Jig and stretch for curved ESL?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Let's have a tread dedicated to the jig needed to build curved ESL's?

Me being a n0ob have some thoughts and ideas but I'm sure there are other people out there interested in learning how to actually build a curved panel?

So, please!
If you have any experience in how to practically build curved esl, please share.
How does the jig look like, how do you tension the diaphragm and how do you mount the front and back stator?
 
Mark I will take some photos and post. I followed Sanders "plans" for the form to attach the spacers and impart the curve to the stators. However, for the stretcher part, I did not follow his method, but rather built a large stretcher frame with ends that had the same radius as my panels and used veneer screws to move one of the ends to tension the diaphram. It is important to only stretch lengthwise, and just provide enough tension on the sides to remove wrinkles.
 
I have attached three photos (or at least tried) showing my stretcher, detail around the use of the veneer screws, and stretcher with the "table top" as Sanders calls it. The stretcher is about 50% larger than the stator panels. Curved diaphragms, which are only stretched lengthwise, develop and "hour glass" shape. To minimize this effect, only the center portion of the diaphragm is used. In my case, I used perforated steel that was 40" long, but from end to end my stretcher is nearly 72". I use carpet tape to attach the membrane to the ends of the stretcher, then using the screws in "reverse direction" stretch the membrane over the "table top" about 1.5%. Using very slight pressure, I pull the membrane to the sides with tape to remove any large wrinkles. Once the membrane looks smooth, I lay the outside stator on top of the membrane (which has double sided tape over the spacers) to attach the stator to the diaphragm. Hope this helps, jerry
 

Attachments

  • Stretcher1.jpg
    Stretcher1.jpg
    57.2 KB · Views: 637
  • Strecther Detail.jpg
    Strecther Detail.jpg
    83.2 KB · Views: 599
  • Stretcher and Frame.jpg
    Stretcher and Frame.jpg
    55.4 KB · Views: 601
I like your stretcher Jerry. I've only built flat panels but I've been contemplating building a curved panel this year and considering how I might build the stretcher. Your stretcher is simpler than what I had imagined and that's the way to go, I figure. I had anticipated the diaphragm would need to be quite a bit wider than final panel size because of the deformation that you mentioned. However, it had not occurred to me that the film would also need to be quite a bit longer as well. Thanks for that tip, Jerry!

Are you using 6 micron film?

BTW, I would love to see your speakers!
 
Thanks everyone for your very kind comments. Yes I am using 6 micron mylar film. I forgot to mention that my stators are 12" wide, but the stretcher will hold 19" wide film, again to minimize impact of the hour glass effect. I am still constructing the speakers, so do not have any final photos. I have one panel completed (and will use your suggestions for mounting Charlie). At present, I am working mostly on the cabinets for the woofers - it will be a hybrid system. I live in the US-Midwest where temperatures are freezing - so I will either need to wait until it warms, or bring the stretcher inside. Following suggestions of others here, the tape adheres best under warm conditions. I will post pictures as I complete them. I have been a member of this forum for nearly 5 years, and have really appreciated the suggestions and ideas that members have shared. Hopefully I am getting to a point where I can give something in return. Jerry
 
I've been busy with drawing up pcb's (the price I pay for trying to multitask) so my mind's been elsewhere but I do follow the thread and I'm still hoping for more people to post their solutions. :)

How do you get the right curvature on the stators? I assume you roll them before mounting?
 
I will share what I did, and hope that others will do as well. I started with 40” x 36” perforated steel sheets, which I cut into 40”x 12” pieces for the stators. This evidently relieved some stress within the sheet, which resulted in a slight curve, but not enough (my design was 20 degrees since Sanders indicated this was a “safe” amount). To get the final correct curve, I either taped the stator to the “20 degree table top” for the inner stator, and then glued the acrylic horizontal spacers to the stator to “fix” the final curve to the spacer, or for the outer stators, taped the acrylic spacers to the "table top", and then attached the stator to the spacers (in both cases using epoxy to glue the spacers to the stators). This puts a compression force on the spacers and effectively sets the curvature of the stators (within a degree or two). I pretty much followed Sanders on this and it worked reasonably well. For those thinking of making curved panels, I personally found that there was more work involved in making the “table tops” and stretcher as there was in construction of the stators. Hope this helps, Jerry
 
Klaatu Audio Elektrostat

I built these speakers to a friend back in -92 who loved the Martin Logan sound, but had no money. I experimented a lot and came up with a design where a special stretching jig wasn't needed. I used flat .8 mm thick steel stators which I had powder coated and then urethane painted. I applied one mm thick double sided foam tape around the stator perimeter and strips of the same material as membrane supports. At the center of the membrane supports, the foam was replaced by 10 mm long plastic pieces for avoiding mechanical compression of the foam strips. The film was mounted with the help of a ordinary stretch frame. After contact stripping, the front stator was laid loosely on top
and the esl package was sunk in a wooden frame as seen on picture and a one meter long 6x50 mm shrink tubed steel bar was mounted from behind, pressing the esl package forward about, if I remember correctly, 30 millimeters or 1,2 inch.

The photo is taken a few months ago when my son discovered my speakers
in an local HiFi equipment shop outside of Malmö.

Remarkably. they still worked as a charm....
 
Let's have a tread dedicated to the jig needed to build curved ESL's?

(Sorry for the longish post)
Me being a n0ob have some thoughts and ideas but I'm sure there are other people out there interested in learning how to actually build a curved panel?

So, please!
If you have any experience in how to practically build curved esl, please share.
How does the jig look like, how do you tension the diaphragm and how do you mount the front and back stator?



Ok, I had an idea. I first started thinking that a curved panel was the way to go so I curved them. The curved panels should improve the head in the vice syndrome. However, there is no free lunch, since when we build them flat the laser directivity in the listening spot would have better imaging then that produced from some curved panel.

So its been about a year since I curved the steel. I have been stuck with the same problem. I was about to flatten them because when you see how easy it is to strected the mylar when there flat using a bike tube and pump you get jealous on how easy it is for them.

What I will do is the following,
1. Do the same thing to strech the mylar as you would if they were flat, using a bike tube as above

2. place marks sperated about 6" on the mylar all over it in many directions, vertical and horizontal in several places.

3. Calculated the elongation, you should see when you acheive the desired tension (youngs' modulus).

4. Measure, verify, coat, measure verify

5. Use 3m double side tape, follow procedure ad naueseum stated around this awesome forum.

6. press the curved panels down on the tensioned, coated mylar.

7. build into the design that you will fail, and have to replace the mylar eventually. make that easy to do down the road.

All the tensional forces are in the horizontal direction, so it would not take much force to vertically press the curved panels down (with some press jig) on the flat mylar. This means the Mylar will have to conform to the new shape of the curved stators. This seems easy enough. If this should work I will let you know. After the matting, the pressing the tension on the mylar should not change as it was before the pressing.

Double sided tape is clean, fast, strong, and exact. Another advantage of using the tape is you can allow for some overhang around the stator edges, this should help prevent breakdown or discharging between the edges of the plates. The tape is not cheap thought, but IMO worth it.

Also, must people coat one side of the mylar, any reason. I worked out the physics and if we coat one side only, the force seems to be nonlinear in flat panels. If you play this freq, you will hear the higher harmonics. NOT GOOD. Also, you seem to get more force per unit area on the mylar if we coat both sides. Considering to coat one side only, will create an induced surface charge on the other uncoated side. Actually, I think they are induced dipole moments that alternate as fast as the freqs, in the signal oscillate.

I have to check my derivation, and think if there is something Im missing. Do we coat one side for any reason to protect from discharging? Anyone know?
 
Also, must people coat one side of the mylar, any reason. I worked out the physics and if we coat one side only, the force seems to be nonlinear in flat panels. If you play this freq, you will hear the higher harmonics. NOT GOOD. Also, you seem to get more force per unit area on the mylar if we coat both sides. Considering to coat one side only, will create an induced surface charge on the other uncoated side. Actually, I think they are induced dipole moments that alternate as fast as the freqs, in the signal oscillate.

I have to check my derivation, and think if there is something Im missing. Do we coat one side for any reason to protect from discharging? Anyone know?

The main advantage to coating just one side is to minimize the mass added to the diaphragm.


Concerning distortion improvement from coating both sides of the diaphragm:
There was a technical article in the February 1956 issue of Wireless World on this topic.
They author claimed that distortion free operation for large diaphragm excursions could only be obtained if the two sides of the diaphragm were charged and insulated from each other by feeding the HV thru independent high resistances.

In the Letters to the Editor in the March 1956 issue, Peter Walker explained that if the capacitance through the diaphragm, from one coating to the other, is large compared to the capacitance between the coatings and adjacent stators then the operating conditions approached those of a single coating on one side of the diaphragm fed through a high resistance. With a thin mylar diaphragm, this is the case.

A followup letter provides the equations showing why separation of charges on either side of the diaphragm has no advantage over a single conductive side with the same net charge.


The key to low distortion with large diaphragm excursions is a high resistivity coating 1e9 ohm/sq or greater to ensure constant charge operation.
 
The main advantage to coating just one side is to minimize the mass added to the diaphragm.


Concerning distortion improvement from coating both sides of the diaphragm:
There was a technical article in the February 1956 issue of Wireless World on this topic.
They author claimed that distortion free operation for large diaphragm excursions could only be obtained if the two sides of the diaphragm were charged and insulated from each other by feeding the HV thru independent high resistances.

In the Letters to the Editor in the March 1956 issue, Peter Walker explained that if the capacitance through the diaphragm, from one coating to the other, is large compared to the capacitance between the coatings and adjacent stators then the operating conditions approached those of a single coating on one side of the diaphragm fed through a high resistance. With a thin mylar diaphragm, this is the case.

A followup letter provides the equations showing why separation of charges on either side of the diaphragm has no advantage over a single conductive side with the same net charge.


The key to low distortion with large diaphragm excursions is a high resistivity coating 1e9 ohm/sq or greater to ensure constant charge operation.

Thank you for your thoughtful response, with great references. As I am building a mid to high panel im not too worry about large excursion distortion. Ok, yes I remember the name Peter Walker from somewhere, he is pretty big, I have to recall where I heard of him.

As far as the mass added to the mylar from coating both sides would of course add more mass to the mylar. Now, relative to the mass of the mylar used do you think that the typical coatings we use that the mass density is larger than that of the Mylar? We need low mass for faster transients. agreed.

If the work of Walker is right and the operation conditions approach that of a single coating, then sure i would agree, just coat one side.

We normally use the mega ohm resistance to prevent charge from moving around on the mylar surface. the electrostatic forces as the membrane moves will cause charge to move, this is bad since it leads to nonlinear non uniform forcing on the mylar.

Thank You again, will most likely coat the one side.
 
right Peter Walker=Quad, in my recent research in looking at Elvamide for the coating of the mylar, they said that they coat both sides of the mylar in the original quads for the mids and highs. Take a look at the link below, not to far from the top of the page list detailed instructions for using elvamide and there you will see something about how they coat the sides of the mylar.

Do It Yourself - Electrostatic Speakers - Material: - The Audio Circuit
 
I did not follow his method, but rather built a large stretcher frame with ends that had the same radius as my panels and used veneer screws to move one of the ends to tension the diaphram. It is important to only stretch lengthwise, and just provide enough tension on the sides to remove wrinkles.

cool, see my post #11 let me know what you think if you have the time. thanks Bryan
 
As far as the mass added to the mylar from coating both sides would of course add more mass to the mylar. Now, relative to the mass of the mylar used do you think that the typical coatings we use that the mass density is larger than that of the Mylar? We need low mass for faster transients. agreed.
The issue with added mass is not one of transient response so much as it is one of limiting the upper bandwidth. For typical 6μm (1/4 mil) mylar film, the weight of the film itself just starts to roll off the response at 20kHz. Adding any mass in the form of coatings and the roll off starts at a lower frequency. In many designs, this roll off is compensated for by peaking in the electrical response by resonance between the transformer leakage inductance and the panel capacitance. The amount of peaking is controlled by a damping resistor which is usually put in series with the transformer primary where it can do double duty and provide transformer core saturation protection as well. Lower series resistance results in boosting high frequencies, higher resistance rolls off the high frequencies. Acoustat describes this method of compensating the high frequency response of their ESL in US patent 4323736 (column 6 starting on line 52 thru column 8)
Step-up circuit for driving full ... - Google Patent Search

As I am building a mid to high panel im not too worry about large excursion distortion... We normally use the mega ohm resistance to prevent charge from moving around on the mylar surface. the electrostatic forces as the membrane moves will cause charge to move, this is bad since it leads to nonlinear non uniform forcing on the mylar.

You are right, for hybrid ESL panels mega(1e6) ohm/square surface resistance will keep the charge from moving much at the frequencies involved and the distortion will be reasonably low. But, if you ever decide to build a full range ESL, you will need a coating that is at least a thousands times less conductive than this(1e9 ohm/square) to keep distortion from charge migration under control for large diaphragm excursions.

If interested in coating resistance effect on distortion at mid and high frequencies you might check out the measurements arend-jan provided in the following thread. He found that higher coating resistance kept distortion low even if a panel was built asymmetrically with one stator spaced further away from the diaphragm than the other. I would think this would be advantageous for curved panels where the natural tendency of the diaphragm is to situate itself closer to the rear stator if even a small amount of lateral tension is applied.

http://www.diyaudio.com/forums/plan...rtion-electrostatic-speakers.html#post1860178
 
Last edited:
I've been down with a cold for a few days now and maybe it's the fever talking?
Is there a reason why you can't stretch the Mylar film on a flat bed and then "roll" the back stator onto it?
You could still stretch it only lengthwise.
If the stator already holds the curved shape all should be fine?
Then just cut the mylar, apply the coating and "roll" on the front stator?

I'm assuming foam tape for mounting.
 
right Peter Walker=Quad, in my recent research in looking at Elvamide for the coating of the mylar, they said that they coat both sides of the mylar in the original quads for the mids and highs. Take a look at the link below, not to far from the top of the page list detailed instructions for using elvamide and there you will see something about how they coat the sides of the mylar.

Do It Yourself - Electrostatic Speakers - Material: - The Audio Circuit

Yeah right, you will consume the double amount of coating.
Don't do it.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.