| The ESL Build Thread - Click HERE for Original Thread |
| BillH |
This thread will detail my first ESL build. I'm still working out a few issues as I go, but the major components are either designed or built. Each ESL will have perforated steel stators, acrylic insulators, a mylar diaphragm, a 50:1 step-up transformer, and a HV power supply based on easily available low voltage transformers.
I wanted to do a multi-way driver. This quote is from an article by R.J. Matthys, printed in 'Electrostatic Loudspeaker Design and Construction' by Ronald Wagner: The efficiency of the two-way system is thirty-two times greater than a one-way system or single diaphragm speaker because of the reduced impedance mismatch. He goes on to say the efficiency of a three way system is about three times more than the two-way. The insulator design didn't allow a three-way, so I'm building a two-way.
Size was an issue, too. The 36" wide perforated sheet set the height. The width needed to be around the width of my RS225S-8 transmission line woofers. After some trial and error, I came up with two sets of 4.5" x 36" and (2) 1.5" x 36" stators per driver.
I'm going to try to get the mylar tensioned to get a resonant frequency of about 90Hz on the 4.5" midrange section and 270Hz on the 1.5" tweeter section. The woofer to ESL crossover is set by the step-up transformers to 225Hz. The ESL crossover will be somewhere between 500 and 1000Hz. Because of its relatively large size, the tweeter's dispersion starts to fall around 3000Hz.
The insulators are described in the Stacked Acrylic ESL Insulator thread. Here's a rendering from Solidworks showing a completed panel. The acrylic insulator is yellow and the stators are red to make it easier to see. The whole driver will be white in real life. |
|
|
| BillH |
The biggest challenge with the perforated stators was deburring all those holes. I settled on a knocking down the burrs with a file and finishing with a round wire brush in my drill press. With the drill press table touching the wire brush I could slide the stator back and forth to round over the perforated holes. My US$80 drill press strained and I had to let the motor cool frequently. It suffered no major damage.:D
Here's a photo of the drill press and one of the smaller stators with a 3/8" brass screw for electrical connection soldered in. The screws were later changed to 1/2" steel. The screws are attached with 60/40 rosin core solder in a 1/4" milled hole. The solder connection is weak, so I'll need to fill the screw's clearance hole in the insulator with epoxy so it doesn't pull out of the stator when tightened down. Next time I'll use a smaller, countersunk hole in the stator. The stators will be powder coated white. |
|
|
| BillH |
Audio step-up transformers with a budget like mine were a problem. In this thread, Calvin shows very positive results for using toroidal voltage step-up transformers as ESL audio transformers. I wasn't able to find the toroids I needed at a reasonable price and the custom transformer winders didn't seem interested in a two piece order.
After searching the 'net I bought a pair of M99B transformers from Russ at JustReal Music. 50:1 turns ratio, 225 Hz low frequency crossover, and very reasonably priced. |
|
|
| furly |
Hi Bill,
In the thread where Calvin discussed his results with the torroids, I remember you had said you had ordered 4 of these toroids . Just wondering if you tried them out? I have just finished building my first set of panels using these. |
|
|
| BillH |
Hi, furly. I never did get those toroids. I posted to diyAudio just before I ordered them. When I got to ordering, they were gone. It was a lesson learned that day, my mistake cost me another US$100 to get a good pair of transformers.
How are they working out for you? |
|
|
| BillH |
Adhesives are the dilemna of the day. I need to attach powder coated perforated steel stators to acrylic insulators. Some options are discussed here. I'm dealing with low surface energy materials which really narrows the choices of adhesives. The other challenge is the small surface area of the stator that will be glued. I've only got a 1/4" wide insulator area to glue a 50% open stator to. The same adhesive thickness needs to be used at all points to keep the planned 0.054" stator to diaphragm spacing.
3M #4932 and #4952 double sided VHB tape look like a good solution. I'll need about 32 yards of it it to hold everything together. The cost is the only thing keeping it from being my first choice.
I've been looking at some readily available adhesives lately to hold the stators to the insulators. Hardware store variety 30 minute cure epoxy was a disappointment. It sheared between the two parts easily.
Regular solvent based contact cement is a maybe. I tried one coat on each part and the bond seemed strong, but didn't end up with much contact area. Two coats on each part would have probably worked better. 3M has a #4693H contact cement (.pdf) made for low surface energy materials that looks interesting.
Tonight I'm trying common silicone sealant/adhesive on a stator and insulator sample. It's the smelly, acetic acid laced style of silicone. I've used it in bathrooms and it has been nearly impossible to remove from any surface it hardens onto. |
|
|
| Bazukaz |
Hi,
I have had success with super glue.I used plexiglass stators , don't know how does it bond to your ones. It then needs to be pressed uniformly for a few minutes.
Lukas. |
|
|
| I_Forgot |
I have used scotchgrip 4693 and it works VERY well, including the most difficult job of holding the tensioned diaphragm. I have some drivers I built years ago using the 4693 and they are as strong as new. The only problem with the stuff is that there is no second chance. If you make a mistake during assembly, you will have to start over...
The only other problem is obtaining the stuff. I have gone through the Thomas Register and called about twenty 3M distributors and not been able to find one that will sell me a small quantity or even stocks the stuff. If you find a source, please let me know.
I_F |
|
|
| tommak |
McMaster's has Scotchgrip 4693 in 1 quart, 1 gallon, or 5 oz. tube. Just go to
http://www.mcmaster.com/
and search for "scotchgrip" or go to page 3229.
Tom |
|
|
| I_Forgot |
Of course! Why didn't I think of that?
Thanks,
I_F |
|
|
| BillH |
Today's build had a couple of failures and one success. | quote: | Bazukaz writes:
I have had success with super glue.I used plexiglass stators , don't know how does it bond to your ones. |
Lukas, thank you for the suggestion. I tried a small powder coated piece of perforated steel glued to a scrap of acrylic. It works great! When I tried to peel the perf. off, the powder coat came off the steel and stayed stuck to the acrylic.
The bare metal behind the white acrylic in the picture is 0.059" thick and was glued on two sides. I bent it when I pulled it off. :D |
|
|
| BillH |
The powder coating of the stators was a learning experience. When the stators came out of the oven, all of the sixteen solder joints that were holding in 6-32 screws for electrical connection failed. I had used soft solder and didn't consider at the time that the melting point of soft solder is very close to the curing temperature of powder coating.
My ego was bruised. Then I thought of something a boss said to me many years ago after I had made a whole order of parts wrong. He said "Everyone makes mistakes, you're not the first. You have to fix your mistake. Salvage all you can, do it quickly and move on." I've got a lot of time invested in the stators and didn't want to start over.
The picture below shows a a pair of small stators before and after repair. They were sandblasted, resoldered, and coated in the sandblasted area with Rustoleum Paint for Plastic. |
|
|
| BillH |
The insulators with stators are all glued up. It went very well. I used a total of 24 grams of super glue to hold the sixteen stators in.
Here's a picture of two completed assemblies. Inside shown on the left and outside on the right. |
|
|
| BillH |
The audio step-up transformers came in today. It's a pair of M99B's from JustRealMusic.
They look well made and came with two pages of photocopied handwritten instructions. |
|
|
| BillH |
I've been working on a two-way crossover circuit for the 1-1/2" wide and the 4-1/2" wide sections of the ESL panels. I've got a first order crossover at 1040 Hz that sims well using available components.
Here's the simplified circuit. The audio step-up transformer is connected on the left. The bias supply is connected on the right. Each of the four stators is connected to one of the center four points. |
|
|
| BillH |
| Because of the high voltages and power dissipation of 2 and 6 watts, the resistors are series connected 1 watt Xicon carbon film. The capacitors are series connected 4700pF 1Kv Murata SMD. I'll be using SMD capacitors on the bias supplies, too. The board measures 2 x 6-5/8". |
|
|
| BillH |
| Here's the bias supply board. It's based entirely on Sheldon Stokes' Simple ESL Polarizing Supply (.pdf file). The only changes were to use SMD capacitors in the voltage multiplier and include a 100 Megohm series resistance on the board. The board measures 2" x 8" |
|
|
| BillH |
Parts came in today. Almost everything is there. Two transformers are back ordered, I ordered the wrong IEC connectors, and forgot to order the 330K resistors for the bias supply. I'll have to dig through my salvaged parts for IEC connectors and resistors.
The SMD capacitors are really small! There's a strip of twenty of them circled in yellow in the photo. |
|
|
| BillH |
Here's a photo of the almost completed circuit boards for one channel.
The high voltage bias supply is at the top. It's missing a transformer that should be here this week. The black wire is a jumper to select output voltage. The inset image in the bottom left shows some of the SMD capacitors on the solder side of the board.
In the middle is the two-way crossover. It has four SMD capacitors on the solder side of the board.
The two boards at the bottom are resistors for the input side of the audio step up transformer. They're 1 ohm 10 watts per board. I wasn't sure of the value I'll need, so I made two for each channel.
The boards were done using the laser toner method. It's the first time I tried it and it worked quite well with only one defect that couldn't be fixed with solder bridges. |
|
|
| Bazukaz |
Hi,
Here is my first wire ESL. I is not finihsed yet , but sounds already:).
For now , i have only a "U" power trafo to test it. Sound is not that bad , even with poor transformer , and it has bass, though i feel that a sub is needed.
Regards,
Lukas. |
|
|
| BillH |
Nice work, Lukas. It looks sturdy, too.
| quote: | | Sound is not that bad , even with poor transformer , and it has bass, though i feel that a sub is needed. |
A larger transformer would help the bass somewhat, but we still need subs unless the ESL panel size is large.
How large is your panel? |
|
|
| Bazukaz |
Hi,
The panel's effective area is 19,5 x 140cm.
For now , i am trying to direct drive panel with FET's , at 400Vrms it sounds no quet , but the amp is still a bit too distortive.The bass sound deeper with direct drive amp.
Regards,
Lukas. |
|
|
| dstockwell |
| BillH any updates.. |
|
|
| BillH |
Hi, dstockwell. I took some time to design the frame for the ESL panels. I tend to overengineer things, so progress has been slow. I'll have more pictures of the finished frame soon.
Here's a work in progress picture of the frame with the ESL panel and electronics temporarily in place.
A closer look at the electronics.
My wife and I got a new dog that's been keeping us busy. She's a German Shepherd named Kola. In between eating and running around the house, she takes a break to sleep.
I'm trying a new image hosting service. Let me know what you think about the pictures. Click the pics to see them full size. |
|
|
| dstockwell |
| Looking good.. Image hosting works good, pics plenty big enough. |
|
|
| BillH |
The enclosure and wiring are almost done. The outer plywood skin measures 43-1/4" x 15-3/4" x 3-3/4". Inside of that, the inner front and back frames hold the ESL panel in place, attached to each other with eighteen 8-32 screws.
The ESL panel is held together with four 4-40 nylon screws, one in each corner, just to hold it together until the panel gets mounted to the frames. The edge of the panel gets squeezed between the frames with a thin, self adhesive foam from the craft section at Walmart between the frames and the panel. I still have to pick out a foam color. :D
I've tried to make the ESL panel as stiff as possible with 1/4" thick acrylic vertical ribs. I needed a way to hold the inner part of the ESL panel together and I didn't want to glue the two halves of the panel together. The small grey horizontal strips in the picture are 5/32" diameter round steel bars, six in each frame, one behind the other. They're deflecting about 1/8" in the middle, squeezing the two halves of the ESL panel together when the frames are tightened. |
|
|
| BillH |
The build is still moving along. There's a lot of small details to take care of before final assembly. I've made another order to Mouser to get the parts I forgot on the first order from them. They should be here in four days. I'm bidding on some walnut veneer from a seller on Ebay to cover the enclosures. The enclosures have been sealed on the inside with clear polyurethane. I still need to paint the steel tension rods.
Here's a drawing of the stretcher I'm building. The outer part of it is glued and the whole stretcher should be done in a couple of days. It's built with 2" x 4" lumber salvaged from pallets. The diaphragm is tensioned by tightening nuts on the running thread on two sides. I'm not sure how well it will work with pull on only two sides, rather than four. The dark bars in the inner four pieces represent slots that will have 7/64" plastic spline pressed in to hold the diaphragm n place while stretching. It's the same spline used to hold the screen in screen doors. |
|
|
| BillH |
Here's a photo of the stretcher with the first piece of Hostaphan diaphragm installed. I learned some things that will need to be fixed on the next try. The first thing I noticed was how hard it was to get the diaphragm tensioned equally as I pushed the plastic spline into the grooves around the outside. It turned out OK, the multiple threaded rods let me skew the tension to get a nice flat film.
I tapped the diaphragm with a finger, tensioned some more, tapped again, tensioned again until it sounded like a drum head. It works!
Then it happened... A small rip revealed itself and that's why you see ripples in the picture. As I was inserting the spline earlier, the spline tool slipped and must have cut or weakened the diaphragm. No problem as this was a test run. One of our dogs put her foot through the diaphragm shortly after this picture was taken and I accidently put my elbow through it as I was bringing it back in the house. Gotta treat the Hostaphan a little better next time. :D
I'm having trouble with the digital camera, it cuts off the bottom of pictures. The bottom of the frame is actually as wide as the rest of it. |
|
|
| I_Forgot |
I think it will be difficult to get good results from such a diaphragm stretcher. There are a couple problems with it, and you can see evidence of them in the photo.
You have secured the film to the stretcher bars by trapping it between a spline and a groove in the bars. The problem with this is that it does not allow the film to stretch at the point of attachment to the stretcher bar. That means the majority of the stretching is occuring at the corners where the film is not so constrained.
If you stretched the diaphragm in one direction, say the y axis, with the edges of the film unattached to the bars that provide the x stretch, then the film would stretch in the y direction, no problem. Now if you attach the film to the x bars and try to stretch in the y direction, what will happen? It will wrinkle because you're not allowing the edges of the film to stretch in the y direction. Now restrict both x and y edges of the film and try to stretch it and what do you get? You will put very large tension on the unrestricted corners of the film while putting almost none at the center. Your film stretching will have to stop before the diaphragm splits, which will occur first at the corners since they are under the most tension.
High tension prevents the diaphragm from moving too far and hitting the stators. In the corners and at the edges of the speakers, the film can't hit the stators because it is supported by the insulators, so having high tension in the corners and at the edges is not so important. The center of the diaphragm is where you have to worry about hitting the stators, so that is where you really need the tension.
The ideal(?) situation would be to have uniform tension over the surface of the diaphragm. This can be achieved by using a circular stretcher, but it is usually not practical due to the width of the diaphragm film that is comercially available (works great for headphones and microphones). Since uniform tension is not practical, the next best thing is to get the tension at the center of the film up so that it will be less likely to contact the stators.
You can make a much simpler stretcher that will provide more uniform and much higher tension by simply stretching a bicycle tire tube around a table that is a little larger than the speaker. Lay the film out over the table and uninflated tube and attach the film to the inside/underside edge of the table using either tape or a spline and groove.
Inflating the tube will stretch the film very tight, with more uniform results than the x-y stretcher. As you inflate the tube you will notice that the film is able to slide/stretch over the tube (unlike the spline and groove in your design). You will also notice that the corners inflate less than the long dimensions of the driver. This means the tension in the central area of the diaphragm will be quite high, while the corners will be lower and less likely to split. The air in the tube can and will go wherever it needs to to make the tension approximately uniform. If the film is pulling tight at one location, it will restrict stretching of the tube and force the air to go where there is less tension, thus equalizing (approximately) the tension everywhere.
I have not tried to make a stretcher table for such a long, narrow driver, so you may run into some problems there, but for drivers that are a little closer to square, the pneumatic stretcher works great. I suggest you give it a try. It only takes an hour or so to make one. It shouldn't matter that the table is bigger than a tire tube. Tubes stretch a lot.
I_F |
|
|
| Calvin |
Hi,
as Forgot said, Your stretcher is not designed for very good results.
I used a quite similar stretcher to good results but with the following differences.
In the Y-direction (long) I divided the moveable parts into 2 to 3 sections, each section with two threaded rods with nuts to screw the diaphragm tight. I used 1 section on each of the X-sides. So there are 6 to 8 sections connected to the outer frame. The threaded rods of the Y-sides run in small slots, so the sections can move a bit when stretching in the Y-direction. While the outer frame has to be very strong the moveable parts can be made from coated MDF-board, 30mm wide. You can glue the diaphragm to it with double sided sticky tape. This way You can get very high tension in one or both directions. The sticky tape is easily removed after stretching. If You replace the straight X-sections by bowed ones You can even stretch the diaphragm for ML-type panels ;)
jauu
Calvin |
|
|
| BillH |
I_F and Calvin, thanks for the input.
I did have a feeling something wasn't quite right with the stretcher. I may be able to equalize the stretch better by securing the y first and stretching it and then securing and stretching the x. The spline seems to pull the film too much when it is installed. I'll try tape to hold the film instead and see what happens.
I_F, I originally wanted to try the inner tube stretcher, but a 27" inner tube, the largest I could find seemed too small to fit around a jig for my panel that measures 14-3/4" x 37-1/16". The other difficulty I had with it conceptually was getting the width and height to stretch unequally. Since the y must stretch more than the x, do you think an inner tube would work as a rectangle? It seems to be a perfect solution as long as the jig is a circle.
Calvin, I could easily separate the stretcher into multiple sections and oval some of the clearance holes for the threaded rods.
I like both of your ideas. I'll try them this weekend and get back to you. |
|
|
| Bazukaz |
Hi,
It is possible to use tape to hold mylar. But you need to bend it over edges , otherwise it will not hold the tension.Also , a well-adhering tape is required. The cheap transparent packing scotch is not suitable here.
I think that you could also use wooden sticks with screws to press the mylar to frame. Also , you could glue some type of rubber on the stick,
so that mylar would not slip out.
Here is the idea that i used to strech mylar.I used slow-bond epoxy.
Note : the transparent tape shown in picture is not good - it does not bond well to wood.But there are many type of tapes that bond much better. |
|
|
| I_Forgot |
| quote: | Originally posted by BillH
I_F, I originally wanted to try the inner tube stretcher, but a 27" inner tube, the largest I could find seemed too small to fit around a jig for my panel that measures 14-3/4" x 37-1/16". The other difficulty I had with it conceptually was getting the width and height to stretch unequally. Since the y must stretch more than the x, do you think an inner tube would work as a rectangle? It seems to be a perfect solution as long as the jig is a circle. |
If a single tube won't stretch to fit, you can use two tubes, one for the upper and one for the lower half of the driver. The middle section, where the two tubes are laying next to each other can be wrapped with duct tape to keep them from stretching and force the air out to where it needs to be.
--
| |
| | top half
--
--
| |
| | bottom half
--
The air in the tube(s) will take care of making the tension even. While inflating, if the tension on the film gets high on one axis, it will keep the tube from stretching that axis any further and addional air will go to stretching the other axis. With two tubes you'll need to match the inflation pressures, but that shouldn't be hard to do.
Let us know how you manage it...
I_F |
|
|
| Capaciti |
Hi Folks,
why messing around with mechanical stretching ?
- Put mylar on an even surface
- Stretch it, so no major wrinkles are visible (it don't need to be uniform)
- Glue it on the spacer frame
- Heat it with a heat gun at about 150°C
- Each panel will be within +/- 10 Hz resonance
Thats it
capaciti |
|
|
| I_Forgot |
Heat shrinking does not provide sufficient tension to keep the diaphragm stable with high bias voltage. You have to use a low bias voltage to prevent the diaphragm from flapping, so the speaker will be very insensitive.
I_F |
|
|
| Capaciti |
Hi I forgot,
Thats true, but a mechanically stretched membrane will increase its compliance and reduce resonance over the time.
people wonder why a new mechanically streched membrane ist stable first and collapses months later.
My experience is that over years the stretched membrane "creeps". It stops exactly at the tension of a heat shrinked . When it stops depends on listening level and membrane excursion.
Capaciti |
|
|
| Brian Beck |
Capaciti,
While I mostly agree with your point, I have witnessed creep relaxation with heat shrunk membranes over time too, although perhaps at a lesser rate than with mechanically stretched membranes. These properties are probably material specific. |
|
|
| Capaciti |
Hi Brian,
the relaxation is less than 10 % for a heat shrinked and more than 50% for a mechanically stretched membrane.
Why does mylar creep ?
at manufacturing the polymer-chains of mylar are stretched and oriented to a certain level, in order to provide the best mechanical performance.
When you stretch mechanically, you give the polymer-chains another orientation and you further stretch the chains. All the chains are linked to each other at several connection points. Those points crack more and more by time and movement, enabling the polymer to expand further, thus lossing tension. The connections brake due to the tension overload of the stretched membrane.
When you heat shrink, you weaken the matrix of the ploymer-chains. Doing this the orientation and stretching of the manufacturing will be reversed and the membrane shrinks. It shrinks exactly to a size, where the internal stress of the polymer-chains is balanced with the increasing tension due to shrinking. By giving enough heat, you allow each chain to "find" ist best fit for orientation and internal stress. This is the only way to expect longterm stability. The heat shrinked membrane will slighty loose tension, since some interconnections of the chain are "damaged" by the heat. After some cycles (a few hours of listening) the damaged connections will brake. After this is finished, the membrane is in its final and stable condition.
Capaciti |
|
|
| Brian Beck |
| Interesting. Thanks for that insight. |
|
|
| BillH |
| quote: | | Capaciti writes: the relaxation is less than 10 % for a heat shrinked and more than 50% for a mechanically stretched membrane. |
So it should be possible to overtension and allow the film to relax to the correct Fo, but how long will it take?
I searched the web for other's experience with the Hostapan film I'm using and found this page which says in part:
A new batch of Hostaphan RE-6 (6 micron, 795 mm in width) has arrived, a perfect material to use as diaphragm material. It is the German equivalent to Mylar C, a tenselized PET film that will heat shrink for about two percent, ideal to get a perfect tension for your membrane.
And from this page:
I do not recommend heat to shrink the mylar, at least in Hostaphan the film the author uses, as the elongation factor is quite low, and is not the same in both dimensions and most importantly the applied tension is not easily controllable. Heat, however may be useful in adjusting small differences in tension to enable the matching of the Fo of a particular panel. If the Fo of a particular panel is too low hot air from hair drier or heat gun can be used to increase tension slightly and bring the Fo up to the other value. The procedure must be carried out with care and the tension increased gradually until the desired Fo is reached.
I think the second method has merit. If heat shrinking two films without an equal starting tension, it could be hard to get them both shrunk to the same Fo. You'd also be able to apply more tension with heat later when they stretch. |
|
|
| BillH |
| quote: | I_Forgot writes: If a single tube won't stretch to fit, you can use two tubes, one for the upper and one for the lower half of the driver.
--
| |
| | top half
--
--
| |
| | bottom half
--
|
How would you would get the diaphragm coated and attached to the stator while it's in the stretching jig? Wouldn't the center, where the two tubes pass each other be in the way? |
|
|
| Capaciti |
Hi Bill H,
the tension applied by 4 weights of about 100-500grams only is much less than mylar or hostaphan can sustain. So this method do not fully utilize the mechanical capabilities.
Thats why the author mentions that heat shrinking will increase fundamental resonance.
You are correct that the shrinkage rate is different depending on axis orientation, but even 2% is more than enough headroom to stretch a membrane which is glued to the frame having visible wrinkles.
Capaciti |
|
|
| BillH |
| quote: | | I_Forgot writes: The tube(s) are under the diaphragm, around the edge of a "table" made of a piece of plywood. |
Ah, now I understand. My situation is a little different. I've been planning to have a stretcher with an open middle section and a 1-1/2" thick removeable table slightly larger than the ESL stator. One side of the table, made from two layers of 3/4" chipboard will be flat and the other side will have spacers to pass my acrylic supports when the stator is set on it.
The order of assembly would be:
1) Mount the film to the stretcher and give it enough tension to get the wrinkles out.
2) Insert table with flat side up into stretcher and apply graphite. 3) Retension to Fo.
4) Remove stretcher with the film still on it, flip table, and set stator on table.
5) Apply glue to stator and drop the stretcher and film on to the stator.
The 'Charge Ring' made of copper foil will be on the other stator, in contact with the graphite.
| quote: | | Capaciti writes: You are correct that the shrinkage rate is different depending on axis orientation, but even 2% is more than enough headroom to stretch a membrane which is glued to the frame having visible wrinkles. |
It would be so much easier to just glue the diaphragm and heat it as you recommend. My only concern would be trying to get a desired Fo with a material that has a finite heat stretching limit. These panels are meant to have a ~100 Hz Fo on the 4-1/2" x 36" sections when assembled, but I need to stretch the diaphragm as a single 12-3/4" x 36" unit. I'm doing the math now to see what that Fo would need to be. |
|
|
| jmateus |
Here's a tension tool that I've been using in stretching the mylar
with great success. It works around the membrane stretching
it as I go around, fixing the stuff with a strong tape. After done
all I have to do is to glue the frame to it, it remains stretched
forever....
This is a smart tool bought from Eraudio.com.au sometime ago. |
|
|
| Few |
jmateus: Could you attach a full-sized image of the tool you're describing? I'm having a hard time making out the details in the thumbnail-sized image.
Thanks. |
|
|
| I_Forgot |
| quote: | Originally posted by Capaciti
Hi I forgot,
Thats true, but a mechanically stretched membrane will increase its compliance and reduce resonance over the time.
people wonder why a new mechanically streched membrane ist stable first and collapses months later.
My experience is that over years the stretched membrane "creeps". It stops exactly at the tension of a heat shrinked . When it stops depends on listening level and membrane excursion.
Capaciti |
Months? I built my first ESL drivers about 15 years ago. They were mechanically stretched. Those drivers still play as new, even with >3 kV bias. I used 6 um thick "lumilar" polyester film made by Toray.
I have tried heat shrinking the film and it works but the tension can never get near what I can do with the pneumatic stretcher. I have used the pneumatic stretcher to replace some diaphragms in my Quad ESL-63s and they are perfectly stable even with the 5.25kV bias voltage and the thin insulators.
Quad used a mechanical stretcher for the original diaphragms. I saw a picture of it in Speaker Builder magazine once (or was it Audio Amateur?). It was some complex mechanism with dozens of little fingers that pulled on the film. When a Quad diaphragm fails, it usually splits near the edge of the driver, or the glue lets go. I have never heard of one "relaxing".
I_F |
|
|
| I_Forgot |
| quote: | Originally posted by BillH
The order of assembly would be:
1) Mount the film to the stretcher and give it enough tension to get the wrinkles out.
2) Insert table with flat side up into stretcher and apply graphite.
3) Retension to Fo. |
How do you determine F0? | quote: |
4) Remove stretcher with the film still on it, flip table, and set stator on table.
5) Apply glue to stator and drop the stretcher and film on to the stator. |
I think you may find that having the table separate from the frame and having to move parts in and out over the tensioned film will make it hard to assemble the speaker without poking holes in the film, especially considering how large the driver is and therefore how heavy the frame and table must be.
| quote: |
These panels are meant to have a ~100 Hz Fo on the 4-1/2" x 36" sections when assembled, but I need to stretch the diaphragm as a single 12-3/4" x 36" unit. I'm doing the math now to see what that Fo would need to be. [/B] |
How did you decide that the resonance should be 100 Hz? How do you translate the tensioning process into a predictable F0? How do you square the tension required for a specific resonance with that required to remain stable at any particular bias voltage?
As far as I know, any rectangular shaped driver will have at least two resonances- one corresponding to the short dimension and another corresponding to the long dimension. There will also be resonances at harmonics of the two fundamental resonances.
Finally, I suggest that you forget the graphite and use Licron. The only complaints I have heard about it is that it doen't give an optically clear result, and it is a little expensive. You won't get that with graphite either, but Licron application is very easy, and doesn't risk tearing the diaphragm. The resistivity is MUCH higher than you can achieve with graphite, which keeps distortion to a minimum.
I started using Licron about 6 months ago, and the drivers are still playing as new. Others have reported a least a couple years (and counting) on this forum.
I_F |
|
|
| jmateus |
Hi Few
Unfortunately I wasn't able to get a better image of the stretcher
no matter what I did, I have a hell of a time trying to get a good
picture every time due to the constrains of size.
Anyway the strecher is a very pratical device that does the job,
the red part is a dinamometer calibrated in kilograms, the white
part is an aluminun plate where you tape the membrane temporarily until you stretch it to your wish. Once you attain
the desired tension you untape the membrane and tape it again against the surface you have it down. And you go around the
membrane doing the same and paying attention to the desired tension
which should be the same around the perimeter of the mylar.
At the end you'll have a tensioned membrane to your requirement
taped all around. Then you need to glue the the structure of the stator to the membrane, that's all.
I hope this long explanation described the stretcher to your contempt. |
|
|
| Brian Beck |
| Your device, as best I can tell from the small picture, looks a lot like a portable fish weight scale. It is a simple spring-loaded thing with scale markings. It measures tension. |
|
|
| jmateus |
Yes, it is, yes it does...
Very efficient, though. |
|
|
| Few |
| Jmateus: Thanks for the description. Even without the larger image I now get the picture (sorry...). |
|
|
| BillH |
| quote: | | How do you determine F0? |
I was going to try setting up a microphone-->pc sound card-->fft software and tap the tensioned film to see if I could read the resulting frequency with the fft software.
| quote: | | How did you decide that the resonance should be 100 Hz? How do you translate the tensioning process into a predictable F0? How do you square the tension required for a specific resonance with that required to remain stable at any particular bias voltage? |
Roger Sanders' book has a table that shows lowest frequency on an ESL panel vs. diaphragm to stator spacing and my ~0.055" spacing gives about 200 Hz. I figured Fo one octave lower should be about right. I may have to increase the 200 Hz Fo a bit. The transformers I'm using have a recommended 225Hz lower limit.
| quote: | | Finally, I suggest that you forget the graphite and use Licron. |
I have to agree. I tried graphite as the low cost alternative for coating, but had questionable results on a few test pieces of film. I had a very hard time getting an even, high resistance coating. It was easy to get it too little or too much graphite on. I'll have to order the Licron. Mouser.com has it in stock. I still need to order four SMD capacitors for one of the crossovers anyway that I've forgotten to order twice. Gotta give it a day or so to see if I need anything else.
Now I have a question for the group. I've built the bias supplies with 100 meg series resistance because of a formula in Sander's book that predicts resistance needed to provide constant charge according to lowest frequency desired. I don't have the book handy right now, but if you need the formula let me know. Most of the bias supplies I've seen on the net are using 10-20 meg resistance. Do you think it makes any difference?
The stretcher jig version 1.1 is done and looking good if I must say. Check back later for a picture. |
|
|
| BillH |
| quote: | | I wrote: These panels are meant to have a ~100 Hz Fo on the 4-1/2" x 36" sections when assembled, but I need to stretch the diaphragm as a single 12-3/4" x 36" unit. I'm doing the math now to see what that Fo would need to be. |
If my math is right, I would need to tension the panel as a whole to ~22Hz to get the 4-1/2" wide sections to 100 Hz Fo. I don't know if I would be able to accurately measure 22Hz. Is this desire to meet a certain Fo a waste of time?
Here's another question for the group, but first a recap of the stator dimensions. Each ESL panel has two pairs of stators arranged as four 36" long strips. The two outside strips are 4-1/2" wide and the inner two strips are 1-1/2" wide. Do you think I'll need any extra horizontal support anywhere along the 36" dimension? I had thought of splitting the length into two or three shorter sections. |
|
|
| Bazukaz |
Hi,
I think you don't need extra support for 4 1/2" x 36" sections.
Regards,
Lukas. |
|
|
| jmateus |
Few
I didn't rest until I was able to upload the picture of the tension tool... I got to convert it to 57 Kb.
I forgot to tell you that you pull the ring to stretch the mylar.... |
|
|
| I_Forgot |
| quote: | Originally posted by BillH
If my math is right, I would need to tension the panel as a whole to ~22Hz to get the 4-1/2" wide sections to 100 Hz Fo. I don't know if I would be able to accurately measure 22Hz. Is this desire to meet a certain Fo a waste of time? |
I have a couple comments here. First, I have never been able to discern any audible difference in panels other than the low frequency output level based on anything other than the combined area of the drivers. Thin insulators limit maximum excursion and so limit the maximum volume at low frequencies. But you can't separate that limit from the front to back phase cancellation at low frequencies. So is what I hear due to F0 or due to the phase cancellation? Both, but I suspect the overall size is far more important than F0 of any driver or section of any driver.
I think your planned method of determining F0 is flawed. I think you would need to apply a pulse to the transformer driving the ESL with the full bias applied in order to measure the F0, then you still have the issue of front to back phase cancellation to deal with. Close miking may solve that problem.
In general, I think worrying about F0 issues is not very productive. F0 takes a back seat to putting enough tension on the diaphragm to keep it stable under bias conditions. You are using very thin insulators which will lead to very sensitive operation, but will require very high tension or low bias voltage. Since you're trying to make a very sensitive driver, you won't want to lower bias voltage, so diaphragm tension needs to be very high.
| quote: |
Do you think I'll need any extra horizontal support anywhere along the 36" dimension? I had thought of splitting the length into two or three shorter sections. |
I don't think you need to add any additional support.
I_F |
|
|
| Few |
jmateus: Thanks for the extra effort. Now I truly have the picture.
BillH: I agree with others that you'll not need extra support points along the 36" length. The most important factor is how far any point on the diaphragm is from the nearest supports. If the nearest supports are quite close in the horizontal direction (as yours are) I think little or nothing will be gained by adding supports along the vertical direction. You would just pick up extra capacitance that loads the amp but doesn't contribute to sound.
On the resonance issue, it's not clear to me a 22" wide diaphragm with a 22 Hz resonance frequency will yield four diaphragm sections each with resonance frequencies of 100 Hz after you've separated the diaphragm into vertical strips. If you only have one diaphragm section and you shrink its horizontal dimension by a factor of four then the resonance frequency should increase by a factor of four. In your situation, though, you'll actually be creating an array of four such diaphragms. Since each section of the array can acoustically couple to its neighboring sections, I'd expect the resonance frequencies to be modified.
This may be more than anyone wants to read about, but I'll throw it in anyway: By analogy with other coupled oscillators I'd expect to see a set of four resonances whose frequencies are similar but not identical to the single resonance expected in the case of zero coupling. In other words, the single resonance is "split" into four distinct resonances when the coupling is included. If the coupling is sufficiently small, and if the resonance is sufficiently damped, then the set of four peaks may appear to be one big resonance peak.
The bottom line is that in practice the factor-of-four prediction may get you close enough, but I wouldn't be surprised if the numbers didn't work out perfectly. In any case, I think the most important thing is that the two panels match and that the resonance frequencies are in the right range.
Few |
|
|
| Few |
I posted my recent response before seeing I_Forgot's post so I'll tack these comments on here.
I agree with I_F that for your application achieving high tension is the key point. However, if you intend to run your panel down to, say, 250 Hz, I don't think you'd want so much tension on your diaphragm that the resonance frequency ended up above 250 Hz. That would considerably complicate the transition between the ESLs and the dynamic woofers that I'm assuming you're using in conjunction with your panels. It would also place a severe limit on the SPLs you can achieve. Any musical content at the diaphragm's resonance frequency would drive the diaphragm into a stator. For those reasons I think the resonance frequency does matter somewhat in your application.
I'm not sure I'd give up on your plan to measure the resonance frequency of the diaphragm by tapping on it and using your soundcard to look at the resulting spectrum. The fundamental resonance is very easy to hear and I'd expect it to be easy to see on a spectrum as well. I_F's idea of using close miking seems like a good one to me. |
|
|
| I_Forgot |
The bias voltage will change the resonance frequency. That is why I suggested that an electrical pulse into the speaker should be used. The problem is that you have to build the driver first, which makes it hard to adjust the resonance if you decide you don't like it for some reason. I suppose you could add pieces of adhesive tape to the diaphragm to increase its mass and bring resonance down, but then you introduce breakup modes and distortion.
Quad uses a thin porous cloth on one side of the drivers to damp the fundamental resonance of the panels. Maybe that's a better idea than trying to achievea specific resonance.
At one time I toyed with the idea of making a speaker in which the pneumatic stretcher was an integral part for the structure. Air would be the only insulator between the diaphragm and the stators. The pressure in the tubes could be altered as the speaker played to test audibility of resonances and stability vs. bias voltage vs. stator spacing. Someday, if I have time...
I_F |
|
|
| Few |
| I understand that the bias voltage causes the diaphragm to be displaced from the central "neutral" position but I've not noticed a difference between the tone generated by tapping on an ESL panel with the bias voltage on versus tapping on the same panel with the bias voltage off. I have to admit I haven't done a careful study, so I could end up eating crow, but it would come as a surprise to me if the change in frequency was significant. I_F, have you done a careful enough experiment to be able to report what the typical magnitude of change is? Are we talking about a 1 Hz change in a 100 Hz resonance or something larger? |
|
|
| Calvin |
Hi,
its good to apply the 1:70 to 1:100 rule. This means that the diaphragm should be supported at points where the shortest distance measured over the diaphragm becomes the 70 to 100 times the d/s. With a d/s of 0.055" the distance between two spacer elements should be 0.055"x70 = 3.85" to 0.55"x100 = 5.5". Having a width of 4.5" You won´t need any horizontal spacer at all. With these dimensions heat shrinking the diaphragm should give just under 100Hz of resonance. So in this caseI´d rather use the heat gun for ease of construction (the pic shows the nearfield response of a small panel with app. 1.5mm d/s and 90x750mm diaphragm. I use 3.5µm film heat shrunk i and coated it with a highly transparent solution I made myself).
When highly resistive coating is used the resistor for the high voltage cascade doesn´t need to be as big as Bill quoted. In fact You wouldn´t need it at all for these kinds of coatings. But its quite useful for experimentation and safety reasons to limit the maximum current in case of touching something while building or leakage. But since the voltage drop over the resistance reduces the bias on the diaphragm it should be kept as small as needed for safe working. 10 to 20meg are sufficient. With low resistance coatings higher resistance values will be needed (as this resistor smooths out the ripple voltage of the cascade high ohmic values filter stronger This might be of interest especially when the high voltage is taken from the ´half-points of the cascade´where a high ripple exists.).
jauu
Calvin |
|
|
| Bazukaz |
Hi,
I have found one type of shoe polish to work well for coating. I apply it with a parlon. Then i take a fresh piece of parlon , moist it and remove excess of polish.
The surface gets slightly whitish , and shoe polish actually dries out well and sticks to mylar quite good.It is much easier to work than graphite , and seems to give higher resistance (depending on amount of polish applied).
I have no idea about long-term reliability.
Best regards.
Lukas. |
|
|
| BillH |
| quote: | | I have found one type of shoe polish to work well for coating. I apply it with a parlon. |
Lukas, what lead you to try shoe polish? I'm not familiar with a parlon. Can you describe it?
I tried a couple of off-the-shelf household items in a search for an easy to apply, inexpensive coating. Anti-static spray meant to be used on clothes gave a suitably high resistance, was easy to apply, but didn't seem to dry enough to be permanent. Rain-X, made for car windshields was easy to apply, but had infinite resistance.
I'm going to try graphite again. Rub it in, remove the excess with alcohol, and see what happens. If it doesn't work, I'll order a can of Licron. |
|
|
| BillH |
Here's a photo of the stretching jig v1.1.
Thanks to all for your ideas and suggestions. I think you'll see many of them used in the latest version of the jig.
The v1.0 moveable parts on the long side have been divided into three pieces and the holes they are mounted to have been slotted to allow up to 5/16" of movement while stretching. I've added some foam blocks between the moveable parts and the main frame to stabilize the moveable parts.
There's no film in the photo. When it's in place, it will be held to the moveable parts with double sided tape. The splines in v1.0 just didn't work out.
In the center of the jig is a table used to apply the graphite. The table is removeable and sitting on six 3/4" thick MDF blocks that are attached to the baseplate. It's made of 5/8" chipboard with 1/8" masonite glued to it and a top layer of white paper. It's hard to see, but there are guidelines on the paper. There's a rectangle the same size as the ESL stator, and two other rectangles, each 1/4" inside the other. The middle guideline is where I'll put masking tape on the diaphragm to end the coating. The inner guideline is in line with the inner edge of the stator as a reference. |
|
|
| BillH |
Here's a photo of the v1.1 jig with the frame opened, the table removed, and the stator in place as it would be for gluing to the diaphragm. There's two hinges in the back that hold it all together.
There's no diaphragm in the picture, but it would be on top of the top section.
Shortly before I took the picture, I knocked the stick I was using to hold it open and the top hit me on the head. :xeye: I'm lucky I made contact with the frame and not the threaded rods sticking out of the frame. |
|
|
| Bazukaz |
| quote: | Originally posted by BillH
Lukas, what lead you to try shoe polish? I'm not familiar with a parlon. Can you describe it?
I tried a couple of off-the-shelf household items in a search for an easy to apply, inexpensive coating. Anti-static spray meant to be used on clothes gave a suitably high resistance, was easy to apply, but didn't seem to dry enough to be permanent. Rain-X, made for car windshields was easy to apply, but had infinite resistance.
I'm going to try graphite again. Rub it in, remove the excess with alcohol, and see what happens. If it doesn't work, I'll order a can of Licron. |
Sorry ,
i used a wrong word. I ment foam rubber.
I have found that with graphite it is hard to archieve uniform coating and the resistance is too low. Also , much of messy work. That's why i tried other coatings. I liked liquid soap much better than graphite , though some say that it "evaporates" over longer periods of time.
Regards,
Lukas. |
|
|
| maudio |
Hi,
my small contribution to the subject of stretching jigs:
Instead of wood I use pvc tubes for the moving inner parts to which the mylar must be attached. This solves the problem of obtaining a good strong bond between the jig and the mylar during stretching, which in my experiences can be troublesome with (unpainted) wood. I use small strips of double sided adhesive tape which sticks very well to both tubes and mylar. Also the rounded surface increases bonding and reduces risk of tearing.
In addition the tubes are mounted in such a way that they extend from the surface of the jig, I hope the picture clarifies.
This allows me to place the mylar on a flat surface and use some tape to lay it out flat without wrinkles (no tension applied yet). Then I place the jig on top of that, tubes facing down, "picking up" the sheet from table to jig.
For measuring tension I draw a rectangle on the untensioned mylar and carefully measure it's dimensions whlie stretching, which allows for pretty good reproducability. Because the tubes can flex a little I can evenly distribute the tension over the entire surface by adjusting the various screws.
About coating: various kinds of glue seem to work as well (dilute and apply by airbrush). Haven't tried it myself no no idea about long-term stability
Just some more ideas.. |
|
|
| Calvin |
Hi,
some glues work and some don´t. Even so they might look the same, smell the same and taste the same ;) You may have to experiment with different manufacturers.
As far as my experiences with my formula are it is stable now for several years. Just today I got my oldest panels out of the houses cellar with its damp -its a very old house with the ground beeing stamped mud- climate, open to frost and heat. I didn´t even dust them and they played instantly. I measured no loss of SPL and couldn´t see any damaged or missing area of coating -admittantly thats difficult, because its highly transparent ;).
The coating can be applied with a piece of foam as well as with an airbrush. The conductivity can be tailored to fit the application and the ingredients are human-friendly, water-based, not poisoneous.
I can recommend the use of glue-based coating.
jauu
Calvin |
|
|
| Paul W |
Calvin,
I am searching for a good transparent coating. Can you share the formula for the coating you use?
Paul |
|
|
| Calvin |
"Contact Paul W:.....Sorry! That user has specified that they do not wish to receive emails through this board."
jauu
Calvin |
|
|
| MJ Dijkstra |
Hi,
For more information about coatings go to audiocircuit.com
Look for section 'loudspeakers', than: 'electrostatic loudspeakers', than 'materials and parts' / coatings
You will find several contributions including professional stuff like the EC-coating. |
|
|
| BillH |
| quote: | Calvin writes:
The coating can be applied with a piece of foam as well as with an airbrush. The conductivity can be tailored to fit the application and the ingredients are human-friendly, water-based, not poisoneous. |
Hallo, Calvin.
I would also like to take a look at your formula if you're still willing to share it. It sounds quite interesting... |
|
|
| BillH |
I spent some time this weekend working on the back panels and veneering the ESL frame.
The back panels will hold IEC power connectors and binding posts for audio in. Somehow I managed to drill and dremel the holes in one panel backwards. :confused: I'll have to redo that one.
The veneering went quite well, no mistakes so far. The walnut veneer came from a seller on Ebay. I'll have pictures after the veneer has a couple of coats of clear on it. I should have the contact cement off my fingers by then, too. |
|
|
| tommak |
| quote: | Originally posted by BillH
Hallo, Calvin.
I would also like to take a look at your formula if you're still willing to share it. It sounds quite interesting... |
Calvin,
I also would be interested in your formula, if you would care to share.
Thank you,
Tom |
|
|
| Silberlot |
Hello
Look for Coating ?
Search Google .eBay Shop - ELEKTROSTATEN-SELBSTBAU-SHOP: Wir verkaufen Highend ...
Regards Dirk |
|
|
| Capaciti |
Hi,
if you are looking for an appropriate coating, the most important is that its not affected by humidity.
Most DIY-coatings are based on any kinds of glues. But all glue-based-coatings are not intrinsically conductive. glue based coatings form a specific surface, which "invite" water molecules to settle on the surface, thus increasing the surface conductivity. But depending on humidity the conductivity varies significantly.
Even it might still work at low humidity, the resistance is too high to allow fully charge of the membrane.
For my knowledge there are just few coating systems avaible which are permanently intrinsic conductive:
- Nylon disolved (but very high resistance, takes weeks to fully charge !!, some people dope it with ultrafine graphite)
- Baytron (modification by TDA,distributed by sigma-aldrich)
- Carbon black (epoxy with carbon-nano-fibers, need to be diluted, which is very difficult without creating clusters, currently used by Quad)
- Licron, Tech Spray (resistance too low for a true ESL, large additional mass)
- ATO (needs sputtering technique, used by ML)
- EC-Coating by Martin-Jan Dijkstra
Do not use ormecon-polyanillin. This material will oxidize by time and resistance will drift away.
Any others known ?
Capaciti |
|
|
| I_Forgot |
| quote: | Originally posted by Capaciti
- Licron, Tech Spray (resistance too low for a true ESL, large additional mass)
Capaciti |
How did you come to this conclusion? It sure seems to work fine in my speakers...
I_F |
|
|
| Capaciti |
Hi I_Forgot,
The surface resistance of licron is between 10e5 and10e6, which is too low for constant charge mode for frequencies lower than 300 Hz. If you use Hybrid-concepts with crossover at about 300 Hz its still OK.
I did some measurements on thickness and added mass of a licron layer.
The thickness for a homogeneous and functional layer is at least 6 micron. due to licrons density this corresponds to the weight of about 9micron mylar. This is much more added mass in comparison to other competitive coatings available. E.g. an EC-coating layer, which is btw nearby fully transparent, adds just a weight of about 2 micron mylar.
So why to use a good coating, when better is avaiable ?
Capaciti |
|
|
| Calvin |
Hi,
I will only share the formula with those I know from, they´re building a ESL. I don´t intend it to become open source. Send me a PN with a small description of Your panel for e.g.
I have to point out, that glue is one of the basic components. Its a material You can easily get in Germany. I don´t know about its availability in Amerika or somewhere else. You might have to test different glues to find the right one. On the other hand there are TNT, FedEX and others who´d be pleased to visit You.
Another component is a conductive Dope. With the add-on of this You can alter the conductivity of the coating considerably.
I don´t see any prob with varying surface resistances and humidity as the conductivity factor as long as the variation stays within certain limits. Nylon has prooved to work as Quads show and its mechanism relies on a inherent part of water (~4-6%) and so with varying resistance. Peter Walker once claimed this behaviour even as an advantage of the Nylon coating against other materials.
I used Licron for hybrid panels. Thow such coated panels perform really well, I probabely won´t use Licron any more, because I found it nearly impossible to get even coatings. Its only sold in spray cans and You can´t get an even high quality surface with those. Since I design most of my panels as optically open designs, I prefer to have highly transparent coatings. As Capacity mentioned the thickness of the Licron layer is quite high -in ESL terms of course. I doubt that with diaphragms of more than 6µm thickness You get enough bandwidth.
jauu
Calvin |
|
|
| BillH |
| Calvin, I'm still interested in your formula. Please contact me via email, PM on the board is disabled. |
|
|
| jmateus |
Hi Calvin
I'm building my second pair of ESL's. The first one I used a coating
supplied by Rob McKinlay of ERAudio with good results, except
that it's not immune to humidity. Besides it's kind of expensive
and with shipping from Australia becomes a little more, of course.
My second pair of ESL's are in the works, the membrane is not
stretched yet and the compound is almost finished and is not enough for these six panels.
So, I would be very much interested in knowing your formula
even if this represents some expense for me.
My e-mail is
Joaomateus11@earthlink.net
I'd appreciate. |
|
|
| Lucius |
I'm getting the "ER" coating. Its been reformulated. I bet "EC" is just as good
but there are some other items ER has that I want. The price is not like buying
an expensive stepup trafos.
The secret coating posts in this thread remind me when a lady in a posh New York
restaurant asked the chef for the recipe on the desert she loved. He gave it to her in writing & there was a big $$ charge on the bottom of her dinner tab. She was so
outraged it made the newspapers. Better than a lemon meringue pie in the face
I always say.
Lucius |
|
|
| Calvin |
Hi,
well guys, I´d like to comment on some things I find quite strange.
The way somebody handles the informations he/she is willing to share is solely up to him/here. To give infos openly always carrys the risk that someone else simply copies and sells the stuff. I dislike such a behaviour and so I decide from case to case who I like to give an info.
What makes me really angry is, that guys asking for infos have theire PMs inactive or even better..mails can´t be delivered without registering -ey Joao!! It usually takes me some time to write good and understandably infos and I have no understanding for guys asking for infos but not providing for easy deliverance. You want infos? Than make it possible!
jauu
Calvin |
|
|
| Lucius |
| quote: | | What makes me really angry is, that guys asking for infos have theire PMs inactive or even better..mails can´t be delivered without registering -ey Joao! |
Mea Culpa.. I didn't realise my E-Mail link for this forum was expired. So if anyone did
try to get a hold of me it works now. Calvin, even if you did not E-mail me, thanks
for bringing this up.
Lucius |
|
|
| jmateus |
Hello Calvin
Thanks very much for the info.
I received this morning.
Thanks again.
Joao |
|
|
| tommak |
| quote: | Originally posted by Calvin
Hi,
well guys, I´d like to comment on some things I find quite strange.
The way somebody handles the informations he/she is willing to share is solely up to him/here. To give infos openly always carrys the risk that someone else simply copies and sells the stuff. I dislike such a behaviour and so I decide from case to case who I like to give an info.
What makes me really angry is, that guys asking for infos have theire PMs inactive or even better..mails can´t be delivered without registering -ey Joao!! It usually takes me some time to write good and understandably infos and I have no understanding for guys asking for infos but not providing for easy deliverance. You want infos? Than make it possible!
jauu
Calvin |
Hi Calvin,
I'm still interested in your coating formula, if you would care to share it. I haven't built any ESL panels yet, as I'm still in the 'imformation collection' stage, reading as much as I can find and trying to learn the benefits and issues of the various approaches. I would be interested in your 'formula' to continue my education and hope of building a 'good' ESL on my first attempt.
I don't have any commercial interest in selling your coating or speakers that would use it.
I can be reached at tommak@mindspring.com
Thanks for your contributions to the diyAudio forum!
Regards,
Tom |
|
|
| Paul W |
Hi Calvin,
I do not know why PM did not work because it is supposed to be enabled. Anyhow, EM address is pnwright3@yahoo.com
I did not intend to frustrate you...sorry.
Best regards,
Paul |
|
|
| BillH |
Here's a photo of the frames with walnut veneer applied. They have two coats of tung oil and still need a topcoat of polyurethane. After that, a buff with fine steel wool and the finishing will be done. It has been humid here so I'll let them sit for another day before topcoating.
The frame on the left is a front view. I'm going to use a neon lamp on the output of each high voltage power supply. I was going to mount the neon lampholders on the front and centered in the lower part. I'm not sure if I want to drill through the front after seeing the frame with veneer on. The lampholders might end up on the back cover.
The frame on the right is a back view and shows the white back cover with IEC and speaker connectors. The IEC connector on the left is the power in and has an integral 5mm x 20mm fuseholder. The other IEC connector feeds power to the second speaker through a male-female IEC cord. |
|
|
| BillH |
Here's a look at the electricals for one speaker. The rails that hold the back cover in place aren't shown in the picture. The high voltage wire is 18 GA sheathed in 7/32" diameter clear aquarium tubing to give a bit of insurance against arcing.
Follow the link to the .pdf file of the power supply. I measured 360 volts with a VTVM at the output of the step-up transformers, less than I expected. No problem, though. There's still enough voltage at the output of the voltage multiplier at the X10 setting to arc between two test plates of perforated steel 0.080" apart. |
|
|
| Capaciti |
Hi Bill H,
nice work !
I looked at the power supply. You shouldn't connect the output of T2 directly to the ground of the cascade, but implement a 20Mohm resistor to avoid hum by capacitive coupling.
What size is your stepup ? it looks like EI 84. I guess with this small sized stepup you wont drive your ESL fullrange ?
Capaciti |
|
|
| BillH |
Thanks, Capaciti. You can't tell from the picture, but the 20 meg resistor is on the HV bias supply board farthest away from you in the photo. It's made up of ten 2 Megohm resistors. It's a modification I made to Roger Sander's design to use readily available components.
The transformer core measures 73mm wide x 64mm high. The manufacturer recommends a low frequency cutoff of 225 Hz. It won't be used as a full range, I've got a TL woofer to provide the low end. |
|
|
| BillH |
The picture below shows a crossover frequency response chart and the simplified ESL circuit. All resistors shown are made up of multiple 1 watt units to withstand the voltage and current demands of the ESL panels.
The frequency response test setup was frequency generator --> audio transformer --> crossover --> VTVM with capacitors of close to the calculated ESL capacitance installed where the ESL sections would be. Although my method was crude, the graph shows the crossover overlap and cutoffs that gave the smoothest summed response without excessive current draw from the amplifier. Note the linear frequency scale on the chart. It's done with MS Excel and I don't know how to get a Log scale on the x axis. |
|
|
| Capaciti |
Hi BillH,
i didn't mean the 20 meg resistor for the output. I meant an additional resistor between ground of the HV-Supply and the center tap of the stepup.
I looked onto your schematic. What is the purpose of R1,C1,R2,C2 ?
Capaciti |
|
|
| BillH |
| quote: | | Capaciti writes: You shouldn't connect the output of T2 directly to the ground of the cascade, but implement a 20Mohm resistor to avoid hum by capacitive coupling. |
I haven't seen a second 20 Meg resistor used in that position before. I could be wrong, but the ground of the cascade seems to be at 0 volts DC in relation to the output. I'll measure the output after the first rectifier to see if there's any AC signal there when I get home later.
| quote: | | I looked onto your schematic. What is the purpose of R1,C1,R2,C2 ? |
They are RC high pass filters for the high frequency sections of the ESL, based on this. |
|
|
| Bazukaz |
Hi,
Sorry for a bit off-topic reply , but i would like to ask what is the difference between using copper versus aluminium foil for contacts to diapraghm. By the way , i have found a simple tool to cut aluminium foil - razor blade.
Regards,
Lukas. |
|
|
| Capaciti |
Hi Bazukaz,
NEVER use aluminium. There will be an isolating oxide layer by time and contact will be compromised !
capaciti |
|
|
| Calvin |
Hi Bill,
is the tweeter-panel with the same d/s as the bass-mid-range panels?
Than I wouldn´t use any high-pass-filtering, because You need every peace of area for a good efficiency of the speaker. In case the tweeter-panel is designed with smaller d/s You my use an RC-filter, but I´d suggest smaller values for C as well as R.
C can be chosen smaller, but not less than 10 times the speaker capacity. A quick simu based on Your speakers capacity value showed that 10nF + 30kOhms give -3dB freq of ~500Hz, 20kohms ~730Hz and 10kOhms ~1.4kHz.
jauu
Calvin |
|
|
| Capaciti |
Hi BillH,
Calvin is right. If you do not have different D/S spacing you should not use the filter. Any filter compromises the phase response of the system.
Especially onESL you can "hear" any filter involved. |
|
|
|
|
|
