it steadily unter temp? And then, in the end, what's the actual difference between a heat treated diaphragm and a non-heat-treated one?
IMO, questions like this would be interesting to be answered and more deeply discussed.
There is much info about physics of mylar and temperature influences on it out there, but not so about tensioned mylar under temperature influence. But I cannot make a real link between these more general informations and the merit of heat treatment of a mylar diaphragm.
https://www.linseis.com/en/wiki-en/glass-transition-determination/
https://www.linseis.com/wp-content/uploads/2018/10/Application-DSC-Glass-point.png
https://www.linseis.com/en/applications/polymers/chip-dsc-10-polymers-pet/
https://omnexus.specialchem.com/polymer-properties/properties/glass-transition-temperature
Most of the information you seek is available from material suppliers, and some, like the jig stretch, are trade secrets.
Do not expect a complete answer.
Do not expect a complete answer.
I am not eager to know about trade secrets or general material information from suppliers.
I would be very happy instead to get some (and in the fact of it's nature incomplete) information specifically from ESL63 about his practical approach to stretch and eventually to bake mylar, as advertised in his thread's title. Infos most welcomed and expected, infos about his practical experiences. Hoping, such kind of infos might not be trade secrets of his. He already showed e.g. some beautiful pictures of advanced lab gear I think nobody else here has handy. And furthermore, after 4km use of mylar, I think he must have a lot of knowledge and experience how to best handle this material he seems willing to share with us. Specific experience out of his "everyday's trial an error" and potentially useful for ESL DIYing , experience a material supplier certainly does not have. As you seem to know ...
I would be very happy instead to get some (and in the fact of it's nature incomplete) information specifically from ESL63 about his practical approach to stretch and eventually to bake mylar, as advertised in his thread's title. Infos most welcomed and expected, infos about his practical experiences. Hoping, such kind of infos might not be trade secrets of his. He already showed e.g. some beautiful pictures of advanced lab gear I think nobody else here has handy. And furthermore, after 4km use of mylar, I think he must have a lot of knowledge and experience how to best handle this material he seems willing to share with us. Specific experience out of his "everyday's trial an error" and potentially useful for ESL DIYing , experience a material supplier certainly does not have. As you seem to know ...
Are there any measurements that show the "advantages" of Mylar heat treatment?
How much does pre-tensilized Mylar-C really creep over what period of time?
If you can't really control the curing process, my guess is that the remedy could be worse than the initial problem (if there is any at all).
Just overstretch a little, so when it settles you have the right tension.
Where is the oven on the ESL-63 stretching jig in the picture? I've only seen an oven used on very early ESL-57 stretcher (pre-tensilized period?).
How much does pre-tensilized Mylar-C really creep over what period of time?
If you can't really control the curing process, my guess is that the remedy could be worse than the initial problem (if there is any at all).
Just overstretch a little, so when it settles you have the right tension.
Where is the oven on the ESL-63 stretching jig in the picture? I've only seen an oven used on very early ESL-57 stretcher (pre-tensilized period?).
Into my brain. And into everybody else's interested in the subject about advanced knowledge about mylar tensioning under heat influence. But you indeed rise an interesting question: What is the meaning, the sense of such kind of obscure info's in general? It's quite a philisophical subject. Maybe best discussed in the lounge.
Shure. That's why I am here.
No problem. But you forgot to enumerate some more items like the coating, special glues and the audio stepup transformer(s). Easy to get. I purchased all this stuff and some other ESL-appropriate DIY-goodies at eraudio.com.au. They have a very decent and high-quality choice and provide a super customer service. In case you want to dive into the subject. Instead, there is no need for several jigs, not even a single one. A plane table and some gaffa tape for stretching the mylar will do. Unless you will heat treat the mylar, and this is the question which might be enlighted here. As for the high voltage gear, you can easily DIY the HT generator (or buy it at eraudio), and you have to spend some $ for a HT probe for your multimeter if you absolutely want to do so. That's it. You see, you easily could join the club if you want to ... and be prepared to have a lot of challenging fun doing so. And I guess you would not end up with a single speaker. It's kind of addicting, this DIY ESL stuff.
Finally: It sounds exciting to me now having the perspective to know something more about not only the make of ESL's, but also about the bake of ESL's. This is why I am insisting a bit.
People who have been there witness about same owen even for 63. Not a surprise which I will show in a couple of weeks.
This information needs some clarification.
What is measured here is how much energy is required from an external heat source to increase the temperature x degrees.
So we can call it "specific temperature" property... If you heat ice it will require a lot of energy to get to melted state, and then when you reach 100 deg it requires a lot of energy to go from liquid state to steam. The DSC graphs shows the quite complex nonlinear property of polymer, and it is informative for sure. (I also updated the first post, which I do now and then)
Last edited: